==> picture [215 x 22] intentionally omitted <==

Single Technology Appraisal

Olaparib in combination with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube and peritoneal cancer after response to first-line platinum-based chemotherapy with bevacizumab [ID1652] Committee Papers

==> picture [215 x 22] intentionally omitted <==

NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE

SINGLE TECHNOLOGY APPRAISAL

Olaparib in combination with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube and peritoneal cancer after response to firstline platinum-based chemotherapy with bevacizumab [ID1652]

Contents:

The following documents are made available to consultees and commentators:

The final scope and final stakeholder list are available on the NICE website.

1. Company submission summary from AstraZeneca

2. Clarification questions and company responses

• a. Additional responses

3. Patient group, professional group and NHS organisation submissions from:

• a. Ovacome

• b. Ovarian Cancer Action

• c. Target Ovarian Cancer

4. Expert personal perspectives from:

• a. Prof. Iain McNeish, Professor of Oncology – clinical expert, nominated by AstraZeneca

• b. Dr Susana Banerjee, Consultant Medical Oncologist – clinical expert, nominated by AstraZeneca

• c. Florence Wilks – patient expert, nominated by Ovarian Cancer Action

• d. Rachel Downing, Head of Policy and Campaigns – patient expert, nominated by Target Ovarian Cancer *see item 3c

5. Evidence Review Group report prepared by BMJ-TAG

6. Evidence Review Group report – factual accuracy check

7. Technical report

8. Technical engagement response from company

• a. Response to queries

9. Technical engagement responses from experts:

• a. Prof. Iain McNeish, Professor of Oncology – clinical expert, nominated by AstraZeneca

• b. Dr Susana Banerjee, Consultant Medical Oncologist – clinical expert, nominated by AstraZeneca

==> picture [215 x 22] intentionally omitted <==

10. Technical engagement responses from consultees and commentators:

• a. British Gynaecological Cancer Society

• b. Ovacome

• c. Target Ovarian Cancer

11. Evidence Review Group critique of company response to technical engagement prepared by BMJ-TAG a. Addendum

Any information supplied to NICE which has been marked as confidential, has been redacted. All personal information has also been redacted.

NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE

Single technology appraisal

Olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer

ID1652

Document B

Company evidence submission

13[th] March 2020

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 1 of 177

Contents

Contents ........................................................................................................................................... 2 Tables and figures............................................................................................................................ 3 Abbreviations ................................................................................................................................... 7 B.1 Decision problem, description of the technology and clinical care pathway ........................... 11 B.1.1 Decision problem .............................................................................................................. 11 B.1.2 Description of the technology being appraised ................................................................ 14 B.1.3 Health condition and position of the technology in the treatment pathway ..................... 16 B.1.4 Equality considerations .................................................................................................... 41 B.2 Clinical effectiveness ............................................................................................................... 42 B.2.1 Identification and selection of relevant studies ................................................................ 42 B.2.2 List of relevant clinical effectiveness evidence ................................................................ 43 B.2.3 Summary of methodology of the relevant clinical effectiveness evidence ...................... 44 B.2.4 Statistical analysis and definition of study groups in the relevant clinical effectiveness evidence ..................................................................................................................................... 54 B.2.5 Quality assessment of the relevant clinical effectiveness evidence ................................ 57 B.2.6 Clinical effectiveness results of the relevant trials ........................................................... 57 B.2.7 Subgroup analysis ............................................................................................................ 74 B.2.9 Indirect and mixed treatment comparisons ...................................................................... 75 B.2.10 Adverse reactions ........................................................................................................... 76 B.2.11 Ongoing studies ............................................................................................................. 88 B.2.12 Innovation ....................................................................................................................... 89 B.2.13 Interpretation of clinical effectiveness and safety evidence ........................................... 94 B.3 Cost effectiveness ................................................................................................................... 99 B.3.1 Published cost-effectiveness studies ............................................................................. 100 B.3.2 Economic analysis .......................................................................................................... 109 B.3.3 Clinical parameters and variables .................................................................................. 118 B.3.4 Measurement and valuation of health effects ................................................................ 137 B.3.5 Health-related quality-of-life data used in the cost-effectiveness analysis .................... 141 B.3.6 Cost and healthcare resource use identification, measurement and valuation ............. 141 B.3.7 Summary of base case analysis inputs and assumptions ............................................. 150 B.3.8 Base case results ........................................................................................................... 151 B.3.9 Sensitivity and scenario analyses .................................................................................. 153 B.3.10 Subgroup analysis ........................................................................................................ 158 B.3.11 Extended regimen analysis .......................................................................................... 158 B.3.12 Validation ...................................................................................................................... 164 B.3.13 Interpretation and conclusions of economic evidence ................................................. 165 References ................................................................................................................................... 167 Appendices .................................................................................................................................. 177

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 2 of 177

Tables and figures

Table 1. The decision problem ...................................................................................................... 13 Table 2. Technology being appraised............................................................................................ 14 Table 3: PAOLA-1 study design .................................................................................................... 43 Table 4. Allowed concomitant medications during study treatment, FAS ..................................... 49 Table 5. Patient characteristics in PAOLA-1 ................................................................................. 53 Table 6. Overview of quality assessments for PAOLA-1 ............................................................... 57 Table 7. Summary of efficacy at the 22 March 2019 DCO, FAS and HRD-positive population ... 58 Table 8. Progression status at time of PFS analysis (22 March 2019 DCO), FAS ....................... 59 Table 9: Summary of PFS analysis by investigator assessment and BICR (22 March 2019 DCO), FAS ................................................................................................................................................ 60 Table 10. Progression status at time of investigator-assessed PFS analysis (22 March 2019 DCO), HRD-positive population ..................................................................................................... 62 Table 11. Summary of PFS analysis by investigator assessment (22 March 2019 DCO), HRDpositive population ......................................................................................................................... 63 Table 12. TFST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ..................................................................................................... 64 Table 13. Post-discontinuation anticancer therapy, AZ Medic review, HRD-positive population . 65 Table 14. Time to second progression (PFS2) for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population .................................................... 66 Table 15. TSST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ..................................................................................................... 67 Table 16. Second post-discontinuation anticancer therapy, investigator review, HRD-positive population ....................................................................................................................................... 68 Table 17. OS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ................................................................................................................ 69 Table 18. Best objective response in patient with radiological evidence of disease, any target or non-target lesions; olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ..................................................................................................... 70 Table 19. Duration of bevacizumab exposure (22 March 2019 DCO), SAS and HRD-positive population ....................................................................................................................................... 77 Table 20. Duration of olaparib or placebo exposure (22 March 2019 DCO), SAS and HRDpositive population ......................................................................................................................... 78 Table 21. Summary of adverse events (22 March 2019 DCO), SAS and HRD-positive population ........................................................................................................................................................ 81 Table 22. Most common AEs (all grades), occurring in ≥10% of patients in either treatment arm (SAS) .............................................................................................................................................. 82 Table 23. AEs of CTCAE Grade 3 or higher, >1% in either treatment arm (SAS) ........................ 83 Table 24. Summary of SAEs (SAS) ............................................................................................... 85 Table 25. AEs of special interest for olaparib (SAS) ..................................................................... 86 Table 26. Bevacizumab ADRs in either treatment arm (SAS) ...................................................... 87 Table 27. All deaths in the PAOLA-1 study (SAS) ........................................................................ 88 Table 28. Results of the population-adjusted indirect comparison (PAIC): PFS outcomes for the weighted BRCA -mutated subset of PAOLA-1 and unweighted SOLO-1 ...................................... 91 Table 29. Results of the population-adjusted indirect comparison (PAIC): PAOLA-1 and PRIMA (HRD-positive) ................................................................................................................................ 93 Table 30. Summary of UK cost-effectiveness analyses considering maintenance treatments for ovarian cancer (N=11) ................................................................................................................. 102 Table 31. Summary of the base case de novo maintenance economic analysis ....................... 110 Table 32. Features of the economic analysis and comparisons with previous appraisals in the first-line advanced ovarian cancer setting ................................................................................... 117 Table 33. Summary of combined AIC and BIC goodness of fit data for PFS ............................. 121 Table 34: Comparison of KM data, empirical data, RWE and long-term extrapolation of PFS for the placebo + bevacizumab arm using fully fitted parametric model methods ........................... 123 Table 35. Literature estimates of PFS in patients with Stage III or IV ovarian cancer ................ 123

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 3 of 177

Table 35: Comparison of KM data and long-term extrapolation of PFS for the olaparib + bevacizumab arm using fully fitted parametric model methods .................................................. 124 Table 37. Goodness of fit for PFS using mixture models ............................................................ 126 Table 39. Comparison of KM data, empirical data, RWE and long-term extrapolation of PFS for the placebo + bevacizumab arm using PMM .............................................................................. 127 Table 40. Comparison of KM data and long-term extrapolation of PFS for the olaparib + bevacizumab arm using PMM ..................................................................................................... 127 Table 38. Fitted parameters for the Weibull mixed survival model distribution ........................... 128 Table 41. Summary of combined AIC and BIC goodness of fit data for PFS2 ........................... 128 Table 42. Comparison of KM data and long-term extrapolation of PFS2 for the routine placebo plus bevacizumab arm using PMM .............................................................................................. 130 Table 43. Comparison of KM data and long-term extrapolation of PFS2 for the olaparib plus bevacizumab arm using PMM ..................................................................................................... 130 Table 44. Summary of combined AIC and BIC goodness of fit data for OS ............................... 132 Table 46. Comparison of KM data, empirical data, RWE and long-term extrapolation of OS for the placebo + bevacizumab arm using parametric models ......................................................... 135 Table 45. Comparison of KM data and long-term extrapolation of OS for the olaparib + bevacizumab arm using parametric models ................................................................................ 135 Table 47. Incremental OS benefit accepted in previous NICE appraisals .................................. 136 Table 48. Summary of AEs included in the economic model ...................................................... 136 Table 49. Utility values associated with specific disease stages/states ...................................... 138 Table 50. Disutility values associated with AEs, and assumed duration of events ..................... 140 Table 51. Summary of utility values for cost-effectiveness analysis ........................................... 141 Table 52. Summary of olaparib drug related costs ...................................................................... 143 Table 53. Subsequent PARPi use in the PAOLA-1 HRD-positive population* ........................... 145 Table 54. Drug acquisition costs – subsequent therapies received by patients in the PAOLA-1 study ............................................................................................................................................. 145 Table 55. Chemotherapy recommended dose and duration of treatment .................................. 146 Table 56: Administration costs ..................................................................................................... 146 Table 57. Unit costs for AEs in the model ................................................................................... 147 Table 58. Unit costs and monthly frequency of resource use associated with the PF and PD states for placebo + bevacizumab ............................................................................................... 148 Table 59. Unit costs and monthly frequency of resource use associated with the PF and PD states for olaparib + bevacizumab ............................................................................................... 148 Table 60. Resource costs (per week) associated with the monitoring and management of patients treated with olaparib or routine surveillance in the model ............................................. 149 Table 61. Summary of assumptions in the model ....................................................................... 150 Table 62. Base case results versus bevacizumab 15 mg/kg maintenance (deterministic) ........ 152 Table 63: Results versus bevacizumab 7.5mg/kg maintenance (deterministic) ......................... 152 Table 64: Results versus routine surveillance (deterministic) ..................................................... 153 Table 65: Base case results versus bevacizumab 15 mg/kg maintenance (probabilistic) .......... 154 Table 66: Analysis results versus bevacizumab 7.5mg/kg (probabilistic) ................................... 155 Table 67: Analysis results versus routine surveillance (probabilistic) ......................................... 155 Table 68. scenario analyses from the CUA (ICERs vs bevacizumab 15 mg/kg maintenance [base case], bevacizumab 7.5 mg/kg maintenance [CDF], and routine surveillance) .......................... 157 Table 69: Assumptions for calculating one-off cost adjustment due to all patients receiving chemotherapy + bevacizumab 15mg/kg ...................................................................................... 160 Table 70: Assumptions for calculating one-off adjustment cost for all patients receiving chemotherapy + bevacizumab 15mg/kg ...................................................................................... 161 Table 71: Assumptions for calculating one-off adjustment cost for all patients receiving chemotherapy + bevacizumab 15mg/kg maintenance ................................................................ 162 Table 72: A summary of the one-off cost adjustments accrued by treatment and comparator arms ...................................................................................................................................................... 163 Table 73. Extended regimen analysis to address the full treatment sequence captured in the NICE scope .................................................................................................................................. 163

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 4 of 177

Figure 1. Rationale for seeking an optimised recommendation for HRD-positive patients .......... 11 Figure 2. The treatment sequence (first-line and maintenance) encompassed in the NICE scope ........................................................................................................................................................ 12 Figure 3: Five-year survival rates for ovarian cancer* in England ................................................. 18 Figure 4. In England, the majority (~60%) of women with ovarian cancer have advanced (FIGO Stage III−IV) disease at the time of diagnosis ............................................................................... 19 Figure 5: Distribution of HRD mutations in advanced ovarian cancer .......................................... 20 Figure 6: Women with HRD-positive tumours achieve significantly better survival outcomes after first-line chemotherapy ................................................................................................................... 21 Figure 7. The patient population covered by the company submission ........................................ 21 Figure 8. Progression-free survival in the overall ICON8 study population .................................. 24 Figure 9: CDF criteria to receive bevacizumab ............................................................................. 25 Figure 10: Proportions of women eligible to receive bevacizumab for advanced ovarian cancer via the CDF .................................................................................................................................... 26 Figure 11. The CDF recommendation for olaparib was described as a “new era” for the treatment of women with newly-diagnosed BRCA -mutated advanced ovarian cancer................................. 27 Figure 12: Gynaecologic Cancer Intergroup (GCIG) responses to platinum chemotherapy ........ 29 Figure 13: Progression-free survival outcomes for women with newly diagnosed and platinum sensitive relapsed ovarian cancer ................................................................................................. 30 Figure 14: Not all women are able to benefit from PARPi therapy upon relapse ......................... 30 Figure 15: The mechanism of action for bevacizumab .................................................................. 33 Figure 16: Comparison of A) PFS and B) OS between 7.5 mg/kg Q3W and 15 mg/kg Q3W bevacizumab doses ....................................................................................................................... 34 Figure 17: Pathways for DNA repair following treatment with PARPi ........................................... 35 Figure 18. Long-term follow-up data show that ~20% of women* do not require subsequent anticancer therapy following maintenance treatment with olaparib for relapse, platinum-sensitive advanced ovarian cancer ............................................................................................................... 36 Figure 19. Study 19 data* show that extensions in TFST and TSST translate into a significant long-term overall survival benefit in favour olaparib (versus placebo) .......................................... 37 Figure 20. Population adjusted analysis: PFS outcomes for the weighted BRCA -mutated subset of PAOLA-1 and unweighted SOLO-1 ........................................................................................... 39 Figure 21. Pooled matching analysis of PRIMA and PAOLA-1 (HRD-positive patients) .............. 40 Figure 22. Overview of study design for PAOLA-1 ........................................................................ 46 Figure 23. Hierarchical testing strategy employed in PAOLA-1 .................................................... 56 Figure 24. Kaplan-Meir plot of investigator-assessed PFS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), FAS (top) and HRD-positive (bottom) .............. 61 Figure 25. TFST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ..................................................................................................... 64 Figure 26. Time to second progression (PFS2) for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population .................................................... 67 Figure 27. TSST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ..................................................................................................... 68 Figure 28. OS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population ................................................................................................................ 69 Figure 29. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group: Global health status/QoL change from baseline (22 March 2019 DCO), HRDpositive population ......................................................................................................................... 71 Figure 30. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group, EORTC-QLQ-C30 functional scale – role functioning; change from baseline (22 March 2019 DCO), HRD-positive population ................................................................................. 72 Figure 31. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group, EORTC-QLQ-C30 functional scale – emotional functioning; change from baseline (22 March 2019 DCO), HRD-positive population ............................................................ 72 Figure 32. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group, EORTC-QLQ-C30 functional scale – social functioning; change from baseline (22 March 2019 DCO), HRD-positive population .......................................................................... 73

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 5 of 177

Figure 33. Mean (± SD) EQ-5D-5L weighted health state index change from baseline across time points by treatment group (22 March 2019 DCO), HRD-positive population ................................ 74 Figure 34. Safety analysis phases ................................................................................................. 76 Figure 35. Time to treatment discontinuation or death (TDT; 22 March 2019 DCO), HRD-positive population ....................................................................................................................................... 79 Figure 36: EMA approvals for PARPi for the maintenance treatment of ovarian cancer .............. 89 Figure 37. Population-adjusted analysis: PFS outcomes for the weighted BRCA -mutated subset of PAOLA-1 and unweighted SOLO-1 (same as Figure 20) ......................................................... 91 Figure 38. Pooled matching analysis of PRIMA and PAOLA-1 (HRD-positive patients) (same as Figure 21) ....................................................................................................................................... 93 Figure 40. Model schematic ......................................................................................................... 113 Figure 41. Illustration of the partitioned survival calculation ........................................................ 115 Figure 42. KM plot of investigator-assessed PFS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population .................................................. 119 Figure 43. Cumulative hazards plot of PFS (HRD-positive population) ...................................... 120 Figure 44. Schoenfeld residuals of PFS (HRD-positive population) ........................................... 120 Figure 45. Visual representation of fitted parametric models to entire HRD positive data set ... 121 Figure 46: PFS in ICON8 trial ...................................................................................................... 122 Figure 47: PFS extrapolation using best fitting model ................................................................. 128 Figure 48. Time to second progression (PFS2) for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population .................................................. 128 Figure 49. Cumulative hazards plot of PFS2 (HRD-positive population) .................................... 129 Figure 50. Schoenfeld residuals of PFS2 (HRD-positive population) ......................................... 129 Figure 51. Visual representation of fitted PFS2 parametric models to entire HRD positive data set ...................................................................................................................................................... 131 Figure 52. OS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population .............................................................................................................. 132 Figure 53 Cumulative hazards plot of OS (HRD-positive population) ......................................... 133 Figure 54 Schoenfeld residuals of OS ......................................................................................... 133 Figure 55. Visual representation of fitted OS parametric models to entire data set ................... 134 Figure 56. Cost-effectiveness plane, olaparib + bevacizumab 15 mg/kg versus bevacizumab 15 mg/kg ............................................................................................................................................ 154 Figure 57. Cost-effectiveness acceptability curve, olaparib + bevacizumab 15 mg/kg versus bevacizumab 15 mg/kg ................................................................................................................ 155 Figure 58. Deterministic sensitivity analysis results (ICERs) ...................................................... 156

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 6 of 177

Abbreviations

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 7 of 177

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 8 of 177

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 9 of 177

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 10 of 177

B.1 Decision problem, description of the technology and clinical care pathway

B.1.1 Decision problem

The regulatory submission for the olaparib indication relevant to this appraisal was provided to the European Medicines Agency (EMA) on the xx[xx] xxxxxxxxxxxxxxxxx. The anticipated marketing authorisation is aligned to the full analysis set (FAS) of the pivotal Phase III PAOLA-1 study, i.e.,

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

This submission will focus on part of the population covered by the PAOLA-1 study, i.e. women whose tumours indicate homologous recombination deficiency (HRD) using a validated test. This decision was made in order to focus the appraisal on the population of women where the addition of olaparib to bevacizumab maintenance treatment has shown a consistent and substantial clinical benefit versus bevacizumab alone, across a range of clinically meaningful endpoints (see Section B.2.6) and is anticipated to be a highly cost-effective use of NHS resources.

Figure 1. Rationale for seeking an optimised recommendation for HRD-positive patients

*, This subgroup included women with both BRCA m and BRCA wt tumours. Importantly, a similar level of PFS benefit (as stated above, HR=0.33), was observed amongst women with HRD-positive but BRCA wt tumours (HR 0.43; 95% CI, 0.28−0.66), indicating that the treatment effect was not driven entirely by the BRCA m population.[†] , An additional PFS

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 11 of 177

benefit for olaparib added to bevacizumab, versus placebo + bevacizumab, was also seen in patients whose tumour HRD status was unknown (for instance, due to failed tests or availability of insufficient tumour samples). Abbreviations : CI: confidence interval; FAS: full analysis set; HR: hazard ratio; HRD: homologous recombination deficiency; NHS: National Health Service; NICE: National Institute for Health and Care Excellence; OS: overall survival; PFS: progression-free survival; PFS2: second progression-free survival; RECIST: Response evaluation criteria in solid tumours; TSFT: time to first subsequent therapy; TSST: time to second subsequent therapy.

The decision problem addressed by this submission is summarised in Table 1. It is important to highlight that the scope of the decision problem is broader than the maintenance setting where the PAOLA-1 regimen (i.e. olaparib, added to bevacizumab) will be used. This extended scope was specified by NICE considering the requirement for all women to have received (and responded to) bevacizumab 15mg/kg every three weeks (Q3W) in combination platinum-taxane chemotherapy, to be eligible for the PAOLA-1 regimen, and the upstream consequences of this requirement on current first-line clinical practice (where women either receive chemotherapy on its own or in combination with a lower bevacizumab dose of 7.5mg/kg) (see Figure 2).

Capturing the full treatment sequence specified by NICE in either the clinical- or cost-effectiveness analysis requires outcome data that were not captured in PAOLA-1 and are unavailable from previous bevacizumab studies (e.g. PFS and OS by response to therapy), where AstraZeneca does not have access to the patient level data. As such, there was no direct way of addressing the full scope (discussed in Section B.2.9). Nonetheless, we provided two different approaches to addressing this – these approaches (which give remarkably similar results) reflect our best attempts to fulfil this challenging scope and are described in Section 3.2 in further detail.

Figure 2. The treatment sequence (first-line and maintenance) encompassed in the NICE scope

==> picture [452 x 264] intentionally omitted <==

Note: It is anticipated that “current” clinical practice in England will change in the coming months, with loss-ofexclusivity of Avastin[®] and multiple biosimilar entries (at lower prices negotiated through national tenders) facilitating bevacizumab use in routine NHS commissioning. Abbreviations: EMA: European Medicines Agency; MA: marketing authorisation. Source: NICE Final Scope.[2 ]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 12 of 177

Table 1. The decision problem

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 13 of 177

Abbreviations: CDF: Cancer Drugs Fund; EMA: European Medicines Agency; HRD: homologous recombination deficiency; LoE: loss of exclusivity; N/A: not applicable; NHS: National Health Service; NICE: National Institute for Health and Care Excellence. Source: NICE Final Scope.[2 ]

B.1.2 Description of the technology being appraised

A description of the technology being appraised is summarised in Table 2. The SmPC for olaparib in this indication was not available at the time of writing this document; AstraZeneca will share this with NICE, when possible.

Table 2. Technology being appraised

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 14 of 177

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 15 of 177

*, Patients with evidence of disease at two years, who in the opinion of the treating physician can derive further benefit from continuous olaparib treatment, can be treated beyond two years. In PAOLA-1, most patients came off-treatment at the first scheduled follow-up visit after two years (week 108 or month 25). Just 5 patients in the olaparib + bevacizumab arm remained on treatment by month 26; by month 30, just 2 patients remained on treatment.

Abbreviations: 1L: first-line; BRCA1/2 : breast cancer susceptibility gene; DNA: deoxyribonucleic acid; DSB: double-strand break; EC: European Commission; EMA: European Medicines Agency; HRD: homologous recombination deficiency; HRR: homologous recombination repair; NICE: National Institute for Health and Care Excellence; PARP: poly-ADP ribose polymerase; PARPi: PARP inhibitor; SmPC: summary of product characteristics; SSB: single-strand break; Source: Olaparib SmPC[4] ; Bevacizumab SmPC.[5]

B.1.3 Health condition and position of the technology in the

treatment pathway

B.1.3.1 Disease overview

• Approximately 7,000 women are diagnosed with ovarian cancer in England every year.[6]

• Due to the non-specific nature of symptoms, the majority of women (~60%) have advanced (FIGO Stage III−IV) disease at the time of diagnosis.[6]

• Advanced disease is typically associated with a poor prognosis:

  • The establishment of specialist gynaecological oncology centres with specialist MDTs and increase in surgical radicality have increased ovarian cancer survival rates over the last two decades. However, there remains an unmet need to further improve outcomes for this disease (5-year survival for advanced disease was <35% between 2013−2017.[6]

• High-grade serous ovarian cancer (HGSOC) is the most common histological subtype in women with advanced (FIGO Stage III or IV) disease, constituting nearly 90% of all cases.[7-] 9

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 16 of 177

• Published data shows that ~50% of HGSOC tumours indicate homologous recombination deficiency (HRD).

• HRD-positive tumours are highly-sensitive to cytotoxic chemotherapy as well as targeted PARPi therapy.[10-12]

• Women with HRD-positive disease achieve significantly longer progression-free survival (PFS) and overall survival (OS) after first-line platinum chemotherapy, and are thus prognostically different to those with HRD-negative diease.[13]

• HRD-positive advanced ovarian cancer – the focus of this submission- represents ~25% of the overall population of women diagnosed with ovarian cancer each year in England (Figure 7).

Ovarian cancer is one of the most common cancers in women

“Ovarian cancer” is a non-specific term used to describe cancers that originate in the ovary, fallopian tube, and primary peritoneum.[14, 15] Approximately 7,000 women are diagnosed with ovarian cancer every year in England (6,902 on average every year between 2015 and 2017; Ovarian Cancer Audit Feasibility Pilot).[6]

• In the most recent National Audit (2015−2017 audit period), the age standardised incidence rates of ovarian cancer across 19 Cancer Alliances in England ranged from 21.8 to 27.5 cases per 100,000 person-years.[6]

• On average, a woman in the UK has a one in 50 chance of being diagnosed with ovarian cancer in her lifetime.[16]

The severity of ovarian cancer is captured by disease stage

In the UK, ovarian cancer is staged according to the International Federation of Gynaecology and Obstetrics (FIGO) classification system (Figure 4; further described in Appendix N).

Advanced ovarian cancer encompasses FIGO Stages III and IV.

• Stage III disease is characterised by extra-pelvic spread of the cancer, into the abdominal cavity, or lymph nodes,

• Stage IV disease involves more distant metastases, for example, to the abdominal viscera or lungs.[15]

Both Stage III and IV ovarian cancer are further classified into sub-stages (Figure 4; Appendix N).

The majority of women in England (~60%) have advanced (FIGO Stage III−IV) disease at the time of diagnosis[6]

Ovarian cancer can be difficult to diagnose due to its asymptomatic nature and non-specificity of symptoms, especially in the early stages of disease[*] .[17, 18] This, together with the absence of validated screening programmes, leads to the majority of women being diagnosed with advanced

• Commonly reported symptoms include frequent and persistent abdominal distention (bloating), loss of appetite, pelvic or abdominal pain, and increased urinary urgency and/or frequency. Other symptoms of ovarian cancer may include irregular periods, lower abdominal and back pain, constipation, nausea, anorexia, dyspepsia, extreme fatigue and post-menopausal or rectal bleeding.[17, 18 ]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 17 of 177

(Stage III or IV) ovarian cancer, when the disease has already spread (Figure 4).[14, 19-21] Further details on ovarian cancer symptoms and diagnostic guidelines are provided in Appendix N.

The late diagnosis of ovarian cancer contributes towards the poor prognosis associated with this condition:[6]

• Five-year survival range from 68%−93% in women with Stage I−II ovarian cancer, versus just 13%−27% in women with Stage III−IV disease[†] .[22]

Although net survival rates for ovarian cancers have continually improved in the last two decades, due to the establishment of specialist gynaecological oncology centres with specialist MDTs and increased surgical radicality for ovarian malignancies (with five-year rates improving from 25.7% in 2001−2005 to 34.7% in 2013−2017; Figure 3)[6] , there remains an unmet need to further optimise outcomes for this aggressive disease.[23]

Figure 3: Five-year survival rates for ovarian cancer in England*

==> picture [411 x 270] intentionally omitted <==

----- Start of picture text -----
50
45
40
35
30
25
20
Diagnosis year range
Net Survival 5 year (excl. borderlines) Net Survival 5 year (incl. borderlines)
Net survival rate (%)
----- End of picture text -----

*Also includes fallopian tube and primary peritoneal carcinomas.

Note : Borderline ovarian tumours are defined histologically by atypical epithelial proliferation without stromal invasion. They tend to grow slowly and in a more controlled way than ovarian carcinomas, and most women are cured following surgery.

Source: Adapted from Ovarian Cancer Audit, Public Health England.[6]

† Survival rates include all recorded cases, including women who would have been too unwell to undergo active anticancer therapy at the time of diagnosis. This population is thus prognostically different (worse) relative to the selected subgroup of women who would be eligible to receive olaparib + bevacizumab maintenance therapy in real-world practice.

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 18 of 177

Figure 4. In England, the majority (~60%) of women with ovarian cancer have advanced (FIGO Stage III−IV) disease at the time of diagnosis

==> picture [694 x 311] intentionally omitted <==

----- Start of picture text -----
% of newly-diagnosed 6%
cases in England:
22%
33% 38%
Focus of this NICE submission
----- End of picture text -----

*Patients with ovary, fallopian tube and primary peritoneal carcinomas excluding borderlines; %s of cases with known / recorded stage. Abbreviations: FIGO: International Federation of Gynecology and Obstetrics; NICE: National Institute for Health and Care Excellence. Source: Image adapted from Cancer Research UK, 2014;[24] Data from the Ovarian Cancer Audit Feasibility Pilot. Disease Profile in England: Incidence, mortality, stage and survival for ovary, fallopian tube and primary peritoneal carcinomas. January 2020.[6]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652]

© AstraZeneca (2020). All rights reserved

High-grade serous ovarian cancer (HGSOC) is the most common histological subtype reported in women with advanced (FIGO Stage III or IV) disease

The vast majority (90%) of ovarian cancers are classified as epithelial cancers (with the remaining being tumours of rarer origins, such as germ cell tumours, stromal tumours, and sarcomas).[8, 9]

Epithelial tumours are grouped by histology into five main sub-types - high-grade serous ovarian cancer (HGSOC) is by far the most common, constituting nearly 90% of advanced (FIGO Stage III and IV) cases (further information provided in Appendix N).[7-9]

Approximately half of HGSOC tumours are deficient in homologous recombination, the main high-fidelity pathway of DNA double-strand break repair in human cells

HGSOC is a highly-mutated cancer.[10] A detailed analysis conducted by The Cancer Genome Atlas (TCGA) programme showed that ~50% of HGSOC tumours have identifiable defects in the homologous recombination (HR) pathway, the main high-fidelity pathway of DNA double-strand break (DSB) repair in human cells (Figure 5)[10, 25]

This creates an opportunity for utilising therapeutic interventions such as PARPi, which, through mechanisms involving “ synthetic lethality ” can selectively target these tumour cells.[26] The mechanism of action of PARPi and the concept of synthetic lethality are described in B.1.3.3.

Figure 5: Distribution of HRD mutations in advanced ovarian cancer

==> picture [367 x 289] intentionally omitted <==

Note: PTEN deletion and EMSY amplification have been reported to confer HRD, but data are evolving and therefore both have been classified as ‘Possibly HRD’. Abbreviations: BRCA , breast cancer susgene; CCNE1, cyclin E1; HRD, homologous recombination deficiency Source: Adapted from Konstantinopoulos et al. 2015;[27] Hollis and Gourley, 2016.[9]

Women with HRD-positive tumours are more sensitive to cytotoxic chemotherapy and achieve enhanced survival outcomes relative to those with HRD-negative disease (Figure 6).[12, 13] This is

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 20 of 177

relevant in the context of this appraisal, which specifically focuses on women with HRD-positive advanced ovarian cancer who have responded to first-line chemotherapy.

Figure 6: Women with HRD-positive tumours achieve significantly better survival outcomes after first-line chemotherapy

==> picture [453 x 146] intentionally omitted <==

Data from the Phase III SCOTROC4 randomised controlled trial (RCT) in women with Stage IC−IV ovarian cancer, who had received first-line carboplatin chemotherapy.

Abbreviations: HGSOC, high-grade serous ovarian cancer; HRD, homologous recombination deficiency; OS, overall survival; PFS, progression-free survival Source: Adapted from Stronach et al., 2018.[13]

This population of women, with HRD-positive advanced ovarian cancer, constitutes ~25% of the overall population diagnosed each year in England and is the focus of this submission (Figure 7).

Figure 7. The patient population covered by the company submission

==> picture [389 x 232] intentionally omitted <==

----- Start of picture text -----
~7,000 [1] women diagnosed with
ovarian cancer every year
~60% [1] have advanced (FIGO
Stage III−IV disease) at diagnosis
90% [2-4] with advanced
disease have high-grade
serous OC (HGSOC)
Of the 7,000 women diagnosed
50% [5, 6 ]
with ovarian cancer every year:
HGSOCs indicate
HRD ~25% have HRD-positive
advanced ovarian cancer
----- End of picture text -----

Note: The HRD-positive population estimate of 25% is calculated as follows: ~7,000*60% * 90% * 50%) Abbreviations HGSOC: high grade serous ovarian cancer; HRD: homologous recombination deficiency. Source : 1. Data from the Ovarian Cancer Audit Feasibility Pilot, 2020.[6] ; 2. Chan et al., 2008[7] ; 3. Bookman et al., 2014[8] 4. Hollis and Gourley, 2016[9] ; 5. The Cancer Genome Atlas Research Network, 2011[10] ; 6. Yemelyanova, 2011[25]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 21 of 177

B.1.3.2 Clinical pathway of care for advanced ovarian cancer

Treatment of newly-diagnosed ovarian cancer

• Treatment plans for women diagnosed with ovarian cancer in England are determined by specialist gynaecological cancer MDTs at specialist gynaecological oncology centres.[28]

• Complete or optimal cytoreduction (where achievable) is the standard-of-care for advanced ovarian cancer patients with good performance status (PS).[29]

  • Primary debulking surgery is recommended in patients where complete or optimal cytoreduction appears achievable.[18]

  • Where this is not possible, neoadjuvant chemotherapy followed by interval debulking surgery is considered non-inferior to upfront surgery.[18][30, 31]

• BGCS and NICE guidelines recommend cytotoxic chemotherapy following surgery, to reduce the risk of disease recurrence.[17, 18] Carboplatin in combination with paclitaxel (NICE TA55) has been the preferred chemotherapy regimen in this setting for multiple decades.[32]

• In 2013, bevacizumab in combination with carboplatin and paclitaxel, followed by bevacizumab maintenance, was made available through the CDF for newly-diagnosed advanced ovarian cancer patients.[33]

The addition of bevacizumab (in combination with chemotherapy followed by maintenance treatment) confers a further progression-free survival advantage , relative to chemotherapy followed by routine surveillance, and is the standard-of-care for eligible patients (estimated to be ~80% of the advanced ovarian cancer patient population) .[6, 34-36]

Treatment plans for women diagnosed with ovarian cancer in England are determined by specialist gynaecological cancer MDTs, housed at specialist gynaecological oncology centres. These were established throughout the country between 2000 and 2005, following the publication of guidelines on Improving Outcomes in Gynaecological Cancers.[28]

Gynaecological cancer MDTs typically include (but are not limited to) specialist gynaecological oncology surgeon(s), clinical / medical oncologist(s), as well as gynae-oncology nurse specialists, radiologists, and pathologists. Treatment decisions are based on: disease stage and grade; histological and molecular subtype; patients’ age, PS, co-morbidities (if any), and preference; as well as quality-assured institutional expertise.[28] An overview of the current treatment pathway is provided below:

==> picture [469 x 69] intentionally omitted <==

----- Start of picture text -----
2. Platinum-based
3. Bevacizumab
1. Surgery ± NACT chemotherapy ± maintenance*
bevacizumab*
*For eligible patients. NACT: neoadjuvant chemotherapy
----- End of picture text -----

Complete or optimal cytoreduction (where achievable) is the standard-of-care for advanced ovarian patients with good PS

The goal of surgery in ovarian cancer is to achieve complete resection of all macroscopic disease.[29] British Gynaecological Cancer Society (BGCS) guidelines recommend primary or

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 22 of 177

upfront debulking surgery (PDS or UDS, respectively) in patients with good PS, where complete or optimal cytoreduction appears achievable.[18]

In instances where this is not achievable (e.g. due to a patients’ PS or spread of disease, such that an optimal debulking procedure is unlikely), neoadjuvant chemotherapy (NACT) followed by interval debulking surgery (IDS) is considered to be non-inferior to upfront surgery.[18]

• This recommendation is based on the results of two randomised trials (CHORUS and EORTC55971) that have shown similar PFS and OS in advanced ovarian cancer patients receiving NACT and IDS, compared with PDS.[30, 31]

The selection of patients for PDS or NACT and IDS is carried out at specialist ovarian cancer centres in an MDT setting, according to the European Society for Gynaecological Oncology (ESGO) quality recommendations (2016).[37]

It should be noted that not all advanced ovarian cancer patients are suitable candidates for surgery (e.g. due a low likelihood of achieving no residual disease with reasonable morbidity, patient’s PS, and disease grade / pathology).[37] Cytotoxic chemotherapy, with or without an anti-angiogenic agent (described below), is considered for these patients, depending on their fitness.

2. Platinum-based 3. Bevacizumab 1. Surgery ± NACT chemotherapy ± maintenance* bevacizumab* *For eligible patients. NACT: neoadjuvant chemotherapy.

BGCS and NICE guidelines recommend cytotoxic chemotherapy following surgery, depending on patient fitness

Surgery is followed by chemotherapy to reduce the risk of disease recurrence.[17, 18, 20] The preferred chemotherapy regimen is carboplatin (area under the curve [AUC] 5/6) alone, or in combination with paclitaxel (175 mg/m[2] ; NICE TA55).[18, 20, 29, 32] Both are administered intravenously every three weeks (Q3W), for six cycles.[32]

The combination of cisplatin and paclitaxel is equally effective but is more toxic and less convenient to administer.[32] For patients who develop an allergy to or do not tolerate paclitaxel, BGCS and the European Society of Medical Oncology (ESMO) guidelines indicate that docetaxel or pegylated liposomal doxorubicin hydrochloride (PLDH) may be considered as alternatives.[17, 18, 20]

Recent data from the ICON8 study, which compared the efficacy and safety of two dose-dense weekly regimens to standard 3-weekly chemotherapy in a mostly UK population of newlydiagnosed ovarian cancer patients (1,397 of 1,566; 89%) showed that a proportion of women achieved a sustained response (and possibly long-term remission), remaining progression-free even after five years of completing surgery and chemotherapy (Figure 8).[36] This group of patients are referred to as “long-term survivors” in the cost-effectiveness section (Section 3.3.2).

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 23 of 177

Figure 8. Progression-free survival in the overall ICON8 study population

==> picture [458 x 183] intentionally omitted <==

Note: >80% of patients in ICON8 had advanced FIGO Stage III or IV disease; 47% of patients had received immediate [upfront] debulking surgery and 50% NACT followed by delayed surgery [3% had not undergone any surgery. Source: Clamp et al., 2019.[36 ]

==> picture [453 x 49] intentionally omitted <==

----- Start of picture text -----
2. Platinum-based
3. Bevacizumab
1. Surgery ± NACT chemotherapy ±
maintenance*
bevacizumab*
----- End of picture text -----

*For eligible patients. NACT: neoadjuvant chemotherapy

Bevacizumab is available for use through the Cancer Drugs Fund (CDF) in combination with chemotherapy and then alone as maintenance therapy

In 2013, bevacizumab, in combination with carboplatin and paclitaxel, was made available for use through the CDF for the first-line treatment of advanced ovarian cancer patients who had:[33, 38]

• FIGO Stage III disease at presentation and required NACT due to low likelihood of optimal primary surgical cytoreduction, OR

• FIGO Stage III ovarian cancer, with residual disease of >1cm following debulking surgery, OR

• FIGO Stage IV disease.

CDF criteria require that bevacizumab treatment is initiated with the first or second cycle of chemotherapy and continued as maintenance therapy at a dose of 7.5 mg/kg every three weeks, for a maximum of 18 cycles in total (as illustrated in Figure 9 below).[33, 38]

In the pivotal Phase III GOG-0218 and ICON7 studies, bevacizumab (15mg/kg or 7.5mg/kg) + chemotherapy followed by bevacizumab maintenance reduced the risk or disease progression or death by 20%−30% in the overall study population (versus chemotherapy followed by routine surveillance).[34, 35] A PFS benefit in favour of bevacizumab maintenance (versus routine surveillance) in women who have responded to their first-line regimen is also supported by additional indirect treatment comparisons conducted by AstraZeneca on available data for women with BRCA m advanced ovarian cancer [as per the SOLO-1 trial], or HRD-positive disease [as per the PRIMA trial]; discussed in Section B.2.12). The use of bevacizumab (7.5mg/kg) in combination with chemotherapy, followed by bevacizumab 7.5mg/kg maintenance therapy, also conferred an OS benefit (HR=0.78, versus chemotherapy followed by routine surveillance) amongst women who

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 24 of 177

had inoperable Stage III ovarian cancer with residual disease of >1cm after PDS, or inoperable Stage III disease, or Stage IV ovarian cancer.[39, 40]

The addition of bevacizumab to chemotherapy, followed by bevacizumab maintenance, thus offers the opportunity to further improve outcomes in advanced ovarian cancer patients, and is expected to increase the proportion of women who achieve long-term remission (relative to chemotherapy followed by routine surveillance; Figure 8). The results of the ongoing ICON8B study, which has 67 UK centres, are expected to provide further evidence on this topic.[36]

Figure 9: CDF criteria to receive bevacizumab

==> picture [427 x 243] intentionally omitted <==

Abbreviations : CDF: Cancer Drugs Fund; IDS: interval debulking surgery; Q3W: once every three weeks; PDS: primary debulking surgery. Source: Adapted from CDF BlueTeq Criteria and National Health Service England Cancer Drugs Fund List.[33, 38]

It is estimated that 80% of the overall population of women with advanced ovarian cancer would currently be eligible to receive bevacizumab treatment, based on their disease stage and surgical status (assuming other eligibility criteria are met[38] ) (Appendix N; Figure 10).

In this context, it is also important to note that multiple bevacizumab biosimilar launches are anticipated in the UK whilst this appraisal is ongoing, following on from the loss-of-exclusivity (LoE) of Avastin[®] . Based on historical precedence, we expect significant reduction in the current price of bevacizumab as a result of these events, in turn facilitating its use in routine commissioning at a dose and population aligned to its EMA marketing authorisation .[5]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 25 of 177

Figure 10: Proportions of women eligible to receive bevacizumab for advanced ovarian cancer via the CDF

==> picture [314 x 177] intentionally omitted <==

Briefly, proportions of women diagnosed with Stage III versus Stage IV ovarian cancer were estimated from the Ovarian Cancer Audit Feasibility Pilot; proportions of PDS versus IDS from ICON8; and proportion of PDS with residual or no residual disease from ICON7. Abbreviations : CDF: cancer drugs fund; IDS: interval debulking surgery; NACT: neoadjuvant chemotherapy; PDS: primary debulking surgery Source : National Cancer Registration and Analysis Service, 2018[41] ; Perren et al., 2011[35] ; Clamp et al., 2019[36] .

In August 2019, NICE recommended olaparib monotherapy (through the CDF) as maintenance treatment for women with advanced ovarian cancer, who had responded to first-line platinumbased chemotherapy and who had deleterious mutations in BRCA1 or BRCA2 genes.[42] Although maintenance olaparib monotherapy is not a formal comparator in this appraisal , the impact of this recommendation is briefly summarised in Figure 11 below for completeness

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 26 of 177

Figure 11. The CDF recommendation for olaparib was described as a “new era” for the treatment of women with newly-diagnosed BRCA -mutated advanced ovarian cancer

In August 2019, NICE recommended olaparib monotherapy as an option for the maintenance treatment for women with advanced ovarian, fallopian tube, or peritoneal cancer, who had responded to first-line platinum-based chemotherapy and who had deleterious mutations in BRCA1 or BRCA2 genes.[42] This recommendation was based on the results of the pivotal Phase III SOLO-1 study, which showed that maintenance treatment with olaparib monotherapy reduced the risk of disease progression or death by a remarkable 70% versus placebo (or routine surveillance).[43]

Kaplan-Meier plot of investigator-assessed PFS in the SOLO-1 study[43]

==> picture [399 x 205] intentionally omitted <==

A benefit of this magnitude had never been achieved previously with a systemic therapy for advanced ovarian cancer. The CDF recommendation for olaparib was described by UK oncologists as being “ the most exciting change in primary management of advanced ovarian cancer in the last 20 years” , marking a “ new era ” for the treatment of 20%–25% of patients, who have mutations in BRCA1/2 genes.[9, 10, 27, 42, 44-46]

Now, tThe PAOLA-1 regimen offers a further improvement in PFS for these women (see Section B.2.12) , in addition to providing a broader group of patients with HRD-positive (including BRCA wt) disease the opportunity to achieve a similar level of benefit:

• PFS HR = 0.33 for HRD-positive patients, including those with BRCA m tumours, versus • PFS HR = 0.43 for HRD-positive patients excluding those with BRCA m tumours (Appendix E).

Abbreviations : BRCA1/ 2: breast cancer susceptibility gene; BRCAm ; BRCA mutated; BRCAwt : BRCA wild-type; CDF: Cancer Drugs Fund; CI: confidence interval; NICE: National Institute for Health and Care Excellence; PFS: progression-free survival.

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 27 of 177

Unmet need:

• Although advances in surgery / surgical radicality, and the addition of bevacizumab (in combination with chemotherapy and as maintenance treatment), have improved long-term PFS in newly-diagnosed advanced ovarian cancer, most women still experience relapse or disease progression after first-line therapy .[20, 47-49]

  • Relapsed ovarian cancer is not only associated with a greater symptom burden / HRQoL impact, but also negatively impacts on emotional wellbeing. Ovarian cancer patients have repeatedly highlighted the devastating nature of relapsed disease in previous NICE appraisals, emphasising that “ any extension to life is incredibly precious ”.[50]

  • Response rates and progression-free intervals shrink with each subsequent round of chemotherapy for relapsed disease, until eventually, the tumour becomes resistant to platinum-based therapy.[51, 52]

  • Conversely, the risks of developing cumulative toxicities (such as, neurotoxicity, alopecia, and ototoxicity) increase, adding to the overall burden of disease.[53, 54]

• For women who relapse with platinum-sensitive disease (Figure 12) and respond to secondline chemotherapy, the standard-of-care is maintenance treatment with one of three recommended PARPi (TA528, TA611, TA620) .[29, 50, 55, 56]

• Although PARPi have greatly improved outcomes for relapsed ovarian cancer (with a small proportion of women experiencing long-term remission even in this advanced setting)[57] :

  • The magnitude of median PFS benefit achieved is much lower compared to that achieved by PARPi in the first-line maintenance setting.[1, 43, 58-60]

  • Women who relapse with platinum-resistant disease or do not respond to their most recent round of chemotherapy are unable to access PARPi for their relapsed disease (Figure 14)[61-63]

• Collectively, these aspects emphasise the importance of effective first-line maintenance therapy, to prevent or delay disease progression, further rounds of cytotoxic chemotherapy, and worsening HRQoL and prognosis for patients.

  • Using PARPi earlier in the treatment pathway – in the first-line maintenance setting - would allow more women to receive and derive maximum benefit from these innovative therapies.

The majority of women relapse, despite initial response to first-line platinum-based chemotherapy (± bevacizumab)

Although most women (~80%) respond to first-line chemotherapy with carboplatin-paclitaxel (with more than half achieving complete remission [i.e. no evidence of disease or complete response(CR) after surgery and chemotherapy), the majority experience relapse or disease progression (Figure 14)[20, 48, 49] The timing of relapse (and length of the progression-free interval) has important implications for both prognosis and response to second-line therapy, and is broadly classified into four categories: platinum-refractory, platinum-resistant, partially (or intermediately) platinum-sensitive, and (highly) platinum-sensitive, as described in Figure 12. [64]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 28 of 177

==> picture [443 x 11] intentionally omitted <==

----- Start of picture text -----
Figure 12: Gynaecologic Cancer Intergroup (GCIG) responses to platinum chemotherapy
----- End of picture text -----

==> picture [455 x 124] intentionally omitted <==

----- Start of picture text -----
Treatment free interval
Primary
treatment 0 months 6 months 12 months
Platinum Platinum Partially Platinum
refractory resistant platinum sensitive
sensitive
----- End of picture text -----

Note: per definitions confirmed by the GCIG 4th Ovarian Cancer Consensus Meeting, ‘platinum-refractory’ refers to those patients progressing during therapy or within 4 weeks after the last dose; ‘platinum-resistant’ to patients progressing within 6 months of platinum-based therapy; ‘partially platinum-sensitive’ to patients progressing between 6 and 12 months; and ‘platinum-sensitive’ to patients progressing with an interval of more than 12 months (GCIG Consensus) [49].[65] Although these definitions are now outdated, they were used to define patient populations in most clinical trials of relapsed ovarian cancer and are thus worth noting. Abbreviations: GCIG: Gynaecologic Cancer Intergroup. Source: Adapted from Ushijima, 2010.[64]

Further treatment options are limited for women with platinum-refractory or -resistant disease and are focused on HRQoL and symptom palliation.[53, 54, 66-69] Life expectancy for this group of women is less than 12 months .[70] BGCS guidelines recommend single-agent chemotherapy with non-platinum agents (such as paclitaxel and PLDH) for these patients, since they are associated with fewer adverse events (AEs) and similar efficacy, relative to combination therapies.[18]

Many women are unable to receive PARPi in the relapsed setting due to platinumresistance / lack of response to chemotherapy

For women with platinum-sensitive disease, ESMO guidelines recommend carboplatin-doublets as the treatment of choice.[20, 69] Paclitaxel and PLDH are both recommended by NICE (in combination with platinum, or as monotherapy) in this setting (TA389)[*] .[71]

Response to chemotherapy and progression-free intervals shrink with each subsequent round of treatment (until eventually, the tumour becomes platinum-resistant), whilst the risks of developing cumulative toxicities (such as, neurotoxicity, alopecia and ototoxicity) increase.[52-54]

• For women who do not respond to second-line platinum-based chemotherapy, treatment options are limited to those described above for platinum-refractory or -resistant disease.

• For those who do respond to platinum-based chemotherapy, the standard-of-care is PARPi maintenance therapy.[29] NICE has recommended three different PARPi for women with relapsed, platinum-sensitive ovarian cancer , namely: olaparib (in those women who have germline or somatic BRCA1 or BRCA2 mutations, TA620), niraparib (TA528), and rucaparib (TA611).[50, 55, 56]

• Gemcitabine in combination with carboplatin, trabectedin in combination with PLDH, or topotecan monotherapy are not recommended (TA389).

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 29 of 177

Although the use of PARPi has transformed patient outcomes in platinum-sensitive relapsed ovarian cancer (with a proportion of women remaining alive and progression-free even after 7 years of completing their chemotherapy),[57] the median duration of PFS achieved is much lower than in the first-line maintenance setting:[58-60]

Figure 13: Progression-free survival outcomes for women with newly diagnosed and platinum sensitive relapsed ovarian cancer

*, Median PFS was similar in other PARPi studies in this setting: 21.0 months in the g BRCA m cohort of ENGOTOV16/NOVA (niraparib), and 16.6 months in the s BRCA m or g BRCA m cohort of ARIEL3 (rucaparib).[59][60] Abbreviations : BRCAm: breast cancer susceptibility gene mutation ; gBRCAm: germline BRCAm; OC: ovarian cancer; PARP: poly-ADP ribose polymerase; PFS: progression free survival. References: Moore et al., 2018[43] ; Pujade-Lauraine et al., 2017[58]

Additionally, women who relapse with platinum-resistant disease or do not respond to their most recent round of chemotherapy are ineligible to receive PARPi therapy in the relapsed setting (see Figure 14). This further emphasises the importance of using PARPi earlier in the treatment pathway - in the first-line maintenance setting – where the likelihood of response to chemotherapy (and therefore eligibility for PARPi maintenance therapy) is highest and magnitude of benefit achieved is the greatest.

Figure 14: Not all women are able to benefit from PARPi therapy upon relapse

==> picture [448 x 252] intentionally omitted <==

Abbreviations: HRD: homologous recombination deficiency; OC: ovarian cancer; PARP: poly-ADP ribose polymerase; PARPi; PARP inhibitor Source : Adapted from Banerjee et al., 2019[61] ; Bruchim et al., 2013[62] ; Aghajanian et al., 2012[63]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 30 of 177

Women with relapsed advanced ovarian cancer face growing physical and emotional burden and worsening HRQoL

Women with recurrent ovarian cancer experience a greater symptom burden (both in terms of the number and severity of symptoms) and worse HRQoL, compared to women with newly diagnosed disease.[72] Many women are too unwell at the point of relapse to undergo further active anticancer therapy.[20]

A 2017 Italian multicentre study in 173 women with ovarian cancer, involving 50 oncologists, reported substantial differences in self-assessed health status between women who had relapsed disease versus those who did not[*] .[72]

• Only 33.6% of women with disease recurrence reported their health as being “ good ” or “ excellent ”, versus 82.4% of women without recurrence (P<0.05).[72]

This was consistent with physician-referred Eastern Cooperative Oncology Group Performance Status (ECOG PS) scores - 91.1% of patients without recurrence had a score of 0 or 1, versus 50.9% of those with recurrent disease (p<0.05).[72]

• Most women with recurrence reported that pain affects their daily activities (71.8%, versus 21% of women with no recurrence).[72]

• Significant differences were also noted in emotional state and wellbeing , with more women with recurrent disease reporting feeling sad or discouraged.[72]

Whereas women without disease recurrence more generally felt that the “ future still [held] many opportunities ”, those with recurrence felt that “ time [was] running out ” and that “ opportunities for the future [were] limited ”.[72]

The negative outlook reported in this study has been echoed by ovarian cancer patients in England, who, in past NICE appraisals of treatments for relapsed ovarian cancer, have highlighted the devastating nature of disease, emphasising that “ any extension to life is incredibly precious ”.[50]

Collectively, these data and insights highlight the impact of disease recurrence on women living with advanced ovarian cancer and underscore the importance of preventing disease progression after first-line therapy, when the chances of achieving long-term remission (or even a cure) are at their highest .

The PAOLA-1 regimen aims to address this unmet need through the addition of olaparib to standard-of-care bevacizumab maintenance therapy. The rationale for the PAOLA-1 regimen and the body of evidence supporting its positioning as a “ new standard of care ” in this setting are further described in the following sections.

*The study defined absence of recurrence as patients that had no detectable symptoms of relapsed disease (clinically or via imaging) following one or more lines of chemotherapy for a minimum of three years following their last cycle of chemotherapy. The study did not consider elevated CA-125 levels. Recurrence was defined as clinical or radiological evidence of disease within six months of the last line of chemotherapy.

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 31 of 177

B.1.3.3 Clinical rationale and proposed positioning of the PAOLA-1 regimen

• Prior evidence from multiple clinical trials highlight a role for both bevacizumab in combination with chemotherapy, followed by bevacizumab maintenance,[34-36] and for olaparib as maintenance monotherapy,[43] for the treatment of patients with newly-diagnosed advanced ovarian cancer.

• The PAOLA-1 regimen allows patients to benefit from both these maintenance treatments after response to first-line chemotherapy, when the volume of disease is at its lowest and the potential magnitude of benefit is highest.

  • By adding olaparib to bevacizumab maintenance treatment, the PAOLA-1 regimen aims to provide patients the maximum benefit achievable in this treating setting.

• Data from the PAOLA-1 study highlight remarkable efficacy for olaparib added to bevacizumab maintenance treatment in the HRD-positive group – the focus of this submission, with a 67% reduction in the risk of disease progression or death, versus an active comparator arm of bevacizumab and placebo.[ 73 ]

  • The PFS data are supported by series of clinically-relevant secondary endpoints, including TFST, PFS2, TSST, and OS (Table 7), which show consistent benefit in favour of olaparib added to bevacizumab (versus placebo + bevacizumab).[74]

• Additional unanchored population-adjusted indirect comparisons show that the PAOLA-1 regimen also provides a meaningful improvement in PFS versus PARPi monotherapy (olaparib [in BRCAm patients] and niraparib [in HRD-positive patients*]).

Collectively, these data support a role for the PAOLA-1 regimen as a “ new standard-of-care” for women with newly-diagnosed HRD-positive advanced ovarian cancer, who are in complete or partial response after first-line chemotherapy with bevacizumab.

*Analysis conducted in women with advanced ovarian cancer who had Stage III disease with visible residual tumour after PDS, inoperable Stage III disease, any Stage IV disease, or had received neoadjuvant chemotherapy (aligned to PRIMA; subset of PAOLA-1 HRD-positive group)

Bevacizumab in ovarian cancer

Ovarian cancers are highly vascularised tumours.[75] Bevacizumab is a humanised monoclonal antibody that targets angiogenesis (i.e. the formation of new blood vessels) in tumours, through inhibiting the pro-angiogenic mediator vascular endothelial growth factor A (VEGF-A).[75-77] The binding and inactivation of VEGF-A by bevacizumab inhibits endothelial cell activation and proliferation, thus preventing tumour growth and metastasis (illustrated in Figure 15).[76]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 32 of 177

Figure 15: The mechanism of action for bevacizumab

==> picture [414 x 260] intentionally omitted <==

Footnotes: A) Process of tumour angiogenesis involving VEGF; B−C) Bevacizumab mechanism of action targeting VEGF.

Abbreviations: VEGF: vascular endothelial growth factor. Source: Adapted from Mukherji et al. (2010).[76]

As highlighted previously (B.1.3.2), the addition of bevacizumab to first-line chemotherapy, followed by maintenance bevacizumab monotherapy, extended PFS by 20%−30% in women with newly-diagnosed advanced ovarian cancer in the pivotal GOG-0218 and ICON7 studies .[34, 35] The ICON7 study also showed an overall survival benefit for bevacizumab in combination with chemotherapy and then as maintenance treatment (versus chemotherapy followed by routine surveillance) amongst women with Stage III ovarian cancer who had residual disease of >1cm after debulking surgery, or inoperable Stage III disease, or Stage IV ovarian cancer.[39]

The EMA marketing authorisation for bevacizumab in the first-line advanced ovarian cancer setting was based on the registrational GOG-0218 trial, which investigated bevacizumab at a dose of 15 mg/kg Q3W, for up to 15 months.[5, 34] However, the CDF recommendation for bevacizumab is aligned to the 7.5 mg/kg for 12 months regimen used in the ICON7 study.[33, 35] Naïve comparisons in similar populations of women from GOG-0218 and ICON7 studies reveal no meaningful differences in PFS or OS achieved with the two bevacizumab doses (Figure 16).[78] This was also confirmed in a meta-analysis investigating the efficacy of bevacizumab + standard chemotherapy stratified by dose, which showed no statistically significant differences (as shown in Figure 16).[78]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 33 of 177

Figure 16: Comparison of A) PFS and B) OS between 7.5 mg/kg Q3W and 15 mg/kg Q3W bevacizumab doses

==> picture [490 x 327] intentionally omitted <==

Note: This analysis investigated the efficacy and safety of bevacizumab stratified by dose in patients with relative high risk for progression (FIGO III, macroscopic >1 cm and IV). However, since the GOG-0218 study only provided PFS and OS curves for all patients with FIGO III−IV disease, patients with FIGO III, and macroscopic disease of ≤ 1 cm could not be separated. A) PFS: in the control arms, median PFS was 11.3 and 11.5 months in ICON7 and GOG-0218 studies (HR, 1.14; CI, 0.96 to 1.34). For bevacizumab + standard chemotherapy arms, median PFS was 16.5 months for the 7.5 mg/kg dose and 15.6 months for the 15 mg/kg dose (HR, 1.04; CI, 0.88 to 1.24). B) OS: Even though difference existed between the two control arms (HR, 1.60; CI, 1.24 to 2.06), no significant difference was shown between the two doses in the bevacizumab + standard chemotherapy arms (HR, 1.15; CI, 0.88 to 1.50).

Abbreviations: CI: confidence interval; FIGO: International Federation of Gynaecology and Obstetrics; HR: hazard ratio; OS: overall survival; PFS: progression-free survival; Q3W, once every three weeks. Source : Zhou et al., 2013[78]

Company evidence submission template for olaparib + bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652]

© AstraZeneca (2020). All rights reserved

Olaparib in ovarian cancer

As described in Section B.1.3.1, approximately half of all HGSOCs have HR deficiency,[10, 25] which renders them amenable to PARPi therapy:

• PARPs are proteins that play an important role in the repair of DNA single-strand breaks (SSBs; described further in Appendix N.4). PARPi, such as olaparib, bind to PARP, trapping it on DNA SSBs (Figure 17).[79] This prevents the ensuing steps of the repair pathway, leading to the persistence of SSBs, and subsequently their conversion to more harmful DSBs during DNA replication.[80]

• Normal cells accurately repair and survive DNA DSBs arising as a consequence of PARP inhibition through the homologous recombination pathway.[80] However, cells that are deficient in homologous recombination utilise the error-prone non-homologous end joining process to repair these breaks.[80] This leads to the accumulation of genomic instability and ultimately, cell death. This phenomenon, whereby the independent loss of two factors permits cell survival, but loss of both factors in combination results in cell death is referred to as “ synthetic lethality ” and underpins the effectiveness of PARPi (Figure 17).[81] The targeted mechanism of action of PARPi limits their toxicity, and favours sustained use (such as in a maintenance setting).[82]

Figure 17: Pathways for DNA repair following treatment with PARPi

==> picture [419 x 354] intentionally omitted <==

Abbreviations: DNA: deoxyribose nucleic acid; DSB: double-strand break; HRD: homologous recombination deficiency; PARP(i): poly-ADP ribose polymerase (inhibitor); RF: replication fork; SSB: single-strand break. Source : Adapted from Dziadkowiec et al., 2016.[79]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

As described in Figure 11, maintenance treatment with olaparib in women with newly-diagnosed, BRCA1/2 -mutation-positive advanced ovarian cancer who were in response following first-line chemotherapy led to an unprecedented 70% reduction in the risk of disease progression or death, relative to placebo (or routine surveillance).[43]

Olaparib has also demonstrated a statistically significant and clinically meaningful PFS and OS benefit (versus placebo or routine surveillance) in platinum-sensitive relapsed ovarian cancer, and is recommended by NICE in this treatment setting.[56-58] Long-term data from the Phase II Study 19, which investigated the efficacy and tolerability of the olaparib capsule formulation versus placebo, showed that the PFS benefit of olaparib translated into long-term improvements in time to first and second subsequent therapy (TFST and TSST, respectively) and ultimately, into extended OS versus placebo (median follow-up = 78 months; ).[57, 83] These data provide important insights into the magnitude of long-term clinical benefit that can be achieved with olaparib maintenance therapy. It is also worth noting that olaparib is the only PARPi for which such longterm data are available.

Figure 18. Long-term follow-up data show that ~20% of women do not require subsequent anticancer therapy following maintenance treatment with olaparib for relapse, platinumsensitive advanced ovarian cancer*

==> picture [357 x 208] intentionally omitted <==

*Unselected by BRCA mutation or HRD status.

Note: Data from Study 19, a randomised, placebo-controlled, Phase II trial that enrolled 265 patients who had received at least two platinum-based chemotherapy regimens and were in complete or partial response to their most recent regimen.

Abbreviations : bid: twice daily; CI: confidence interval; HR: hazard ratio; TFST: time to first subsequent therapy. Source: Friedlander et al., 2018 (Supplementary Material).[83]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 19. Study 19 data show that extensions in TFST and TSST translate into a significant long-term overall survival benefit in favour olaparib (versus placebo)*

==> picture [387 x 221] intentionally omitted <==

*Unselected by BRCA mutation or HRD status.

Note: Data from Study 19, a randomised, placebo-controlled, Phase II trial that enrolled 265 patients who had received at least two platinum-based chemotherapy regimens and were in complete or partial response to their most recent regimen.

Abbreviations : bid: twice daily; BRCA : breast cancer susceptibility gene; CI: confidence interval; HR: hazard ratio; HRD: homologous recombination deficiency; OS: overall survival; TFST: time to first subsequent therapy; TSST: time to second subsequent therapy. Source: Friedlander et al., 2018.[83]

Rationale for the PAOLA-1 regimen (i.e. addition of olaparib to bevacizumab maintenance)

Pre-clinical data have suggested a potential synergistic benefit of combining PARP and VEGF inhibitors (see Appendix N for further detail). This hypothesis is supported by clinical data from two Phase II RCTs:

• The first investigated olaparib in combination with cediranib (a small molecular VEGF receptor and c-kit tyrosine kinase inhibitor) versus olaparib monotherapy , in women with platinum-sensitive relapsed ovarian cancer.[84, 85]

• The second evaluated the PARPi niraparib + bevacizumab , versus niraparib alone, in women with platinum-sensitive relapsed ovarian cancer.[86]

In both studies, the combination of a PARPi (i.e. olaparib or niraparib) and an anti-angiogenic agent (bevacizumab or cediranib) significantly extended PFS, relative to PARPi monotherapy in a biomarker unselected population (including both BRCA1/2 -mutated and wild-type tumour types).[84-86] Importantly, neither study included an anti-VEGF monotherapy arm, a limitation that was addressed by the PAOLA-1 study.[1]

PAOLA-1 is the first and only Phase III clinical trial that investigated the efficacy and tolerability of adding olaparib to bevacizumab therapy, an established standard-of-care for maintenance treatment of newly-diagnosed advanced ovarian cancer.[1] In doing so, the study aimed to maximise the clinical benefit that could be achieved in this setting, at a time when the disease burden is at its lowest (i.e. following surgery and/or response to chemotherapy) and the potential to achieve long-term remission or even cure is at its highest.[87, 88]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Bevacizumab was selected for combination with olaparib in PAOLA-1 since it was an established standard-of-care for the first-line and maintenance treatment of advanced ovarian cancer in Europe and Japan when the study was initiated. A biomarker unselected population was chosen for inclusion in PAOLA-1 based on the observation of clinical benefit regardless of BRCA1/2 mutation status in the platinum-sensitive relapsed setting.[84-86]

Data from the PAOLA-1 study, in conjunction with further indirect treatment comparisons, support the position of the PAOLA-1 regimen as a “new standard-of-care” for HRDpositive patients who have responded to first-line chemotherapy

2. Platinum-based 3. Olaparib, added 1. Surgery ± NACT chemotherapy ± to bevacizumab bevacizumab maintenance*

*Proposed position of the PAOLA-1 regimen. Abbreviations : HRD: homologous recombination deficiency; NACT: neoadjuvant chemotherapy.

The addition of olaparib to bevacizumab maintenance treatment demonstrated a remarkable PFS benefit in HRD-positive population, with a 67% reduction in the risk of disease progression or death versus an active control arm of bevacizumab given with placebo, with a median duration of PFS of >3 years (versus 17.7 months with placebo + bevacizumab).[1] The PFS data were supported by a series of clinically-relevant secondary endpoints, including TFST, PFS2, and TSS - all of these showed meaningful improvements in favour of olaparib added to bevacizumab, and provide confidence in the OS result, which (albeit immature) indicates a xxx reduction in the overall risk of death versus bevacizumab with placebo.[1]

Following on from the publication of PAOLA-1 results and feedback from clinical community, AstraZeneca have conducted a series of population-adjusted indirect comparisons to show the incremental benefit of the PAOLA-1 regimen versus placebo and olaparib maintenance monotherapy and contextualise its positioning in the clinical pathway of care. The results of these analyses are briefly described below (and discussed further in Section B.2.12).

It is important to note that these indirect comparisons involve treatments that are outside the scope of the decision problem for this appraisal; these data are shown purely to aid the interpretation of PAOLA-1 results and to allow an understanding and appreciation of the incremental benefit of the PAOLA-1 regimen versus other treatment options.

The incremental benefit of olaparib added to bevacizumab versus olaparib monotherapy using data from SOLO1 and PAOLA-1 ( BRCAm tumours)

In the absence of a common control arm across studies (placebo in SOLO-1 and placebo plus bevacizumab in PAOLA-1), the relative efficacy of the olaparib plus bevacizumab and olaparib monotherapy arms of PAOLA-1 and SOLO-1 was evaluated using an unanchored populationadjusted indirect comparison method, as described in DSU TSD18. The results of this analysis showed a meaningful PFS benefit in favour of olaparib + bevacizumab, versus olaparib monotherapy (HR=0.71; 95% confidence interval (CI): 0.45, 1.09) in patients with BRCAm advanced ovarian cancer (Figure 20).

In addition, this analysis showed:

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

• A statistically significant PFS benefit for olaparib monotherapy versus bevacizumab (+ placebo) (HR=0.48; 95% CI: 0.30, 0.75), and

• A statistically significant PFS benefit for bevacizumab (+ placebo) versus placebo (HR=0.65; 95% CI: 0.43, 0.95)

Figure 20. Population adjusted analysis: PFS outcomes for the weighted BRCA -mutated subset of PAOLA-1 and unweighted SOLO-1

==> picture [422 x 56] intentionally omitted <==

==> picture [374 x 231] intentionally omitted <==

Abbreviations : Bev: bevacizumab; BRCA: breast cancer susceptibility gene; PFS: progression-free survival. Shaded area indicates 95% CI.

A key limitation of this unanchored population-adjusted indirect comparison is that it is limited to women who have BRCAm advanced ovarian cancer. There are no studies that allow indirect comparisons between olaparib + bevacizumab, versus olaparib or bevacizumab maintenance monotherapy, or placebo, in the HRD-positive population that is relevant to this appraisal (see appendix D).

However, using data from the PRIMA study, which investigated the efficacy and tolerability of maintenance therapy with another PARPi, niraparib, may provide an indication of relative benefit[†] . These data are summarised below.

The incremental benefit of olaparib added to bevacizumab, versus PARPi (niraparib) or bevacizumab maintenance monotherapy using data from PAOLA-1 and PRIMA

† Note: Unlike PAOLA-1, which did not restrict inclusion by prior surgery / surgical outcomes, the PRIMA study only included those Stage III patients who had received NACT and IDS, or had visible residual tumour after PDS, or inoperable disease, in addition to patients with Stage IV disease. See Section B.2.12 for further detail on this analysis.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 39 of 177

(HRD-positive tumours; inoperable or sub-optimally debulked Stage III or Stage IV disease only)

As with the comparison of PAOLA-1 to SOLO1, an unanchored population-adjusted indirect treatment comparison was performed between the active and control arms of PRIMA (niraparib versus placebo) and PAOLA-1 (olaparib plus bevacizumab versus placebo plus bevacizumab). The results of this analysis show that the addition of olaparib to bevacizumab significantly improved PFS versus:

• Niraparib (HR=0.57; 95% CI: 0.41, 0.80),

• Bevacizumab (+ placebo) (HR=0.40; 95% CI: 0.28, 0.57), and

• Placebo (HR=0.23; 95% CI: 0.16, 0.33),

in patients with HRD-positive tumours, who had - Stage III disease with visible residual tumour after PDS, inoperable Stage III disease, any Stage IV disease, and those who had received neoadjuvant chemotherapy (Figure 21; this analysis is described further in Section B.2.12.).

Figure 21. Pooled matching analysis of PRIMA and PAOLA-1 (HRD-positive patients)

==> picture [429 x 60] intentionally omitted <==

x

Shaded area indicates 95% CI. Abbreviations : HRD, homologous recombination deficiency.

Although subject to the limitations of non-randomised comparisons (and lack of access to individual patient data from PRIMA), these data support the proposed positioning of olaparib added to bevacizumab maintenance treatment as a “ new standard-of-care ” for women with HRD-positive advanced ovarian cancer, who have responded to first-line chemotherapy.

In this respect, it is worth noting that the additional budget impact of introducing the PAOLA-1 regimen in clinical practice will be minimised from the following factors:

• Significant reduction to the current bevacizumab price following loss of exclusivity of Avastin[®] and multiple biosimilar launches will:

  • Reduce the costs incurred by the NHS due to more patients receiving bevacizumab 15mg/kg in combination with chemotherapy followed by bevacizumab 15mg/kg maintenance, as part of the PAOLA-1 regimen. This assumes that PAOLA-1 data will be practice-changing, and more women will receive bevacizumab (than in current clinical practice) due to the remarkable efficacy of olaparib + bevacizumab maintenance treatment.

• The costs associated with the increased use of olaparib in the first-line maintenance setting will be at least partially off-set by lower use of PARPi (olaparib, niraparib, and rucaparib) in second- and greater lines of treatment as maintenance therapy for platinum-sensitive relapsed disease.

• The use of PARPi in multiple lines of treatment for the same patient is not permitted under existing NICE guidance.[33]

• Finally, ~50% of HRD-positive patients, who have BRCA m disease, are already receiving maintenance therapy with olaparib in current NHS practice (TA598).[42]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.1.4 Equality considerations

It is not considered that the introduction of olaparib is likely to lead to recommendations which differentially impact any patients protected by equality legislation or disabled persons.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2 Clinical effectiveness

B.2.1 Identification and selection of relevant studies

A systematic literature review (SLR) was conducted to identify studies relevant to this submission. The SLR search strategy, study selection criteria, and results are provided in Appendix D.

A broad SLR search was conducted capture any published clinical trial evidence on first-line and maintenance treatments for newly-diagnosed advanced ovarian cancer patients. This approach was selected to ensure no relevant studies were accidentally missed; however, in preparing for this submission, we had filtered search results to focus specifically on the maintenance setting that is relevant to the PAOLA-1 regimen.

Following discussions with NICE and the Evidence Review Group (ERG) during the checkpoint meeting on 24 February 2020, the inclusion criteria applied to the searches were broadened to cover the full treatment sequence captured in the intervention and comparator statements of the decision problem (Table 1). The study selection process was repeated to capture any of the following regimens:

• Intervention: Platinum-based chemotherapy with bevacizumab (15 mg/kg every 3 weeks) followed with olaparib and bevacizumab (15mg/kg Q3W) maintenance therapy only in responding patients

• Comparators:

  • Platinum based chemotherapy followed with routine surveillance.

  • Platinum-based chemotherapy with bevacizumab (7.5 mg/kg Q3W) followed with bevacizumab (7.5mg/kg Q3W) maintenance therapy.

In addition, studies that randomised patients after response to first-line chemotherapy were included if they contained any of the maintenance regimens specified in the intervention or comparator statements, i.e. olaparib and bevacizumab (15mg/kg Q3W), bevacizumab monotherapy (15mg/kg Q3W or 7.5mg/kg), or routine surveillance (i.e. placebo).

A total of 74 publications, reporting on 51 clinical trials, were identified using this strategy and highlight the breadth of the scope specified by NICE . All of these studies were evaluated for feasibility of inclusion in a network meta-analysis / indirect treatment comparison. The results of this feasibility analysis are described in Section B.2.9.

Due to fundamental differences in the patient population included in PAOLA-1 versus the other identified studies, indirect treatment comparisons linking the intervention and comparators, as specified in the decision-problem, were not deemed possible (see Section B.2.9. and Appendix D for further details). The clinical effectiveness evidence presented in this Section therefore focuses solely on findings from the PAOLA-1 study - the only RCT evaluated the intervention of interest in the NICE scope and the pivotal study for olaparib in this indication (i.e.

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx).

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.2 List of relevant clinical effectiveness evidence

A brief overview of the PAOLA-1 study, where the clinical effectiveness evidence presented in this submission are derived from, is presented in Table 3. Further details are provided in Section B.2.3.−B.2.10.

Table 3: PAOLA-1 study design

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

• TFST (outcomes in bold have • TSST been incorporated into • TDT the HE model’s base• Adverse effects of treatment case results) • HRQoL (EORTC QLQ-C30, EORTC QLQ-OV28, and EQ-5D ) Note: Investigator-assessed PFS (primary endpoint) data are reported for the FAS and the HRD-positive group; data on key secondary efficacy endpoints and PROs are presented for the HRD-positive group only. Safety summaries are presented for the SAS and HRD-positive group.

*, As per the 1988 FIGO classification. Using the 2014 FIGO classification for Stage III disease, women in PAOLA1 would be classified as having Stage IIIA–IV ovarian cancer

†, The study protocol required ≥3 cycles of bevacizumab to be administered in combination with chemotherapy; maximum duration of bevacizumab = 15 months in total. For clarity, patients enrolled into the PAOLA-1 study were randomised to olaparib + bevacizumab or placebo + bevacizumab groups.

§, Patients with evidence of disease at two years, who in the opinion of the treating physician can derive further benefit from continuous olaparib treatment, can be treated beyond two years. In PAOLA-1, most patients came offtreatment at the first scheduled follow-up visit after two years (week 108 or month 25). Just 5 patients in the olaparib + bevacizumab arm remained on treatment by month 26; by month 30, just 2 patients remained on treatment. Abbreviations: BID: twice daily; ENGOT: European Network for Gynaecological Oncological Trial; EORTC: European Organisation for the Research and Treatment of Cancer; EQ-5D: EuroQoL five dimensions; FIGO: International Federation of Gynaecology and Obstetrics; GCIG: Gynaecologic Cancer InterGroup; HE: health economic; HRD: homologous recombination deficient; HRQoL: health-related quality of life; OS: overall survival; PFS: progression-free survival; PFS2: second progression-free survival; Q3W, once every three weeks; QLQ-C30: Quality of Life Questionnaire for Cancer Patients (Core 30 item module); QLQ-OV28: Quality of life questionnaire for ovarian cancer patients; TDT: time to treatment discontinuation or death; TFST: time to first subsequent therapy or death; TSST: time to second subsequent therapy or death. Source : PAOLA-1 CSR.[73]

B.2.3 Summary of methodology of the relevant clinical

effectiveness evidence

B.2.3.1 Trial design

PAOLA-1 was a large, multicentre, randomised, double-blind, placebo-controlled, Phase III externally sponsored study that assessed the efficacy and safety of olaparib, added to bevacizumab versus placebo added to bevacizumab in women with newly-diagnosed advanced ovarian cancer who were in complete or partial response following first-line platinum-taxane chemotherapy with bevacizumab. An overview of the study design is shown in Figure 22 ; details of the eligibility criteria are provided in Section B.2.3.1.[73]

The rationale for the PAOLA-1 study was discussed in detail in Section B.1.3.3; however, it is important to reiterate here that the study was designed so that patients could complete their platinum-taxane–containing regimen combined with bevacizumab and continue bevacizumab maintenance treatment post-enrolment, as per the standard-of-care in participating countries, at the time of study protocol development.

Randomisation was performed in a 2:1 ratio to olaparib or matching placebo, added to bevacizumab (in both arms). Olaparib 300 mg tablets were administered twice daily (BID) for up to 2 years. As part of the intervention, treatment with intravenous bevacizumab 15 mg/kg every three weeks (Q3W) was initiated in combination with chemotherapy (with a minimum of 3 cycles of overlap) and continued after randomisation as maintenance therapy for up to 15 months in

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

total.[89] An Interactive Voice Response System / Interactive Web Response System (IVRS / IWRS) was used to allocate patients to the two study arms. The study was conducted in a double-blind manner; patients, investigators, and study centre staff were blinded to the study drug allocation.[89]

Randomisation was stratified by:[89]

• First-line treatment outcome at screening:

  • No evidence of disease (NED[‡] ), with complete macroscopic resection at initial or primary debulking surgery (PDS)

  • NED* / CR, with complete macroscopic resection at interval debulking surgery (IDS)

  • NED* / CR at screening, in patients who had either incomplete resection (at initial or interval debulking surgery) or no debulking surgery (debulking surgery considered as not feasible).

  • o Partial response (PR)

• Tumour BRCA (t BRCA ) status, as determined by BRCA testing on tumour tissue[§] : o Deleterious mutation (t BRCA m).

  • Absence of deleterious mutation (non-t BRCA m: tumour BRCA wildtype [t BRCA wt] / variant of unknown significance / unknown). Note: this group also included those patients whose tests had failed.

Patients were randomised at least 3 weeks and no more than 9 weeks after the last dose / infusion of chemotherapy and only if all major toxicities from the previous chemotherapy had resolved to Common Terminology Criteria for Adverse Events (CTCAE) Grade 1 or better (except alopecia and peripheral neuropathy).

Further details on the study are provided in the following sections. At the time of data cut-off (DCO; 22[nd] March 2019), the median duration of follow-up for the primary efficacy endpoint (of investigator-assessed PFS) was 22.7 months and 24.0 months in the olaparib + bevacizumab and placebo + bevacizumab arms, respectively.[1]

‡, Patients without assessable disease after initial debulking surgery were considered to have NED at the end of first-line chemotherapy and surgery strategy if the disease had not progressed. *, Patients with measurable or assessable disease after initial surgery or at the start of neo-adjuvant chemotherapy, whose disease was no longer detectable at the end of the chemotherapy and surgery strategy were considered to have achieved CR.

§, Prospective t BRCA testing was conducted in all screened patients at one of five French institutions recommended by the French National Cancer Institute (INCa), France. Information about g BRCA mutation status was requested (not mandated) for all patients.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 22. Overview of study design for PAOLA-1

==> picture [439 x 190] intentionally omitted <==

aBevacizumab 15 mg/kg once every 3 weeks, for up to 15 months in total. The study required ≥3 cycles of bevacizumab to be administered in combination with chemotherapy.[b] HRD score by Myriad myChoice[®] HRD plus test.[c] BRCA m status by laboratory testing of tumour sample.[d] NED, CR and PR. Abbreviations: BID: twice daily; BRCA m: breast cancer susceptibility gene mutation; CR: complete response; HRD: homologous recombination deficiency; IDS: interval debulking surgery; NED: no evidence of disease; PDS: primary debulking surgery; PR: partial response; t BRCA : tumour breast cancer susceptibility gene. Source: Ray-Coquard et al., 2019, ESMO Congress Presentation.[90 ]

B.2.3.2 Eligibility criteria

Inclusion and exclusion criteria for the PAOLA-1 study are detailed in Appendix L.1 and in Section 4 (pages 39–42) of the Clinical Study Protocol (CSP).[89] Importantly, only adult women (≥18 years of age) with newly-diagnosed, histologically-confirmed[**] , advanced (FIGO Stage III–IV) ovarian cancer, primary peritoneal cancer and/or fallopian tube cancer were enrolled onto the study.

Patients must have:

• Completed platinum-taxane chemotherapy prior to randomisation (minimum 6, maximum 9 cycles [unless discontinuation due to non-haematological toxicity after at least 4 cycles]), including:

  • Including, a minimum of 3 cycles of bevacizumab (15 mg/kg Q3W) in combination with the last 3 cycles of platinum-taxane chemotherapy. Those patients who had undergone IDS must have received a minimum of 2 cycles of bevacizumab (15 mg/kg Q3W) in combination with the last three cycles of platinum-taxane chemotherapy.

• Had NED or be in CR or PR following first-line treatment.

  • There should have been no clinical evidence of disease progression (physical exam, imaging, or CA-125) throughout the first-line treatment and prior to study randomisation.

• Been randomised at least 3 weeks and no more than 9 weeks after their last dose of chemotherapy.

**, As high-grade serous, or high-grade endometrioid, or other epithelial non-mucinous ovarian cancer in a patient with germline BRCA1 / 2 deleterious mutation.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

• All major toxicities from previous chemotherapy must have resolved to CTCAE Grade 1 or better (except alopecia and peripheral neuropathy).

• Had ECOG performance status 0 to 1.

Availability of formalin-fixed, paraffin-embedded samples from the primary tumour were mandated for centralised t BRCA testing; a test result was required for stratification.

Patients whose tumours were of non-epithelial origin (i.e. germ cell tumours) or of low malignant potential (e.g. borderline tumours) or mucinous carcinoma were excluded. Patients had to have recovered from the effects of any major surgery (surgery within 4 weeks of starting study treatment was not permitted).

Patients with clinically significant cardiovascular disease, prior history of hypertensive crisis (CTCAE Grade 4) or hypertensive encephalopathy, history / evidence of haemorrhagic disorders within 6 months of randomisation, or current / clinically-relevant bowel obstruction were excluded, as were pregnant or lactating women.

Any previous treatment with a PARPi, including olaparib, was not permitted.

B.2.3.3 Settings and locations where the data were collected

PAOLA-1 was a multicentre study, conducted in 137 study centres in 11 countries, including: Austria (6 centres), Belgium (3 centres), Denmark (1 centre), Finland (2 centres), France and Monaco (44 centres), Germany (51 centres), Italy (9 centres), Japan (7 centres), Spain (13 centres) and Sweden (1 centre).[73]

Of the patients randomised, 97.0% were in Europe and 3.0% in Japan.

B.2.3.4 Trial drugs, “background”, and concomitant medications

Study drugs: patients were assigned to receive either olaparib 300 mg BID, or matching placebo, for up to 24 months.[1, 89]

• Crossover to olaparib was not permitted in the PAOLA-1 study; however, after discontinuation of the intervention, patients could receive other treatments (including PARPi) at the investigators’ discretion.

• Patients, who, in the opinion of the treating physician could derive further benefit from continuous treatment, could be treated beyond two years; however, at the time of DCO, just three patients had exceeded the protocol-defined two years of treatment.

“Background” medication: Patients in both arms received bevacizumab 15 mg/kg intravenously Q3W, for up to 15 months in total (including the period of pre-randomisation in combination with chemotherapy and post-randomisation in combination with olaparib or placebo).

• Bevacizumab was a “non-investigational drug” since it was administered in accordance to its marketing authorisation, as the standard-of-care therapy in this setting.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Concomitant medications: When it was believed that it would not interfere with study medication, investigators could prescribe concomitant medications or treatments that were considered necessary for patients’ welfare.[89]

Permitted concomitant medications included:

• Anticoagulants (including warfarin and subcutaneous heparin).

• Anti-emetics.

• Contraceptives.

• Palliative radiotherapy (for brain metastases).

• Bisphosphonates or denozumab (for bone disease).

• Corticosteroids (for the symptomatic control of brain metastases).

Disallowed concomitant medications included:

• Other anticancer therapy (including chemotherapy, immunotherapy, hormonal therapy††, biological therapy and novel agents; exceptions for certain products for the treatment of brain metastases and bone disease). Simultaneous radiotherapy was also not permitted within 6 weeks or during the treatment period.

• Aspirin (chronic use [>325 mg/day] which is ongoing or within 10 days prior to randomisation).

• Potent CYP3A4/5 inhibitors and inducers.

Full details of permitted and disallowed concomitant medications during the study are available in Appendix L and in Section 5.7 (pages 54–56) of the PAOLA-1 CSP.[89] The administration of all medication (including investigational products), and any unplanned diagnostic, therapeutic, or surgical procedures performed during the study period (including blood transfusions) were recorded.

The most commonly-used concomitant medications in PAOLA-1 were antibiotics, antihypertensive drugs, and antiemetic agents. The categories of concomitant medications were generally well balanced in the two study arms, with the following exceptions:[73]

• A lower proportion of patients (xxxxx) in the olaparib + bevacizumab arm received antihypertensives (versus xxxxx in the placebo + bevacizumab) (Table 4).

This was due to a higher incidence of hypertension amongst patients who received olaparib, added to bevacizumab (versus placebo added to bevacizumab; see Section B.2.10). Hypertension is a known AE associated with bevacizumab; fewer instances of hypertension in the olaparib + bevacizumab group versus the placebo + bevacizumab group may indicate a potentially protective impact of olaparib on bevacizumab-associated hypertension.

• A higher proportion of patients in the olaparib + bevacizumab arm (xxxxx) received antiemetics (versus xxxxx of patients in the placebo + bevacizumab arm) (Table 4). This was due to a higher incidence of nausea in the olaparib + bevacizumab arm, versus the placebo + bevacizumab arm (discussed further in Section B.2.10), which was expected and in line with

†† Hormone replacement therapy was acceptable.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

previous trials of olaparib.[43] The majority of cases of nausea was resolved with antiemetic therapy.

• A higher proportion of patients in the olaparib + bevacizumab arm received a red blood cell transfusion (xxxxx, versus xxxxx in the placebo + bevacizumab arm; Table 4). This is attributable to a higher rate of anaemia in the olaparib + bevacizumab arm, versus the placebo + bevacizumab arm (Table 24). Again, this was as expected and consistent with the known safety profile of olaparib and is discussed further in Section B.2.10.[43]

Overall, the concomitant treatments administered were generally representative of those commonly prescribed to manage side effects of olaparib and / or bevacizumab and treat concomitant conditions in the target population and were not considered to have impacted the study results.

The number of patients with disallowed concomitant medications during study treatment (including the 30-day safety follow-up) was low (xxxxx and xxxxx in the olaparib + bevacizumab and placebo + bevacizumab group, respectively) (further detail provided in Appendix L). The use of disallowed concomitant medication did not raise concerns about the conduct of the study.[73]

Table 4. Allowed concomitant medications during study treatment, FAS

aIncludes medication with an onset date on or after the date of first dose and up to and including 30 days following the date of last dose of olaparib or placebo. Also includes medication with an onset date prior to the date of first dose but continued after the date of first dose. Abbreviations: FAS: full analysis set. Source: PAOLA-1 CSR.[73]

Details of first and subsequent therapies and / or surgery for the treatment of the cancer, after discontinuation of treatment, were also collected. Reasons for starting subsequent anti-cancer therapies were also collected and included in the exploratory assessments of OS.

B.2.3.5 Discontinuation of study treatment or withdrawal from study

Administration of olaparib or placebo continued for up to 24 months from randomisation, until confirmed radiological disease progression (according to investigator assessment, per RECIST

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

v1.1), or until unacceptable toxic effects[‡‡] , whichever occurred first, as long as the patient had a benefit and did not meet other discontinuation criteria (provided in Appendix L and Section 5.9 (pages 57–58 of the PAOLA-1 CSP).[89]

Patients discontinuing from treatment were not seen as withdrawing from the study and were followed-up for disease progression (if discontinued in the absence of progression), PFS2, and OS (final analysis), as per the protocol schedule. Reasons for withdrawal from the study are also provided in Appendix L and Section 5.9 (page 59) of the PAOLA-1 CSP.[89]

At the time of the first DCO, 331 patients (61.9%) in the olaparib + bevacizumab arm and 194 patients (72.7%) in the placebo + bevacizumab arm (FAS) had discontinued olaparib or placebo, respectively.[1, 73]

The most common reason for discontinuation of study treatment was disease progression as per RECIST v1.1 criteria:

• Olaparib + bevacizumab: 34.0%

• Placebo + bevacizumab: 58.1%.

Further information on patient disposition at the time of the first DCO is provided in Appendix D1.

B.2.3.6 Primary, secondary and exploratory endpoints

RECIST v1.1 criteria were used to evaluate tumour responses at each visit, to determine when a patient experienced disease progression and their best objective response (BOR).[1, 73] Further information on tumour assessments, including the frequency of scans, is provided in Appendix L1.4 and in Section 6.3 (pages 65–67) of the PAOLA-1 CSP.[89]

The primary objective of the PAOLA-1 study was:

“To determine the efficacy of olaparib maintenance compared with placebo, by investigatorassessed PFS (according to modified RECIST version 1.1) in patients with high-grade epithelial ovarian, fallopian tube, or peritoneal cancer who are in clinical CR or partial response following first-line platinum-taxane chemotherapy plus bevacizumab, and were planned to pursue bevacizumab in the maintenance phase up to a total of 15 months”.

PFS was defined as the time from randomisation until the date of objective radiological disease progression according to modified RECIST v1.1 or death (by any cause in the absence of progression), regardless of whether the patient discontinued randomised therapy or received another anticancer therapy prior to progression.

Secondary efficacy and safety objectives of the PAOLA-1 study included:[1, 73]

• Time to earliest progression by RECIST or CA-125 or death , defined as the time from randomisation to the earlier date of modified RECIST v1.1 or CA-125 progression or death by any cause

‡‡ Any toxicity observed during the study treatment phase was managed using dose interruption if deemed appropriate by the investigator; further information on this is provided in Section B.2.10.2.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

• Time from randomisation to first subsequent therapy or death (TFST), defined as the time from randomisation to the earlier of first subsequent therapy start date following study treatment discontinuation, or death

• Time from randomisation to second progression (PFS2), defined as the time from the date of randomisation to the earliest progression event subsequent to that used for the primary PFS, or death

• Time from randomisation to second subsequent therapy or death (TSST), defined as the time from randomisation to the earlier of the second subsequent therapy start date following study treatment discontinuation, or death

• Overall survival (OS), defined as time from the date of randomisation until death due to any cause)

• Health-related quality-of-life (HRQoL) measures, including:

  • EORTC QLQ-C30: an integrated system for assessing the HRQoL of cancer patients, composed of 5 functional scales, 9 symptom scales and 1 global status/quality of life scale

  • EORTC QLQ-OV28 questionnaire: a specific ovarian cancer module, composed of 28 questions, including 10 symptom scales and 3 sexual functioning scales

  • EQ-5D-5L: a standardised measure of health status. The questionnaire comprises 6 questions that cover 5 dimensions of health (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and how the patient feels.

• Safety and tolerability analyses , including AEs, serious adverse events (SAEs), discontinuation of investigational product due to AE(s), deaths, laboratory data, vital signs and echocardiograms (ECGs)

Further information on key endpoints, including definitions, can be found in Appendix L and Section 3 (page 38) of the PAOLA-1 CSP.[73]

All efficacy endpoints were evaluated in the FAS. The efficacy and tolerability of olaparib versus placebo, when added to bevacizumab, in the HRD-positive subgroup – the focus of this submission - was analysed in an exploratory pre-specified subgroup analysis.

B.2.3.7 Biomarker analyses

Tumour samples from PAOLA-1 patients were tested post-randomisation (but prior to database lock) using the Myriad myChoice[®] HRD Plus test , which detects and classifies the following biomarkers simultaneously in tumour tissue:[73]

• Sequence variants and large rearrangements in BRCA1 and BRCA2 , as well as an additional 13 HR-repair genes: ATM, BARD1, BRIP1, CHEK1, CHEK2, CDK12, FANCL, PALB2, PPP2R2A, RAD51B, RAD51C, RAD51D and RAD54L .

• Myriad HRD score, which is designed to identify a comprehensive signature / genomic scar for HR deficiency by testing genome-wide single nucleotide variants. It is determined by measuring three elements, namely, loss of heterozygosity, telomeric allelic imbalance, and large-scale state transitions.

The overall Myriad HRD status is based on the Myriad HRD score and t BRCA m status. A positive Myriad HRD status is determined either by presence of a t BRCA1/2 mutation, or by an HRD score

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

at or above a pre-specified cut-off in the absence of a BRCA1/2 mutation. The Myriad HRD cutoff of 42 was used for the “HRD-positive” subgroup data presented in this submission, unless otherwise stated – this cut-off detects 95% of BRCAm tumours and has been extensively investigated as a biomarker of PARPi benefit in ovarian cancer.[59, 91, 92]

Of the 806 randomised patients, 664 (82.4%) had an available Myriad HRD status. The proportion of Myriad t BRCA mutations and Myriad HRD status was well balanced between treatment arms (please see the PAOLA-1 clinical study report [CSR] Section 10.4.2.1, Table 18, for further information). In patients where t BRCA mutation status was determined both by on-study prospective (screening laboratory) testing and by post-randomisation central t BRCA testing at Myriad, there was high (96.3%) concordance between test results.

142 (17.6%) patients had an unknown Myriad HRD status - 3.0% of patients had no available sample to send to Myriad and 14.6% of patients had a cancelled or failed test.

B.2.3.8 Baseline characteristics

Between July 2015 and September 2017, 806 patients were randomised in a 2:1 ratio to olaparib + bevacizumab and placebo + bevacizumab arms of the PAOLA-1 study. The disposition of patients in the study and a description of the analysis sets (including the [FAS] and the HRDpositive group) are provided in Appendix D.2 and Section B.2.4.1, respectively.

Randomisation in the PAOLA-1 study was stratified by first-line treatment outcomes and t BRCA status, to ensure balanced allocation to the olaparib + bevacizumab, or placebo + bevacizumab groups (see Section B.2.3.1). Other prognostically-important baseline characteristics, such as patient age, performance status, disease stage, and histology, were also well-balanced between olaparib + bevacizumab and placebo + bevacizumab groups (Table 5; Section 10.4 [page 98] of the PAOLA-1 CSR).[73]

HRD testing was conducted post-randomisation; however, similar proportions of patients (47.5% and 49.1% in the olaparib + bevacizumab and placebo + bevacizumab arms, respectively), were “HRD-positive” (as per the Myriad myChoice[®] HRD Plus test cut-off score of 42). This proportion was as expected and aligned to published data that show that approximately half of all ovarian carcinomas have mutations that confer HR-deficiency (detailed in Figure 5, B.1.3.1). This prespecified subgroup of patients who are HRD-positive are the focus of this submission .

HRD-positive patients who received olaparib + bevacizumab or placebo + bevacizumab, were wellbalanced across key baseline characteristics, and reflective of the FAS. HRD-positive patients in the olaparib + bevacizumab and placebo + bevacizumab arms were also very well matched in terms of prior surgery (upfront debulking surgery, interval debulking surgery, or no surgery), surgical outcomes (presence or absence of residual macroscopic disease), and response to firstline chemotherapy (NED, CR, PR). These data are summarised in Table 5. Baseline characteristics for the FAS are also shown alongside for completeness, and to highlight the consistency / similarities between the FAS and HRD-positive populations. Generalisability of patients enrolled onto PAOLA-1 versus the real-world cohort of patients in the UK is discussed in Section B.2.13.2.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Table 5. Patient characteristics in PAOLA-1

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

*Percentages may not total 100 because of rounding

†Other defined as clear cell (n=2, olaparib + bevacizumab), undifferentiated (n=1, olaparib + bevacizumab; n=6, placebo + bevacizumab) or other (n=3, olaparib + bevacizumab; n=2, placebo + bevacizumab)

‡No evidence of disease defined as complete macroscopic resection after initial cytoreductive surgery, no radiologic evidence of disease, and a normal CA-125 level after chemotherapy

§Clinical complete response defined as the disappearance of all measurable/assessable disease and normalisation of CA-125 levels

¶Clinical partial response defined as radiologic evidence of disease and/or an abnormal CA-125 level

**Tumor BRCA mutation or HRD score ≥42

††HRD score <42

‡‡Unknown defined as an inconclusive, missing or failed test

§§HRD score ≥42; t BRCA m determined by Myriad® MyChoice HRD Plus Test

Abbreviations: CA: cancer antigen; CR: complete response; eCRF: electronic case report form; ECOG: Eastern Cooperative Oncology Group; FIGO: International Federation of Gynaecology and Obstetrics; HRD: homologous recombination deficiency; HRR: homologous recombination repair; ITT: intention-to-treat; NED: no evidence of disease; PR: partial response; t BRCA m: tumour breast cancer susceptibility gene mutation. Source: Ray-Coquard et al., 2019.[1]

B.2.4 Statistical analysis and definition of study groups in the

relevant clinical effectiveness evidence

All analyses were performed in accordance with a comprehensive statistical analysis plan (SAP), which details the analyses to be conducted, summaries produced, and the analysis sets upon which they would be based (Sections 1–3 of the PAOLA-1 SAP).[93]

The main hypothesis evaluated in the PAOLA-1 study was that olaparib added to bevacizumab achieves improved efficacy versus placebo added to bevacizumab (assessed through the primary endpoint of investigator-assessed PFS, per RECIST1.1), in women with newly-diagnosed advanced ovarian cancer who were in complete or partial response following first-line platinumtaxane chemotherapy with bevacizumab. As per the SAP, the study would have met this objective upon reporting a statistically significant PFS benefit of olaparib versus placebo.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.4.1 Analysis sets

Two main analysis sets were defined for the PAOLA-1 study.[73]

All efficacy and HRQL data were analysed using the FAS, which included all randomised patients on an intention-to-treat (ITT) basis (i.e. based on treatment assigned at randomisation, regardless of whether treatment was received). The FAS for the PAOLA-1 study included 806 patients in total (N=537 and 269, for olaparib + bevacizumab and placebo + bevacizumab arms, respectively).

Summaries of safety and tolerability assessments were based on the safety analysis set (SAS) , which included all patients who received at least one dose of randomised study medication and had at least one safety follow-up assessment. Two patients randomised to the olaparib + bevacizumab arm and two patients randomised to the placebo + bevacizumab arm did not receive any doses of study treatments, and were therefore, excluded from the SAS (total, N=802; olaparib + bevacizumab, N=535; placebo + bevacizumab, N=267).[73] Erroneously treated patients, i.e. those who were randomised to olaparib but actually received placebo, and vice versa, were accounted for by actual treatment received. Patients receiving treatment from more than one treatment arm were accounted for based upon the initial treatment started.

As stated previously, the evaluation of the efficacy of olaparib + bevacizumab versus placebo + bevacizumab, in patients whose tumours tested HRD-positive, or HRD-negative was a prespecified subgroup analysis in the PAOLA-1 study. HRD-negative or unknown, and HRD-unknown groups were also analysed in post-hoc exploratory analyses. Of the 806 patients in the FAS:

• 387 were HRD-positive (olaparib + bevacizumab, N=255; placebo + bevacizumab, N=132),

• 277 were HRD-negative (olaparib + bevacizumab, N=192; placebo + bevacizumab, N=85),

• 142 were of “unknown” status (olaparib + bevacizumab, N=90; placebo + bevacizumab, N=52).

B.2.4.2 Statistical analyses

Statistical analyses were performed by the Biostatistics Group, AstraZeneca. All calculations were performed with SAS[®] software Version 9.4 (SAS Institute, Inc, Cary, North Carolina), unless otherwise stated. Further information on sample size calculation and analysis of key outcome variables (including supporting sensitivity and subgroup analyses, and censoring) are provided in Appendix L1.6 (and described in detail in Section 11 and Section 9.8 PAOLA-1 CSP and CSR respectively).[73, 89]

Briefly, the study planned to randomise 762 patients (with an additional 24 patients randomised in Japan by GOTIC (Gynaecologic Oncology Trial and Investigation Consortium)). The progressionfree survival (PFS) analysis was planned to occur when approximately 458 investigator-declared progression events had occurred (~57% maturity), which would have >80% power to show statistically significant PFS at a 2-sided 5% level, assuming that the true treatment effect was a hazard ratio (HR) of 0.75. This would translate to a median PFS improvement from 15.8 months (in the placebo + bevacizumab arm) to 21.1 months (in olaparib + bevacizumab arm).

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Global recruitment to the study closed when 806 patients were randomised. DCO for the primary analysis of PFS (22 March 2019) took place when 474 progression events had occurred (58.8% maturity), approximately 45 months after the first patient was randomised. The PAOLA-1 study met its primary endpoint at the time of this analysis, demonstrating a statistically significantly and clinically-meaningful improvement in investigator-assessed PFS in the FAS, in favour of olaparib + bevacizumab versus placebo + bevacizumab (HR=0.59; 95% CI: 0.49, 0.72; p<0.0001; median PFS of 22.1 months in the olaparib + bevacizumab arm, versus 16.6 months in the placebo + bevacizumab arm). The robustness of these data were confirmed in a range of sensitivity analyses (including PFS by blinded independent central review [BICR]; described in the PAOLA-1 CSR[73] ); the PFS benefit of olaparib versus placebo, when added to bevacizumab, was evident across all pre-specified subgroups (Appendix E).

The PAOLA-1 study employed a multiple testing procedure to strongly control for type I error at 2.5% (1-sided) across the primary endpoint of PFS and the key secondary endpoints of PFS2 and OS. Specifically, PFS2 will be tested only after statistical significance is shown for PFS. OS will be tested only after the null hypotheses is rejected for both PFS and PFS2 (Figure 23). Further information on this is presented below and in Appendix L1.6.

Figure 23. Hierarchical testing strategy employed in PAOLA-1

==> picture [389 x 176] intentionally omitted <==

All endpoints are investigator-assessed. Abbreviations: OS: overall survival; PFS: progression-free survival; PFS2: second progression-free survival. Source: PAOLA-1 Protocol.[89]

An interim analysis of PFS2 was planned at the time of the primary PFS analysis. However, these data were just 39% mature at this time, and the study continues to final analysis of PFS2. The latter is planned when data are ~53% mature or after a maximum duration of one year following the primary PFS analysis, whichever occurs first.

An interim OS analysis will be performed at the time of final PFS2 analysis.[89, 94] If PFS2 data are not statistically significant, a final summary of OS will be performed when the OS data are ~60% mature, or three years after the main PFS analyses (xxxxxxxxxx), whichever comes first.[89, 94]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.5 Quality assessment of the relevant clinical effectiveness

evidence

PAOLA-1 was performed in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, under the auspices of an independent data and safety monitoring committee.[1, 73] This study was conducted by ARCAGY (Association de Recherche Cancers Gynécologiques) Research on behalf of the European Network for Gynaecological Oncological Trial (ENGOT) and the Gynaecologic Cancer InterGroup (GCIG).

A complete quality assessment in accordance with the NICE-recommended checklist for assessment of bias in RCTs is presented in Table 6 and Appendix D.3. The risk of bias in the PAOLA-1 study is confirmed as being low.

Table 6. Overview of quality assessments for PAOLA-1

Adapted from Systematic reviews: CRD’s guidance for undertaking reviews in health care (University of York Centre for Reviews and Dissemination).

B.2.6 Clinical effectiveness results of the relevant trials

As explained previously, the Phase III PAOLA-1 clinical trial is the only study that assessed the clinical effectiveness of olaparib added to bevacizumab (15mg/kg Q3W) as maintenance treatment for women with newly-diagnosed advanced ovarian cancer who are in complete or partial response after first-line platinum-based chemotherapy with bevacizumab (15mg/kg Q3W).

The PAOLA-1 study met its primary endpoint of investigator-assessed PFS in the FAS, demonstrating a statistically-significant and clinically-meaningful benefit for olaparib, when added to bevacizumab maintenance (versus placebo + bevacizumab; HR=0.59; 95% CI: 0.49, 0.72, p<0.0001).

As described in Figure 1, pre-planned subgroup analyses by biomarker status showed that the observed PFS benefit was primarily driven by those women whose tumours were HRD-positive (using the myChoice[®] HRD Plus assay [Myriad Genetic Laboratories, Inc.]) (HR = 0.33, 95% CI: 0.25, 0.45). At the time of analysis, no additional PFS or OS benefit was observed from adding olaparib to bevacizumab maintenance treatment in women with HRD-negative tumours. These results are currently being followed-up in order to better understand the clinical effectiveness of adding olaparib to bevacizumab as maintenance treatment in the FAS.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

However, considering the evidence available at present, we are seeking reimbursement in the HRD-positive group of patients, where the addition of olaparib to bevacizumab has shown a robust and compelling benefit across a range of clinically-meaningful endpoints – these data are summarised below and discussed in detail in the following sections.

Table 7. Summary of efficacy at the 22 March 2019 DCO, FAS and HRD-positive population

a, Calculated using KM techniques; *, estimated from a stratified Cox proportional hazards model stratifed by first line treatment outcome and t BRCA status

Abbreviations: BICR: blinded independent central review; CA-125: cancer antigen-125; CI: confidence interval; HR: hazard ratio; KM: Kaplan-Meier; OS: overall survival; PFS2: second progression-free survival; RECIST v1.1: Response Evaluation Criteria in Solid Tumors version 1.1; t BRCA : tumour breast cancer susceptibility gene; TDT: time to treatment discontinuation or death; TFST: time to first subsequent therapy; TSST: time to second subsequent therapy. Source: Ray-Coquard et al., 2019;[1] PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.6.1 Primary endpoint: investigator-assessed PFS (per RECIST v1.1)

FAS:

The PAOLA-1 study met its primary endpoint of investigator-assessed PFS (according to RECIST v1.1) during a pre-planned analysis (22 March 2019 DCO), demonstrating a statistically-significant and clinically-meaningful benefit for olaparib added to bevacizumab (versus placebo + bevacizumab) in the FAS (HR=0.59, 95% CI: 0.49, 0.72; p<0.0001) (Figure 24; Table 8).[1] Median duration of PFS was nearly two years (22.1 months) in the olaparib + bevacizumab arm versus 16.6 months in the placebo + bevacizumab arm. Nearly half (47.9%) of the patients in the olaparib + bevacizumab arm were progression-free at the time of DCO (versus 27.9% of patients in the placebo + bevacizumab arm; Table 8).[73 ]

This is a remarkable result versus an active comparator , in a population of women unselected by surgical outcome or biomarker status, and thus representative of the real-world UK cohort of patients (Table 8; see Section B.2.13.2 for further information on generalisability). The sensitivity analysis of PFS by BICR was conducted, which showed consistent results with the investigator assessment of PFS and confirmed its robustness (both shown in Table 9). For results of other sensitivity analyses, see CSR (Section 11.1.5); results of the pre-specified subgroup analyses of PFS are discussed in Appendix E.

Table 8. Progression status at time of PFS analysis (22 March 2019 DCO), FAS

a , Does not include RECIST progression events that occur after two or more missed visits or within two visits of baseline where the patient has no evaluable visits or does not have a baseline assessment. b , Not necessarily mutually exclusive categories. c , Death in the absence of RECIST progression or death occurring within two visits of baseline where the patient has no evaluable visits or does not have a baseline assessment. Does not include deaths that occur after two or more missed visits. d , RECIST progression event occurred after two or more missed visits or within two visits of baseline where the patient has no evaluable visits or does not have a baseline assessment.

e , Death which occurred after two or more missed visits in the absence of RECIST progression. f , Patients known to be alive and without RECIST progression. g , Patients at last evaluable RECIST assessment. Note: This analysis is based on investigator review of radiological scans. Progression status at time of PFS analysis using BICR is provided in CSR Table 14.2.1.1.3. Abbreviations: BICR: blinded independent central reviewer; DCO: data cut-off; FAS: full analysis set; PFS: progression-free survival; RECIST: response evaluation criteria in solid tumours. Source: PAOLA-1 CSR;[73] Ray-Coquard et al., 2019.[1]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Table 9: Summary of PFS analysis by investigator assessment and BICR (22 March 2019 DCO), FAS

a , progression-free survival is defined as time from randomisation until date of RECIST progression or death b , calculated using Kaplan-Meier techniques c, time from randomisation to date of censoring d , estimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and tBRCA status e , determined using log-rank test stratified by first line treatment outcome and t BRCA status ***,** This analysis uses programmatically derived overall visit response based on investigator review of radiological scans Note: Progression includes deaths in the absence of RECIST progression, progression-free includes patients who have not progressed or died. Abbreviations: BICR: blinded independent central review; CI: confidence interval; DCO: data cut-off; ITT: intention-to-treat; HR: hazard ratio; PFS: progression-free survival; IQR: interquartile range; RECIST: response evaluation criteria in solid tumours. Source: PAOLA-1 CSR;[73] Ray-Coquard et al., 2019.[1]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

==> picture [455 x 503] intentionally omitted <==

----- Start of picture text -----
Figure 24. Kaplan-Meir plot of investigator-assessed PFS for olaparib + bevacizumab
versus placebo + bevacizumab (22 March 2019 DCO), FAS (top) and HRD-positive (bottom)
FAS:
HRD-positive (focus of submission):
----- End of picture text -----

Abbreviations : CI, confidence interval; DCO, data cut-off; FAS: full analysis set; HRD: homologous recombination deficiency; PFS, progression-free survival. Source : Ray-Coquard et al., 2019.[1]

HRD-positive population (focus of this submission):

Pre-specified subgroup analysis showed an even greater PFS benefit from the addition of olaparib to bevacizumab maintenance treatment in the HRD-positive group of patients than in the FAS ( HR=0.33 , 95% CI: 0.25, 0.45 [unstratified][§§] ;Table 10, Table 11).[1, 73]

§§ The HR (0.35; 95% CI: 0.26, 0.47), estimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and t BRCA status was consistent with the unstratified analysis and shows that the effect is stable to stratification on first-line outcome and t BRCA .

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

The median duration of investigator-assessed PFS in the olaparib + bevacizumab arm was > three years (37.2 months) and over twice as long versus patients in the placebo + bevacizumab arm (median PFS = 17.7 months) (Table 11).

Of note, the investigator-assessed PFS data for the HRD-positive group described in this section were used to inform the health economic modelling described in Section B.3.

Table 10. Progression status at time of investigator-assessed PFS analysis (22 March 2019 DCO), HRD-positive population

a , Does not include RECIST progression events that occur after two or more missed visits or within two visits of baseline where the patient has no evaluable visits or does not have a baseline assessment.

b , Not necessarily mutually exclusive categories.

c , Death in the absence of RECIST progression or death occurring within two visits of baseline where the patient has no evaluable visits or does not have a baseline assessment. Does not include deaths that occur after two or more missed visits.

d , RECIST progression event occurred after two or more missed visits or within two visits of baseline where the patient has no evaluable visits or does not have a baseline assessment.

e , Death which occurred after two or more missed visits in the absence of RECIST progression.

f , Patients known to be alive and without RECIST progression.

g , Patients at last evaluable RECIST assessment.

Note: This analysis is based on investigator review of radiological scans.

Abbreviations : DCO: data cut-off; HRD: homologous recombination deficiency; PFS: progression-free survival; RECIST: response evaluation criteria in solid tumours

Source: PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Table 11. Summary of PFS analysis by investigator assessment (22 March 2019 DCO), HRDpositive population

aProgression-free survival is defined as time from randomisation until date of RECIST progression or death bCalculated using Kaplan-Meier techniques cTime from randomisation to date of censoring dEstimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and tBRCA status

Note: Progression includes deaths in the absence of RECIST progression, progression-free includes patients who have not progressed or died. Based on investigator RECIST assessment. Abbreviations: CI: confidence interval; DCO: data cut-off; IQR: interquartile range; ITT: intention-to-treat; HR: hazard ratio; HRD, homologous recombination deficiency; PFS: progression-free survival. Source: Ray-Coquard et al., 2019;[1] PAOLA-1 CSR;[73] PAOLA-1 HRD-positive subgroup data.[74]

The olaparib + bevacizumab and placebo + bevacizumab Kaplan-Meier curves separated early and remained separated, consistent with a sustained PFS benefit for olaparib versus placebo (Figure 24).[74] This is also supported by the non-overlapping 95% CIs of 1-year, 2-year, and 3-year landmark PFS assessments for olaparib versus placebo, when added to bevacizumab maintenance treatment (Table 11).

At the time of the primary PFS analysis, xxxxxxxxxxxxxxxxxxxxxxxxx of patients in the olaparib + bevacizumab arm were progression-free after >2 years of follow-up (xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx), versus just xxxx% of patients in the placebo + bevacizumab arm (xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx; Table 10).[74 ]

B.2.6.2 Key secondary endpoints: TFST, PFS2, TSST, and OS in the HRD-

positive population

The observed PFS benefit of olaparib added to bevacizumab (versus placebo + bevacizumab) in the HRD-positive group of patients - the focus of this submission - is supported by data on the key secondary efficacy endpoints of TFST, PFS2, TSST, and OS, and collectively present a compelling body of evidence for the clinical effectiveness of olaparib added to bevacizumab maintenance in the population of interest for this appraisal.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Time to first subsequent therapy or death (TFST); HRD-positive population

Consistent with the PFS benefit, the addition of olaparib to bevacizumab maintenance treatment also extended TFST (relative to placebo + bevacizumab) in the HRD-positive group (xxxxxxxxxxxxxxxxxxxxxxxxxxx).[74] This is evident from the KM-curves for TFST, which separated early in favour of olaparib + bevacizumab and remained separated for the duration of the followup period (Figure 25). Median TFST was xxxxxxxxxxx in the olaparib + bevacizumab arm, despite xxxxxxxx of follow-up (xxxxxxxxxxxxxxxxxxxxxxx). The benefit of olaparib added to bevacizumab in extending TFST versus placebo + bevacizumab is also supported by landmark assessments between 6 months and 3 years (Table 12).

An extension to TFST was also observed in the FAS (xxxxxxxxxxxxxxxxxxxxxxxxxxx),[73] and was consistent with the PFS benefit observed in this dataset (described in further detail in Section 11.1.2.3 of the CSR).

Table 12. TFST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

aTime to first subsequent therapy is defined as time from randomisation until first subsequent anti-cancer therapy or death bCalculated using Kaplan-Meier techniques cTime from randomisation to date of censoring dEstimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and tBRCA status

Abbreviations: CI: confidence interval; DCO: data cut-off; HR: hazard ratio; HRD: homologous recombination deficiency; TFST: time to first subsequent therapy. Source: PAOLA-1 HRD-positive subgroup data.[74]

Figure 25. TFST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

x

Abbreviations: bd: twice daily; DCO: data cut-off; HRD: homologous recombination deficiency; TFST: time to first subsequent therapy. Source: PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

At the time of the 22 March 2019 DCO, xxxxxxxxxxxxxxxxx HRD-positive patients who received olaparib in addition to bevacizumab, and xxxxxxxxxxxxxxxxx HRD-positive patients who received placebo in addition to bevacizumab, had started a first subsequent anticancer therapy (Table 12).[74]

The most commonly-used first subsequent therapies in both arms were carboplatin or pegylated liposomal doxorubicin (Table 13), which is consistent with UK clinical practice.[74] Although crossover to olaparib was not permitted in the PAOLA-1 study, patients could receive a PARPinhibitor following disease progression (e.g. outside of the study) through other clinical trials or commercially available products. More patients in the placebo + bevacizumab arm received a PARPi as their first subsequent therapy relative to the olaparib + bevacizumab arm (xxxxxxxxxxxxxxx respectively).[74] Although greater use of subsequent PARPi therapy in the placebo + bevacizumab arm can extend post-progression survival in the subgroup of patients who receive these treatments and underestimate the PFS2, TSST, and OS benefit achieved with the addition of olaparib to bevacizumab maintenance treatment, this is broadly reflective of real-world practice.

More patients in the placebo + bevacizumab arm also received an anti-angiogenic agent as their first subsequent therapy (xxxxxxxxxxxxxxxxx);[74] currently, no anti-angiogenic therapies are recommended in England in the relapsed ovarian cancer setting. The use of anti-angiogenic treatments in the placebo + bevacizumab is likely to further bias PFS2, TSST, and OS data in favour of the control arm and underestimate the true benefit of the PAOLA-1 regimen.

Table 13. Post-discontinuation anticancer therapy, AZ Medic review, HRD-positive population

Note: Patients who received subsequent therapy are counted once per category and type. Patients may appear under more than one subsequent treatment type. For two patients the investigator recorded the first subsequent therapy in subsequent therapy number 2. Abbreviations: AZ: AstraZeneca; RD: homologous recombination deficiency; PARPi: poly-ADP ribose polymerase inhibitor; PLD: pegylated liposomal doxorubicin. Source: PAOLA-1 HRD-positive subgroup data.[74]

Time to second progression or death (PFS2), HRD-positive population

The use of multiple subsequent therapies after disease progression can impact on the overall survival benefit of new interventions. Intermediate clinical endpoints, such as PFS2 and TSST

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

provide information about the long-term benefits of a treatment after disease progression, and are important measures of real-life treatment decisions and patient experience.

PFS2 events were based on radiological, CA-125, or symptomatic progression as assessed by the investigator, or death. Consistent with PFS data, the addition of olaparib to bevacizumab maintenance treatment also substantially extended PFS2 versus placebo + bevacizumab in the HRD-positive group of patients (xxxxxxxxxxxxxxxxxxxxxxxxxxx; Table 14).[74] Median PFS2 was xxxxxxxxxxx in the olaparib + bevacizumab arm, despite xxxxxxxx of follow-up (xxxxxxxxxxxxxxxxxxxxxxx).

KM-curves for PFS2 separated early in favour of olaparib + bevacizumab and remained separated for the duration of follow-up (Figure 11).[74] Although these data are currently immature (xxxxx maturity across both arms), they are supportive of the PFS benefit of olaparib added to bevacizumab translating into an extension (delay) to time to second progression or death – this has important implications for patients (who are spared further courses of cytotoxic chemotherapy, added to the substantial physical and psychological impact of disease progression), as well as their family and carers. At the time of DCO, xxxxx of patients in the olaparib + bevacizumab arm and xxxxx of patients in the placebo + bevacizumab arm were classified as not having had a second progression.

Table 14. Time to second progression (PFS2) for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

aSecond progression-free survival is defined as time from randomisation until second progression as recorded in the CRF.[b] Calculated using Kaplan-Meier techniques.[c] Time from randomisation to date of censoring.[d] Estimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and tBRCA status. Abbreviations: CI: confidence interval; DCO: data cut-off; HR: hazard ratio; HRD: homologous recombination deficiency; IQR: interquartile range; PFS2: second progression-free survival. Source: PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 26. Time to second progression (PFS2) for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

x

Abbreviations: bd: twice daily; DCO: data cut-off; HRD: homologous recombination deficiency; PFS2: time to second progression-free survival. Source: PAOLA-1 HRD-positive subgroup data.[74]

Time to second subsequent therapy or death (TSST), HRD-positive population

Consistent with the PFS2 data described above, the addition of olaparib to bevacizumab also prolonged TSST, versus placebo + bevacizumab, in the HRD-positive group (Table 15). The TSST KM curves separately early in favour of olaparib + bevacizumab, and remained separated, demonstrating a sustained benefit versus placebo + bevacizumab during the study period (Figure 27). This was further supported by non-overlapping 95% CIs of landmark TSST assessments from 18 months onwards (through to 3 years; Table 15).

The hazard ratio for TSST was consistent with that of PFS2 (xxxxxxxxxxxxxxxxxxxxxxxxx versus xxxxxxxxxxxxxxxxxxxxxxxx]).[74] Median TSST was xxxxxxxxxxx in the olaparib + bevacizumab arm, despite maximum follow-up of xxxxxxxxxxx (median=xxxxxxxxxxx; Table 15).[74]

Table 15. TSST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

a , time to first subsequent therapy is defined as time from randomisation until first subsequent anti-cancer therapy or death

b , calculated using Kaplan-Meier techniques c, time from randomisation to date of censoring d , estimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and tBRCA status Abbreviations: CI: confidence interval; DCO: data cut-off; HR: hazard ratio; HRD: homologous recombination deficiency; IQR: interquartile range; TSST: time to second subsequent therapy. Source: PAOLA-1 HRD-positive subgroup data.[74]

Data from the FAS showed a similar trend in TSST (xxxxxxxxxxxxxxxxxxxxxxxxxxxx), although of a lesser magnitude, consistent with HRD-positive patients deriving the most benefit from the PAOLA-1 regimen. These data are described in more detail in Section 11.1.2.4 of the CSR.[73]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 27. TSST for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

x Abbreviations: bd: twice daily; DCO: data cut-off; HRD: homologous recombination deficiency; TSST: time to second subsequent therapy. Source: PAOLA-1 HRD-positive subgroup data.[74]

At the time of the 22 March 2019 DCO, xxxxxxxxxxxxxxxxx HRD-positive patients in the olaparib + bevacizumab arm and xxxxxxxxxxxxxxxxx HRD-positive patients in the placebo + bevacizumab arm had received a second subsequent therapy (Table 16).[74] Greater use of a second subsequent therapy in the placebo + bevacizumab arm was consistent with more patients experiencing disease progression in this arm. The most frequently used second subsequent therapies in both arms were non-platinum cytotoxic drugs (such as paclitaxel, gemcitabine, and pegylated liposomal doxorubicin). More patients in the placebo + bevacizumab arm received carboplatin; however, this slight imbalance is unlikely to have had significant impact on the overall study results due to its modest efficacy in this setting.[73]

More patients in the placebo + bevacizumab arm received targeted therapies, including PARPi – the use of the latter in this setting is as per routine NHS practice. Other targeted therapies, that are not currently licensed / recommended in England are unlikely to significantly impact on the overall results due to the small patient numbers who received these (Table 16).

Table 16. Second post-discontinuation anticancer therapy, investigator review, HRDpositive population

*, According to the AZ Medic

Note: Patients who received subsequent therapy are counted once per category and type. Patients may appear under more than one subsequent treatment type. For two patients the investigator recorded the first subsequent therapy in subsequent therapy number 2.

Abbreviations: ; PARPi: polyadenosine 5’diphospho ribose polymerase inhibitor. Source: PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Overall survival (OS); HRD-positive population

OS is the main endpoint that is routinely used to demonstrate superiority of antineoplastic therapies. OS data in the HRD-positive population of PAOLA-1 had reached xxxxx maturity (xxxx% in the olaparib + bevacizumab arm, and xxxx% in the placebo + bevacizumab arm) at the time of DCO and the vast majority of patients were still alive in both study arms of the HRD-positive group (Table 17).[74]

Data from the PAOLA-1 study show a clear OS benefit in favour of olaparib added to bevacizumab versus placebo + bevacizumab (xxxxxxxxxxxxxxxxxxxxxxxxxxx), despite low maturity. A xxxxreduction in the risk of death is remarkable in this setting, in a population of women who were unselected by prior surgical outcomes and included those who did not have BRCA m tumours.

A clear separation of OS KM-curves, in favour of olaparib + bevacizumab, was observed from ~6 months onwards; the KM-curves remained separated for the duration of follow-up, consistent with a sustained OS benefit for olaparib + bevacizumab versus placebo + bevacizumab (Figure 28). Median OS had not been reached for either arm.[73]

Table 17. OS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

aOverall survival is defined as time from randomisation until death. bCalculated using KM techniques. cTime from randomisation to date of censoring.[d] Estimated from a stratified Cox proportional hazards model stratified by first line treatment outcome and t BRCA status. Abbreviations: CI: confidence interval; DCO: data cut-off; HR: hazard ratio; HRD: homologous recombination deficiency; IQR, interquartile range; OS: overall survival. Source: PAOLA-1 HRD-positive subgroup data.[74]

Figure 28. OS for olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

x

Abbreviations: DCO: data cut-off; HRD: homologous recombination deficiency; OS: overall survival. Source: PAOLA-1 HRD-positive subgroup data.[74]

An OS benefit in favour of olaparib + bevacizumab was not observed in the FAS in this data set, although this was based on highly immature data xxxxxxx xxxxxxxxxxxxx xxxxxxxxxxxxx xxxxxxxxxx.[73] The trend towards an improved OS benefit with olaparib plus bevacizumab in HRD

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved Page 69 of 177

positive patients when compared with the FAS, is consistent with the same trends observed for PFS, where a greater effect was seen in HRD-positive patients versus the FAS. Trends in OS in both the FAS and the HRD-positive subgroup will be reassessed in subsequent data cuts. An interim analysis of OS is planned at time of the final PFS2 analysis, if the final PFS2 is statistically significant in the FAS. Otherwise, a final OS summary will be performed when the OS data are ~60% mature or three years after the primary PFS analysis, whichever comes first.

Best objective response (BoR), HRD-positive population

Among the HRD-positive patients who had evidence of disease at randomisation (i.e. presence of target or non-target lesions at baseline), a greater ORR of xxxxx was achieved amongst those who received olaparib + bevacizumab (versus xxxxx for those who received placebo + bevacizumab). Of these, majority of patients had a CR (Table 18).[74] These results illustrate that the clinical benefit of olaparib extends beyond delaying progression, and includes reducing tumour volume beyond that which can achieved with bevacizumab alone.

The majority of patients who did not achieve a response had stable disease for ≥24 weeks (Table 18). Disease progression was recorded in just x patients in the olaparib + bevacizumab arm and x patients in the placebo + bevacizumab arm (amongst HRD-positive patients with evidence of disease at randomisation).

Table 18. Best objective response in patient with radiological evidence of disease, any target or non-target lesions; olaparib + bevacizumab versus placebo + bevacizumab (22 March 2019 DCO), HRD-positive population

a, Response does not require confirmation.

Note: This analysis was based on investigator CRF assessment per modified RECIST version 1.1. Patients with evidence of disease at baseline were considered evaluable for response.

Abbreviations: CI: confidence interval; DCO: data cut-off; HRD: homologous recombination deficiency; RECIST: response evaluation criteria in solid tumours Source: PAOLA-1 HRD-positive subgroup data.[74]

Health-related quality of life (22 March 2019 DCO); HRD-positive group

Compliance rates were high for both EORTC QLQ-C30 and EORTC QLQ-OV28 instruments (xxxx in both arms; FAS); patients missing data / visits were well-balanced. EORTC QLQ-C30 and

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

EORTC QLQ-OV28 data for the FAS are presented in the CSR (Section 11.1.3); summary results for the HRD-positive group (EORTC QLQ-C30) are shown below.

EORTC QLQ-C30

EORTC QLQ-C30 scores range from 0 to 100, with higher scores in global health status/QoL and functional scales indicating better HRQoL.[73] A clinically-meaningful change was pre-specified as requiring a 10-point difference in adjusted means.[73] HRQoL remained stable across the 24-month treatment period (until “EoT” in Figure 29 below) in both olaparib + bevacizumab and placebo + bevacizumab groups.[74] No clinically meaningful changes from baseline in HRQoL global health status/QoL score were observed across timepoints in either treatment arm.[74] Similar results were also observed in the following EORTC QLQ-C30 functional scales: role functioning (Figure 30), physical functioning (data not shown), emotional functioning (Figure 31), and social functioning (Figure 32). Collectively, these data show that the addition of olaparib to bevacizumab does not negatively impact on the HRQoL of patients and are consistent with the manageable safety profile of olaparib + bevacizumab treatment (discussed in Section B.2.10).

Global health/QoL scores as well as role-, social-, and emotional-functioning scores also remained stable in the olaparib + bevacizumab group in the follow-up period (although these data should be interpreted with caution given small sample sizes).[74] EORTC QLQ-C30 summary data in the HRDpositive group were consistent with that in the FAS, confirming its robustness.

Figure 29. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group: Global health status/QoL change from baseline (22 March 2019 DCO), HRDpositive population

==> picture [375 x 232] intentionally omitted <==

Abbreviations : EoT: end of treatment; EORTC: European Organisation for the Research and Treatment of Cancer; FUP: follow-up; HRD: homologous recombination deficiency; QLQ-C30: Quality of Life Questionnaire for Cancer Patients (Core 30 item module); QoL: quality of life; SD: standard deviation. Source: PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 30. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group, EORTC-QLQ-C30 functional scale – role functioning; change from baseline (22 March 2019 DCO), HRD-positive population

==> picture [380 x 242] intentionally omitted <==

Abbreviations : EoT: end of treatment; EORTC: European Organisation for the Research and Treatment of Cancer; FUP: follow-up; HRD: homologous recombination deficiency; QLQ-C30: Quality of Life Questionnaire for Cancer Patients (Core 30 item module); QoL: quality of life; SD: standard deviation. Source: PAOLA-1 HRD-positive subgroup data.[74]

Figure 31. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group, EORTC-QLQ-C30 functional scale – emotional functioning; change from baseline (22 March 2019 DCO), HRD-positive population

==> picture [409 x 254] intentionally omitted <==

Abbreviations : EoT: end of treatment; EORTC: European Organisation for the Research and Treatment of Cancer; FUP: follow-up; HRD: homologous recombination deficiency; QLQ-C30: Quality of Life Questionnaire for Cancer Patients (Core 30 item module); QoL: quality of life; SD: standard deviation. Source: PAOLA-1 HRD-positive subgroup data.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 32. Mean (±SD) EORTC QLQ-C30 scores change from baseline across time points, by treatment group, EORTC-QLQ-C30 functional scale – social functioning; change from baseline (22 March 2019 DCO), HRD-positive population

==> picture [444 x 275] intentionally omitted <==

Abbreviations : EoT: end of treatment; EORTC: European Organisation for the Research and Treatment of Cancer; FUP: follow-up; HRD: homologous recombination deficiency; QLQ-C30: Quality of Life Questionnaire for Cancer Patients (Core 30 item module); QoL: quality of life; SD: standard deviation. Source: PAOLA-1 HRD-positive subgroup data.[74]

EQ-5D-5L

The impact of treatment and disease state on health state utility as assessed by the EQ-5D-5L was a secondary variable in this study.[73] The compliance rates for the planned on-treatment visits of EQ-5D-5L were high xxxxxx in both arms from baseline to Week 96, reflecting the protocoldefined treatment cap of two years on olaparib.[73]

The weighted health state index score showed no worsening / deterioration in patients who received olaparib + bevacizumab versus those treated with placebo + bevacizumab, in both the FAS (see CSR Section 11.1.3.2) and in the HRD-positive population (Figure 33). The EQ-5D-5L analyses were used in the cost-effectiveness model and are described in further detail in Section B.3.4.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 33. Mean (± SD) EQ-5D-5L weighted health state index change from baseline across time points by treatment group (22 March 2019 DCO), HRD-positive population

==> picture [452 x 289] intentionally omitted <==

Abbreviations : EoT: end of treatment; EQ-5D-5L: EuroQoL five dimensions, five level; FUP: follow-up; HRD: homologous recombination deficiency; QoL: quality of life; SD: standard deviation. Source: PAOLA-1 HRD-positive subgroup data.[74]

B.2.7 Subgroup analysis

A summary of PFS data in the pre-specified subgroups of the PAOLA-1 FAS , based on stratification factors (first-line treatment outcome and t BRCA m status), clinical characteristics, and biomarker subgroups, is provided in Appendix E and Section 11.5 of the PAOLA-1 CSR[73]

Briefly:

• A PFS benefit with olaparib versus placebo, when added to bevacizumab, was observed regardless of first-line treatment outcome and t BRCA m or t BRCA wt status (Figure 5, Appendix E).

• Clinically meaningful reductions in the risk of disease progression or death (ranging from xxxxxxxxxx) were also observed for olaparib versus placebo, when added to bevacizumab, across all pre-specified clinical characteristics, including patient age and disease stage (Figure 5, Appendix E).

• As highlight previously, pre-planned post-randomisation testing (using the Myriad myChoice[®] HRD Plus test), revealed a benefit with olaparib versus placebo, when added to bevacizumab, in the HRD-positive subgroup (HR=0.33, 95% CI: 0.25, 0.45). Importantly, a similar benefit was also observed in the HRD-positive subgroup excluding Myriad t BRCA m patients xxxxxxxxxxxxxxxxxxxxxxxxxxxx), indicating that the PFS benefit in the HRD-positive group was not driven entirely by the t BRCA m population. This is an important result that

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

demonstrates the efficacy of olaparib in the first-line maintenance setting in a broader population of women with HRD-positive tumours, regardless of whether they have mutations in the BRCA1/2 genes . The incremental benefit of olaparib + bevacizumab versus maintenance with olaparib monotherapy in the t BRCA m population is discussed in Section B.2.12.

B.2.8 Meta-analysis

The PAOLA-1 study is the only clinical trial that has evaluated the efficacy and safety of olaparib, when added to bevacizumab, in the population of interest for this appraisal; therefore, a metaanalysis of available evidence is not applicable to this appraisal.

B.2.9 Indirect and mixed treatment comparisons

The feasibility of performing an indirect or mixed treatment comparison between the PAOLA-1 study and those representing the comparators in the NICE scope was assessed, and reported in full in Appendix D.1.5.

Overall, due to the following differences in study design and populations between trials of ‘first-line followed by maintenance treatment or routine surveillance’ (e.g. GOG-0218) and ‘maintenance only’ treatments (e.g. PAOLA-1), it was not feasible to conduct an indirect comparison of the intervention and comparator treatments outlined in the scope:

• Design: the point of randomisation is different across the ‘first-line followed by maintenance treatment or routine surveillance’ studies and the ‘maintenance only’ studies, leading to misalignment in the types of interventions given.

  • For instance, the GOG-0218 study (which included first-line platinum-taxane chemotherapy with bevacizumab 15mg/kg followed by bevacizumab 15mg/kg maintenance treatment) randomised patients at the start of chemotherapy,[34] whilst the PAOLA-1 (maintenance only) study randomised patients at the point of completing first-line platinum-taxane chemotherapy with bevacizumab 15mg/kg.[1 ]

• Population: ‘maintenance only’ studies (such as PAOLA-1) include only those patients who have had either complete or partial response to chemotherapy,[1] while ‘first-line followed by maintenance treatment or routine surveillance’ studies (such as GOG-0218) include all patients (even those who have stable- or progressive-disease).[34]

These issues prohibit the use of both conventional network-based indirect comparison methods (e.g. NICE TSD2) due to the lack of common comparators across studies, and of populationadjusted methods (e.g. NICE TSD18) due to the lack of overlap in study populations.

In addition, within the maintenance only setting, no studies were identified to bridge between PAOLA-1 and other relevant maintenance only studies (e.g. SOLO-1)[43] . In the absence of a network of studies, we have performed a series of unanchored population adjusted indirect comparisons of PAOLA-1 versus other PARPi maintenance studies, including SOLO-1 and PRIMA. The results of these analyses are provided in Section B.2.12 and are used to give additional context to the results of the PAOLA-1 study. They are not presented within the evidence synthesis section of the submission because they include comparators outside the scope of the appraisal (e.g. olaparib and niraparib monotherapy), and only partially address the decision scope outlined by NICE (e.g. only the maintenance setting).

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.10 Adverse reactions

Safety data summarised in this section are derived from the full SAS, comprising all patients who received at least one treatment dose and had at least one safety follow-up assessment, regardless of their HRD status (further details on data analysis sets in PAOLA-1 are provided in Section B.2.4.1 and in Section 9.8.2 of the PAOLA-1 CSR[73] )

Safety and tolerability were assessed in terms of AEs (including SAEs), deaths, laboratory data, vital signs, ECGs and treatment exposure. All safety data are summarised by actual treatment arm, including patients who had dose reductions for the blinded period of study, and no formal statistics were performed. Safety results were analysed for both the overall study duration phase and the combination phase (Figure 34):

• The overall study duration phase was defined as time from initiation of olaparib or placebo treatment, including the 30 day follow-up after the last dose.

• The combination phase was defined as time from initiation of olaparib or placebo until the last dose of olaparib or placebo and bevacizumab given concurrently, plus 21 days.

Unless otherwise specified, discussions of safety data relate to the overall study duration, although the data for the shorter combination phase are also presented where relevant.

Figure 34. Safety analysis phases

==> picture [460 x 135] intentionally omitted <==

Source : PAOLA-1 CSR[73]

B.2.10.1 Treatment exposure

Treatment exposure to bevacizumab (SAS and HRD-positive population)

The median duration of bevacizumab treatment was similar in both olaparib + bevacizumab and placebo + bevacizumab arms (xxxx months and xxxx months, respectively; SAS and HRD-positive group), demonstrating that combination treatment with olaparib did not negatively impact on the administration of bevacizumab (Table 19). The median number of cycles of bevacizumab (excluding in the period prior to randomisation) was xx cycles and xx cycles in the olaparib + bevacizumab arm and placebo + bevacizumab arms, respectively.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Table 19. Duration of bevacizumab exposure (22 March 2019 DCO), SAS and HRD-positive population

If a patient was ongoing treatment, DCO was used to calculate duration. aTotal exposure = last infusion date - first infusion date + 21. Summary excludes prior bevacizumab infusions. bPre-randomisation cycles of bevacizumab include those given in combination with chemotherapy. cSummary excludes prior bevacizumab infusions which were summarised separately. One patient received olaparib within 21 days of their last prior bevacizumab infusion but did not receive a bevacizumab infusion after randomisation. Abbreviations: DCO: data cut-off; HRD: homologous recombination deficiency; SAS: safety analysis set; SD: standard deviation. Source: PAOLA-1 CSR;[73] PAOLA-1 HRD-positive subgroup data.[74]

Treatment exposure to olaparib or placebo (SAS and HRD-positive population)

For the overall study duration, the median duration of exposure to olaparib in the olaparib + bevacizumab arm and placebo in the placebo + bevacizumab arm was 17.3 months and 15.6 months, respectively in the SAS, consistent with the time to first progression and the two-year treatment cap for olaparib or placebo.[1] The median total duration of olaparib treatment was very similar to the actual duration of treatment (i.e. excluding dose interruptions) (Table 20).

The median total and actual duration of treatment with olaparib in the “combination phase” was comparable between the olaparib + bevacizumab and placebo + bevacizumab arms (xxxx and xxxx months, and xxxx and xxxx months, respectively), showing that combining olaparib with bevacizumab did not negatively impact upon the duration of olaparib dosing.[73]

Median duration of exposure to olaparib in the olaparib + bevacizumab arm and placebo in the placebo + bevacizumab arm of the HRD-positive group, was xxxx months and xxxx months, respectively, again consistent with the time to progression and the two-year treatment cap for olaparib or placebo.[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Table 20. Duration of olaparib or placebo exposure (22 March 2019 DCO), SAS and HRDpositive population

Dose interruptions include those where the patient forgot to take all doses on a given day. If patient was ongoing, data-cut-off has been used to calculate duration. aTotal treatment duration (months)=(last dose date-first dose date+1)/30.4375. Abbreviations: HRD: homologous recombination deficiency; SD: standard deviation. Source: PAOLA-1 CSR;[73] PAOLA-1 HRD-positive subgroup data.[74]

At the time of DCO, xxxxx of patients in the olaparib + bevacizumab arm and xxxxx of patients in the placebo + bevacizumab arm (HRD-positive population) had discontinued treatment. The median time to study treatment discontinuation or death (TDT) was xxxx months in the olaparib + bevacizumab arm (95% CI: xxxxxxxxxx months) and xxxx months in the placebo + olaparib arm (xxxxxxxxxx months). The KM-curves for both treatment arms (HRD-positive group) are shown in Figure 35.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 35. Time to treatment discontinuation or death (TDT; 22 March 2019 DCO), HRDpositive population

==> picture [409 x 265] intentionally omitted <==

Abbreviations : bd: twice daily; TDT: time to treatment discontinuation or death. Source: PAOLA-1 HRD-positive subgroup data.[74]

B.2.10.2 Dose interruptions and reductions

Toxicities in the PAOLA-1 study were managed either through dose interruptions or dose reductions (to 250 mg twice daily as a first step, and a further reduction to 200 mg twice daily, if needed); no dose escalations were permitted.[73] All reductions, interruptions or deviations from the protocol-defined dose of 300 mg twice daily, including single missed or forgotten doses, were captured as a dose reduction or dose interruption in the dosing eCRF.

Overall, more patients in the olaparib + bevacizumab arm had dose reductions, relative to the placebo + bevacizumab arm (xxxx% versus xxx%, respectively), however, just one reduction was required in the majority of cases (xxxxxxx reductions[***] ; olaparib + bevacizumab arm, SAS).[73] Most first dose reductions occurred within the first three months of treatment.

xxxx% of patients in the olaparib + bevacizumab arm had at least one dose interruption, versus xxxx% of patients in the placebo + bevacizumab arm. The majority of patients had just one or two dose interruptions (xxxxxxx and xxxxx events in the olaparib + bevacizumab and placebo + bevacizumab arms, respectively).

AEs were the most common cause of dose reductions and interruptions in both treatment arms and are further described below in Section 0.

***70 patients required two dose reductions, while two patients required three dose reductions (olaparib + bevacizumab arm; SAS).

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Dose interruptions and reductions were not analysed separately in the HRD-positive group, since there were no reasons to suspect any underlying differences from the SAS. Treatment exposure and safety profiles in the HRD-positive were as expected and reflective of the PAOLA-1 SAS.

B.2.10.3 Summary of AEs (SAS and HRD-positive population)

Overall, olaparib + bevacizumab was well-tolerated and had a manageable safety profile relative to placebo + bevacizumab. At the first DCO, most patients in both treatment arms had experienced at least one AE (Table 21).[73] The majority of AEs were non-serious and did not necessitate discontinuation of study treatment. Grade ≥3 AEs were reported in xxxx% of patients in the olaparib + bevacizumab arm and xxxx% of patients in the placebo + bevacizumab arm in the overall study period (SAS). The proportions of patients reporting serious adverse events (SAEs) was similar between treatment arms. There were five fatal AEs in total; one in the olaparib-treated arm and four in the placebo-treated arm (SAS).[1]

An overview of common AEs, CTCAE Grade ≥3 AEs, SAEs, and AEs leading to discontinuation of study treatment or death is provided in the sections below for the SAS. Overall, the safety profile of the olaparib + bevacizumab treatment arm was consistent with previous trials of each drug; the combination treatment did not impact on the tolerability of either bevacizumab or olaparib.[1]

A summary of key safety analyses in the HRD-positive population are also shown in Table 21 (alongside data for the SAS) and highlight no meaningful differences in the two datasets . This is as expected, since underlying biomarker status is not expected to impact upon patient’s tolerability of study treatments.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Table 21. Summary of adverse events (22 March 2019 DCO), SAS and HRD-positive population

Dose interruptions, reductions and discontinuations reported are from olaparib and placebo. Abbreviations: AEs: adverse events; SAEs: serious adverse events. Source: PAOLA-1 CSR;[73] Ray-Coquard et al., 2019;[1] HRD-positive subgroup data[74]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652]

© AstraZeneca (2020). All rights reserved

Common adverse events (SAS)

The majority of patients in both treatment arms had experienced ≥1 AE by the time of the first DCO in the overall study period: 531 of 535 patients (99.3%) in the olaparib + bevacizumab arm and 256 of 267 patients (95.9%) in the placebo + bevacizumab arm).[1] xxx (xxxx%) and xxx (xxxx%) patients in olaparib + bevacizumab and placebo + bevacizumab arms, respectively, experienced AEs that were deemed by the Investigator as being causally related to the study treatment.[73] The most commonly occurring AEs (occurring in ≥10% of patients in either arm) are summarised in Table 22.

The most common AE experienced in the olaparib + bevacizumab arm (overall study period) was nausea (285/535 patients [53.3%]). The vast majority of these events (272 of 285) were of low grade (<Grade 3) and could be resolved with antiemetic therapy.[1] All of the events that were reported at a frequency of ≥10% in the olaparib + bevacizumab arm and also occurred at more than a 5%-point greater frequency in the olaparib + bevacizumab arm than the placebo + bevacizumab arm, were known adverse drug reactions (ADRs) for olaparib and included nausea, fatigue, anaemia, lymphopenia, vomiting and leukopenia.

The most common AE in the placebo + bevacizumab arm was hypertension (160/267 patients [59.9%]) (overall study duration; Table 22).[1] Hypertension and proteinuria AEs were reported at a ≥5%-point greater frequency in the placebo + bevacizumab arm than the olaparib + bevacizumab arm; both are listed as ADRs for bevacizumab.

The majority of AEs first occurred within the first 28 days of treatment (xxxxxxx patients [xxxxx] in the olaparib + bevacizumab arm and xxxxxxx patients [xxxxx] in the placebo + bevacizumab arm.[73] The frequencies of commonly-reported AEs in the combination phase are also provided in Table 22 for completeness and are consistent with the data for the overall study period.

Table 22. Most common AEs (all grades), occurring in ≥10% of patients in either treatment arm (SAS)

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

aPreferred term, MedDRA Version 22.0. bIncludes AEs with an onset date on or after the date of first dose and up to and including 30 days following the date of last dose of olaparib or placebo. Abbreviations : AEs: adverse events; MedDRA: Medical Dictionary for Regulatory Activities. Source: Ray-Coquard et al., 2019;[1] Ray-Coquard et al., 2019 Supplementary Appendices;[94] PAOLA-1 CSR.[73]

CTCAE Grade ≥3 AEs (SAS)

Grade ≥3 AEs were reported in xxxx% of olaparib + bevacizumab-treated patients and xxxx% of placebo + bevacizumab treated patients (overall study period; Table 23).

Hypertension (xxxx%), anaemia (xxxx%), lymphopenia (xxx%) and fatigue (xxx%) were the only AEs of Grade ≥3 reported in ≥5% of patients in the olaparib + bevacizumab arm. Hypertension (xxxx%) was the only AE of Grade ≥3 reported in ≥5% of patients in the placebo + bevacizumab arm. All AEs of Grade ≥3 reported in ≥2% of patients dosed with olaparib + bevacizumab are known ADRs for these interventions.

A high proportion of AEs of Grade ≥3 AEs occurred during the combination phase and are captured in Table 23 for completeness.

Table 23. AEs of CTCAE Grade 3 or higher, >1% in either treatment arm (SAS)

==> picture [452 x 316] intentionally omitted <==

----- Start of picture text -----
Overall study duration Combination phase only
Olaparib + Placebo + Olaparib + Placebo +
System organ class
bevacizuma bevacizumab bevacizumab bevacizumab
MedDRA preferred term
b (N=535) (N=267) (N=534) (N=267)
n (%) n (%) n (%) n (%)
Patients with AE CTCAE xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx
Grade ≥3 [a]
Blood and lymphatic xxxxxxxxxx xxxxxxxx xxxxxxxxxx xxxxxxx
system disorders
Anaemia xxxxxxxxx xxxxxxx xxxxxxxxx xxxxxxx
Lymphopenia xxxxxxxx xxxxxxx xxxxxxxx xxxxxxx
Neutropenia xxxxxxxx xxxxxxx xxxxxxxx xxxxxxx
Leukopenia xxxxxxx xxxxxxx xxxxxxx xxxxxxx
Thrombocytopenia xxxxxxx xxxxxxx xxxxxxx x
Vascular disorders xxxxxxxxxx xxxxxxxxx xxxxxxxxx xxxxxxxxx
Hypertension xxxxxxxxxx xxxxxxxxx xxxxxxxxx xxxxxxxxx
Gastrointestinal disorders xxxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx
Nausea xxxxxxxx xxxxxxx xxxxxxxx x
----- End of picture text -----

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

AEs Grade ≥3 for overall study duration, includes AEs affecting >1% of patients in either treatment arm.[a] Patients with multiple AEs of Grade ≥3 are counted once for each system organ class/preferred term. Includes AEs with an onset date on or after the date of the first dose and up to and including 30 days following the date of last dose of olaparib or placebo. CTCAE Version 5.0, MedDRA Version 22.0.

Abbreviations: AE: adverse event; CTCAE: Common Terminology Criteria for Adverse Events; MedDRA: Medical Dictionary for Regulatory Activities. Source: PAOLA-1 CSR.[73]

Serious AEs (SAEs; SAS)

Similar frequencies of SAEs were reported in the olaparib + bevacizumab and placebo + bevacizumab arms (31.2% and 31.1%, respectively; overall study period (Table 24).[1, 73] Hypertension was the most commonly-reported SAE, with a similar incidence between the two study arms (48 patients [9.0%] in the olaparib + bevacizumab arm and 35 patients [13.1%] in the placebo + bevacizumab arm).[1]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

In the olaparib + bevacizumab arm, xxx patients (xxxx%) experienced SAEs in the combination phase, compared with xx patients (xxxx%) in the placebo + bevacizumab arm (further details in Table 24).[73]

Table 24. Summary of SAEs (SAS)

aPreferred term, MedDRA Version 22.0. SAEs for overall study duration, includes SAEs affecting >1% of patients in either treatment arm. Patients with multiple SAEs are counted once for each system organ class/preferred term. Includes SAEs with an onset date on or after the date of the first dose and up to and including 30 days following the date of last dose of olaparib or placebo. MedDRA Version 22.0. Abbreviations: MedDRA: Medical Dictionary for Regulatory Activities; SAE: serious adverse event. Source: PAOLA-1 CSR;[73] Ray-Coquard et al., 2019;[1] Ray-Coquard et al., 2019 Supplementary Appendices.[94]

Adverse events leading to discontinuation of study treatment, dose reductions, or dose interruptions (SAS)

AEs leading to discontinuation of study treatment (olaparib or placebo) were reported in 109 (20.4%) patients in the olaparib + bevacizumab arm and 15 (5.6%) patients in the placebo + bevacizumab arm.[1] AEs leading to discontinuation of treatment in ≥2 patients are presented in Table 57 (page192) of the CSR.[73] The most common AEs (reported in ≥2% of patients) leading to discontinuation of olaparib were anaemia (19 [3.6%]) and nausea (18 [3.4%]) (overall study period).[1, 94] The most common AEs (reported in ≥0.5% of patients) leading to discontinuation of placebo were dyspnoea (xxxxxxxx) and myocardial infarction (2 [0.7%]) (overall study period).[73, 94] The majority of AEs leading to discontinuation of olaparib or placebo occurred during the combination phase (reported in xxxxx of patients in the olaparib + bevacizumab arm and xxxx of patients in the placebo + bevacizumab arm).[73]

Overall, AEs leading to olaparib or placebo dose reductions occurred in 220 (41.1%) patients in the olaparib + bevacizumab arm and 20 (7.5%) patients in the placebo + bevacizumab arm.[1] The

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

most common AEs leading to dose reduction of olaparib (in ≥5% of patients) were anaemia (xxxxxxxxxxxxxxxxxx) and nausea (xxxxxxxxxxxxxxxxx).[73] Diarrhoea was the most common AE leading to dose reduction of placebo (xxxxxxxxxxxxxxxxxxx). xxxxx of AEs leading to dose reductions in olaparib + bevacizumab-treated patients occurred in the combination phase, compared with xxxx of AEs leading to dose reductions in placebo + bevacizumab-treated patients.[73]

AEs leading to olaparib or placebo dose interruptions occurred in 54.4% of patients in the olaparib + bevacizumab arm, and 24.3% of patients in the placebo + bevacizumab arm.[1] The most common AEs leading to dose interruption of olaparib (in ≥5% of patients) were anaemia (xxxxxxxxxxxxxxxxxxx) and nausea (xxxxxxxxxxxxxxxx).[73] Headache (xxxxxxxxxxxxxx xxxx), diarrhoea and nausea (xxxxxxxxxxxxxx xxxx) were the most common AEs leading to dose interruption of placebo.

xxxxx of AEs leading to dose interruptions in olaparib + bevacizumab-treated patients occurred in the combination phase, compared with xxxxx of AEs leading to dose reductions in placebo + bevacizumab-treated patients.[73] The AEs leading to treatment interruption of olaparib were generally consistent with the known safety profile of olaparib.

AEs of special interest (SAS)

AEs of special interest for olaparib

AEs of special interest for olaparib are summarised in Table 25.

Myelodysplastic syndrome (MDS), acute myeloid leukaemia (AML) and aplastic anaemia (AA) were reported for six patients (1.1%) who received olaparib + bevacizumab and one patient (0.4%) who received placebo + bevacizumab, based on long-term collection of data beyond treatment discontinuation and 30-day follow-up. This demonstrates no evidence of an association of MDS/AML/AA with olaparib treatment, in line with previous studies.[1, 94]

New primary malignancies were reported in seven patients (1.3%) in the olaparib + bevacizumab arm and three patients (1.1%) in the placebo + bevacizumab arm.[1, 94]

All events of pneumonitis (two patients), interstitial lung disease (three patients) and bronchiolitis (one patient) occurred in the olaparib + bevacizumab arm.[1, 94]

Table 25. AEs of special interest for olaparib (SAS)

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Abbreviations: AA: aplastic anaemia; AE: adverse event; AML: acute myeloid leukaemia; ILD: interstitial lung disease; MDS: myelodysplastic syndrome.

Source: Ray-Coquard et al., 2019;[1] Ray-Coquard et al., 2019 Supplementary Appendices.[94]

AEs of special interest for bevacizumab

AEs of special interest which are known to be associated with bevacizumab treatment are shown in Table 26. Patients receiving olaparib + bevacizumab had a similar or lower incidence of bevacizumab ADRs than patients receiving placebo + bevacizumab.[73] In particular, Grade ≥3 hypertension was reported in 30.3% of patients in the placebo + bevacizumab arm, compared with 18.7% of patients in the olaparib + bevacizumab arm.[94]

Blood pressure data xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx of hypertension in olaparib-treated patients compared with placebo-treated patients, suggesting that olaparib may have a xxxxxxxxxxxxxxxxxxxxxxxxx the hypertension known to be associated with bevacizumab treatment.[73] In the placebo + bevacizumab arm, xxxxx of patients had a shift in systolic blood pressure from normal to high, and xxxxx had a shift in diastolic blood pressure form normal to high. In comparison, xxxxx and xxxxx of patients in the olaparib + bevacizumab arm had a shift in systolic and diastolic blood pressure respectively, from normal to high.[73]

Table 26. Bevacizumab ADRs in either treatment arm (SAS)

aIncludes multiple MedDRA preferred terms. bPatients with multiple events in a category are only counted once in that category. Patients with events in more than one category were counted once in each of those categories. CTCAE Version 4.0, MedDRA Version 22.0.

Abbreviations: ADRs: adverse drug reactions; AE: adverse event; CTCAE: Common Terminology Criteria for Adverse Events; GI: gastrointestinal; MedDRA: Medical Dictionary for Regulatory Activities; PRES: posterior reversible encephalopathy syndrome.

Source: PAOLA-1 CSR;[73] Ray-Coquard et al., 2019 Supplementary Appendices.[94 ]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Deaths

Overall, xxx (xxxxx) patients treated with olaparib and xx (xxxxx) patients treated with placebo died during the study (Table 27).[73] The majority of deaths were due to ovarian cancer; deaths due to disease progression are not reported as AEs.

There were five fatal AEs (one in the olaparib + bevacizumab arm and four in the placebo + bevacizumab arm), which occurred during treatment or within the 30-day follow-up period.[1] A further xxxx fatal AEs occurred after the 30-day follow-up period (xxxxx in the olaparib + bevacizumab arm and xxx in the placebo + bevacizumab arm).[73] Further information on deaths related to AEs are provided in Section 12.3 of the CSR.

Table 27. All deaths in the PAOLA-1 study (SAS)

Deaths are reported for the FAS and patients are only reported in one category.[a] Death related to disease is determined by the investigator. Abbreviations: AE: adverse event; FAS: full analysis set. Source: PAOLA-1 CSR;[73] Ray-Coquard et al., 2019.[1]

B.2.11 Ongoing studies

Other than PAOLA-1, there are no ongoing studies relevant to the decision problem for this appraisal.

As per the PAOLA-1 protocol, a final OS analysis will be performed when the OS data are approximately 60% mature, or three years after the primary PFS analysis, whichever comes first.[73, ] 89 This analysis is anticipated in xxxxxxxxxx.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.12 Innovation

The last decade has marked a remarkable period of innovation in the treatment of advanced ovarian cancer, with PARPi at the centre of the practicechanging developments across all lines of treatment. Olaparib has been at the forefront of this innovation, with the first regulatory approvals (amongst PARPi) in both platinum-sensitive relapsed (SOLO-2 and Study 19) and first-line maintenance settings ( BRCA m only; SOLO-1; as illustrated in Figure 36).

Figure 36: EMA approvals for PARPi for the maintenance treatment of ovarian cancer

==> picture [474 x 34] intentionally omitted <==

----- Start of picture text -----
Abbreviations: 1L, first-line; 2L, second line; BRCA, breast cancer gene; HGSOC, high grade serous ovarian cancer;
PARP, poly-ADP ribose polymerase.
Source: Olaparib SmPC (and draft update); [4] Rucaparib SmPC; [95] Niraparib SmPC. [96]
----- End of picture text -----

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652]

© AstraZeneca (2020). All rights reserved

As summarised in Section B.1.3.2 (Figure 11), the CDF recommendation for olaparib monotherapy as maintenance treatment for women with newly-diagnosed BRCA1/2 -mutated advanced ovarian cancer has been practice-changing.[42] Data from the SOLO-1 study showed a 70% reduction in the risk of disease progression or death for olaparib maintenance versus placebo, and a minimum estimated three-year improvement in median PFS.[43] At the time, this was by far the greatest PFS benefit that had ever been observed in trials of first-line treatments for advanced ovarian cancer and may have been one of the largest improvements in PFS to have been reported in solid tumours.

Building on this, the Phase III PAOLA-1 study, the pivotal clinical trial relevant to this submission, evaluated efficacy and tolerability of adding olaparib to bevacizumab maintenance treatment, with the aim of further improving patient outcomes in this setting. In the HRD-positive patient population (which is broader than the SOLO1 population and includes women with BRCA wt status), the addition of olaparib to bevacizumab reduced the risk of disease progression or death by 67% versus an active comparator arm of bevacizumab given with placebo (HR = 0.33, 95% CI: 0.25, 0.45).

Unanchored population-adjusted indirect comparison (PAIC) performed using individual patient data (IPD) on investigator-assessed PFS (per RECIST version 1.1) from the SOLO-1 study (olaparib versus placebo in BRCA m patients) pooled with the BRCA m subset of patients from the PAOLA-1 study (olaparib + bevacizumab versus placebo + bevacizumab) showed that the addition of bevacizumab to olaparib was also associated with a meaningful improvement in PFS versus olaparib alone (HR=0.71; 95% CI: 0.45, 1.09) (Table 28). The adjusted KM-curves separated early in favour of olaparib + bevacizumab (versus olaparib monotherapy, or placebo + bevacizumab), and remained separated throughout the majority of the follow-up period (Figure 37).

Although subject to the inherent limitations of this type of analysis (further detail available on request), these data have important implications for treatment decision-making and suggest that the addition of olaparib to bevacizumab maintenance treatment has the potential to further improve treatment outcomes for women with newly-diagnosed BRCAm advanced ovarian cancer (relative to the already practice-changing SOLO-1 [olaparib monotherapy] maintenance regimen). These data will be presented at an upcoming Society of Gynaecologic Oncology Annual Meeting on Women’s Cancer (March 28[th] −30[th] , Toronto, Canada).

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Figure 37. Population-adjusted analysis: PFS outcomes for the weighted BRCA -mutated subset of PAOLA-1 and unweighted SOLO-1 (same as Figure 20)

==> picture [407 x 56] intentionally omitted <==

==> picture [350 x 216] intentionally omitted <==

==> picture [357 x 20] intentionally omitted <==

----- Start of picture text -----
Shaded area indicates 95% CI.
Abbreviations : BRCA : breast cancer susceptibility gene; PFS: progression-free survival.
----- End of picture text -----

Table 28. Results of the population-adjusted indirect comparison (PAIC): PFS outcomes for the weighted BRCA -mutated subset of PAOLA-1 and unweighted SOLO-1

Note:[a] Results based on weighted outcomes after matching tumour location status, ECOG status, FIGO stage, type of surgery (interval versus initial), residual disease status after surgery (yes or no), response to first-line treatment and age to SOLO1;[b] CIs generated via bootstrapping. Abbreviations: BRCA: breast cancer susceptibility gene; CI: confidence interval; ECOG: Eastern Cooperative Oncology Group; FIGO: International Federation of Gynaecology and Obstetrics; HR: hazard ratio; PAIC: population-adjusted indirect comparison; PFS: progression-free survival.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

The PAIC described above compared the relative efficacy of olaparib in combination with bevacizumab, versus olaparib monotherapy in newly-diagnosed patients with BRCA m advanced ovarian cancer, which is narrower than the population we are currently seeking reimbursement in. However, a separate analysis leveraging recent data from the Phase III PRIMA study40 show a similar result in HRD-positive patients.

The PRIMA study evaluated the efficacy and tolerability of the PARPi niraparib versus placebo in newly-diagnosed advanced ovarian cancer patients who had , responded to first-line chemotherapy.[40] Unlike PAOLA-1, which did not restrict inclusion by prior surgery / surgical outcomes, the PRIMA study only included those Stage III patients who had received NACT and IDS, or had visible residual tumour after PDS, or inoperable disease, in addition to patients with Stage IV disease.[40]

• The primary endpoint of the study was PFS in patients who were HRD-positive (using the same test / cut-off as in the PAOLA-1 study) and in the overall study population (which included HRDnegative patients), as determined on hierarchical testing. The PAIC was performed using: IPD from a subset of patients in PAOLA-1 (who met the staging and surgical inclusion criteria for PRIMA), and data digitised from published PRIMA PFS curves. The analysis showed that the addition of olaparib to bevacizumab maintenance treatment significantly improved PFS versus niraparib monotherapy in HRD-positive patients (

• Table 29). The adjusted KM-curves separated early in favour of olaparib + bevacizumab and remained separated throughout the follow-up duration (Figure 38). These data will have been submitted for presentation at an upcoming American Society of Clinical Oncology Annual Meeting (March 29[th] −June 2[nd] , Chicago, US). Further detail on methodology and results are available on request.

Importantly, both indirect treatment comparisons showed a significant PFS benefit for bevacizumab relative to placebo, supporting a role for bevacizumab maintenance monotherapy in this setting and the rationale for the PAOLA-1 study, i.e. maximising the extent of clinical benefit achieved through adding olaparib to bevacizumab maintenance therapy.

The PFS data from PAOLA-1 are supported by a series of clinically-relevant intermediate endpoints such as TFST, PFS2, and TSST, all of which show a consistent benefit in favour of olaparib added to bevacizumab (versus placebo + bevacizumab) (Table 7). These findings also provide confidence in the remarkable OS data, which (albeit immature) demonstrate a 45% reduction in the overall risk of death with the PAOLA-regimen versus placebo + bevacizumab.

Collectively, these data support the positioning of olaparib added to bevacizumab as a “new standard-of-care” for women with newly-diagnosed HRD-positive advanced ovarian cancer, who are in complete or partial response after first-line platinum-taxane chemotherapy and bevacizumab. The use of the PAOLA-1 regimen in this population of women is anticipated to be a highly cost-effective use of NHS resources (see Section 3.8, 3.13) , especially considering significant (anticipated) reduction in bevacizumab prices, following Avastin[®] LoE and multiple biosimilar entries (further discussed in Section B.3).

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

==> picture [455 x 23] intentionally omitted <==

----- Start of picture text -----
Figure 38. Pooled matching analysis of PRIMA and PAOLA-1 (HRD-positive patients) (same
as Figure 21)
----- End of picture text -----

==> picture [430 x 60] intentionally omitted <==

==> picture [371 x 247] intentionally omitted <==

Shaded area indicates 95% CI. Abbreviations : HRD: homologous recombination deficiency; PFS: progression-free survival.

Table 29. Results of the population-adjusted indirect comparison (PAIC): PAOLA-1 and PRIMA (HRD-positive)

a , Results from estimated IPD. Olaparib + bevacizumab and bevacizumab results from IPD after matching to PRIMA.

Abbreviations: CI: confidence interval; ESS: effective sample size; HR: hazard ratio; IPD: individual patient data; PAIC: population-adjusted indirect comparison; PFS: progression-free survival.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

B.2.13 Interpretation of clinical effectiveness and safety evidence

This appraisal requests a recommendation for the addition of olaparib to bevacizumab maintenance treatment in women with advanced (FIGO Stage III and IV) high-grade epithelial ovarian, fallopian tube, or primary peritoneal cancer who are in complete or partial response after first-line platinum-based chemotherapy with bevacizumab, and whose tumours are HRD-positive.

The clinical effectiveness evidence for olaparib in this indication is derived from the pivotal, randomised, double-blind, placebo-controlled, international, Phase III PAOLA-1 study. Results from the primary analysis of the PAOLA-1 study have demonstrated that the addition of olaparib to bevacizumab provides superior efficacy versus placebo + bevacizumab in the population of interest (as defined above), in conjunction with a manageable safety profile, and no detrimental impact on patients’ HRQoL. Key clinical efficacy and safety evidence from the PAOLA-1 study, including strengths and limitations of the evidence-base, and generalisability to the UK population of patients are briefly discussed below.

B.2.13.1 Principal findings from the evidence base

Clinical efficacy and HRQoL

The PAOLA-1 study met its primary endpoint of investigator assessed PFS, demonstrating a statistically significant and clinically meaningful benefit for olaparib added to bevacizumab maintenance treatment in the FAS (HR = 0.59; 95% CI, 0.49, 0.72; p<0.001).[1] The median duration of PFS achieved by adding olaparib to bevacizumab (22.1 months) is unprecedented in this treatment setting, in a population of women unselected by biomarker status or outcomes of prior surgical intervention.[1]

As explained in Figure 1, although the PAOLA-1 study was positive for the FAS see Section B.2.4.1), pre-planned subgroup analyses showed that women whose tumours were HRD-positive experienced a substantial benefit from the addition of olaparib to bevacizumab maintenance treatment:

• The median duration of PFS in the olaparib + bevacizumab arm was > 3 years, and over twice as long as that achieved with bevacizumab given with placebo.[1] Approximately 60% of women who received olaparib added to bevacizumab were progression-free at the 3-year landmark assessment of PFS, providing hope of long-remission or even cure in this group of patients .[1]

• The PFS data were supported by meaningful extensions in TFST xxxxxxxxxxxxxxxxxxxxxxx xxxxx), PFS2, and TSST.[74] These intermediate endpoints provide important insights into the long-term benefits of treatment (beyond disease progression) and reflect real-life treatment decisions and patient experience.

  • Meaningful extensions to PFS2 and TSST in favour of olaparib added to bevacizumab maintenance treatment demonstrates that the PAOLA-1 regimen does not negatively impact upon the efficacy of second-line treatments.

  • The longer duration of TFST and TSST represent extended periods free from cytotoxic chemotherapy, which negatively impacts upon patients’ HRQoL (adding to the significant physical and psychological burden of disease progression itself).[53, 54]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

• The addition of olaparib to bevacizumab maintenance treatment achieved a greater ORR than placebo + bevacizumab (xxxxx versus xxxxxx respectively, in those patients who had evidence of disease at randomisation); most patients achieved a CR (Table 18).[74] These data highlight an important benefit of olaparib beyond delaying disease progression, through reducing tumour volume to a greater extent than is possible with bevacizumab maintenance alone.

• Finally, the PFS and PFS2 benefit achieved from the olaparib to bevacizumab maintenance translates into a meaningful improvement in OS. Treatment with olaparib + bevacizumab reduced the overall risk of death by xxx versus placebo + bevacizumab in the HRD-positive population[74] , with KM-curves (Figure 28) showing clear and sustained separation in favour olaparib + bevacizumab, despite:

  • Low maturity of data (xxxxx mature; xxxx% in the olaparib + bevacizumab arm, and xxxx% in the placebo + bevacizumab arm) and

  • Greater use of targeted therapies (such as PARPi) in subsequent lines of therapy in the placebo + bevacizumab arm, which will underestimate the true OS benefit achieved with the PAOLA-1 regimen.

Importantly, these remarkable clinical benefits were achieved with no detrimental impact on patients’ HRQoL from the addition of olaparib to bevacizumab maintenance treatment:

• No clinically meaningful differences in global health status/QoL scores were observed between olaparib + bevacizumab and placebo + bevacizumab groups during the 24-month treatment period.

• The EQ-5D-5D weighted health state index score showed no worsening or deterioration in patients who received olaparib + bevacizumab versus placebo + bevacizumab.

Safety and tolerability

The median duration of exposure to olaparib or placebo (in olaparib + bevacizumab and placebo + bevacizumab arms, respectively) was consistent with the two-year treatment cap and time to first progression. The median total duration of exposure to bevacizumab was similar between the two arms, indicating that the addition of olaparib did not affect patients’ ability to receive bevacizumab.

The high median relative dose intensity (>95%) showed most patients were able to take the full dose of olaparib.

The safety data from PAOLA-1 were consistent with the known safety profiles of olaparib

and bevacizumab. The most commonly-reported AEs in the olaparib + bevacizumab arm were known ADRs for olaparib or bevacizumab, such as:

• Nausea, fatigue, anaemia, lymphopenia, vomiting and leukopenia (ADRs for olaparib), and

• Hypertension and proteinuria (ADRs for bevacizumab).

Interestingly, incidences of hypertension and proteinuria were lower when olaparib was added to bevacizumab (Table 22). The exact reason for this is not known; although one hypothesis suggests that olaparib may have a protective effect on some cardiovascular AEs.[97]

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

Importantly, the majority of AEs were non-serious and did not necessitate discontinuation of study treatment. The proportions of patients reporting serious adverse events (SAEs) was similar between treatment arms. No new safety signals being identified.

Safety data in the HRD-positive population was consistent with the SAS, with no clinically meaningful differences in the different categories of AEs.

Overall, the safety analyses showed that the PAOLA-1 regimen was tolerable. This is further corroborated by patient reported outcome (PRO) data, which show that the addition of olaparib to bevacizumab maintenance treatment had no detrimental impact on patients’ HRQoL (relative to bevacizumab given with placebo). Taken in the context of the substantial and sustained efficacy of the regimen, these data support a favourable risk:benefit ratio for the addition of olaparib to standard-of-care bevacizumab maintenance treatment.

B.2.13.2 Strengths and limitations of the evidence base, and generalisability to

the UK

PAOLA-1 was a well-designed, multicentre, randomised, double-blind, placebo-controlled, Phase III, externally sponsored study (Table 6) that provided comparative evidence for the addition of olaparib to bevacizumab maintenance treatment, which was an established standard-of-care when the study was initiated.[1, 73] The study was designed in close collaboration with the academic community and conducted by ARCAGY Research on behalf of ENGOT and GCIG.[73]

PAOLA-1 was performed in line with the Declaration of Helsinki, applicable regulatory requirements, International Code of Harmonisation / Good Clinical Practice (ICH / GCP), and relevant ARCAGY, study-centre, and local guidelines.[73, 89] The study was approved by the independent Institutional Review Board / Independent Ethics Committee associated with each study centre. Quality of data was assured through monitoring of investigational sites, appropriate training for study personnel, and use of data management procedures.[73, 89] In addition, an independent data monitoring committee was created to assess the safety of the study on a regular basis.[73, 89]

The PAOLA-1 population can be considered broadly generalisable to the UK population of patients in terms of demographics, prior surgery / surgical outcomes, and chemotherapy:

• Disease stage: Approximately 70% and 30% of patients in PAOLA-1 had Stage III and IV ovarian cancer,[1] respectively – these proportions are broadly representative of the UK population of newly-diagnosed advanced ovarian cancer patients (~64% of whom have Stage III disease at the time of diagnosis) (data from the Ovarian Cancer Audit Feasibility Pilot).[6]

• Age: The median age of patients was ~60 years (with a range of 26 years to 87 years, across both treatment arms).[1] This is consistent with the average age of patients in previous studies that included mostly UK patients (such as ICON8; median age: 61−63 years; range: 53−68 years, across study arms)[36] and is representative of the real-world population of women are likely to be treated with bevacizumab and / or olaparib.

• Prior surgery: Approximately 50% of the patients enrolled into the PAOLA-1 study had undergone upfront / primary debulking surgery (with ~42% receiving NACT followed by interval (i.e. delayed) debulking surgery and the remaining 8% of patients not undergoing any

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

surgery).[1] This split is very similar to the patients enrolled onto the ICON8 study, which was conducted across 87 UK centres and included 1,397 UK patients[98] enrolled between 2011 and 2014 - in the overall ICON8 population, 47% of patients had undergone immediate debulking surgery, 50% delayed debulking surgery, and 3% had inoperable disease.[36]

Whilst there is substantial variation in surgery rates and the use of upfront versus interval debulking procedures at regional (or even at individual centre) levels, the ICON8 data can be considered broadly representative of UK practice (while National audit data on this metric are unavailable).

• Outcome of surgical procedure: The proportions of patients in PAOLA-1 who had no macroscopic residual disease following surgery (~65%)[1] was lower than the proportion reported in the ICON8 study (84%),[36] although the latter only reported this for the proportion of patients who underwent delayed debulking surgery. Other studies involving large numbers of UK patients (such as ICON7) have also reported broadly similar surgical outcomes as ICON8 (with no residual disease recorded for 74% of patients included).[99]

A higher proportion of patients with no residual disease in studies with high UK representation may be due to the fact all surgical procedures for ovarian cancer are conducted at specialist gynaecological oncology centres by specialist surgeons, supported by specialist MDTs. Since lack of macroscopic disease at baseline is associated with better prognoses in advanced ovarian cancer, the slightly lower proportion of women with no macroscopic residual disease in PAOLA-1 may mean that study outcomes are conservative relative to what could be potentially achieved in UK practice.

• First-line chemotherapy: The use of carboplatin and paclitaxel as first-line chemotherapy regimen is aligned to the marketing authorisation and real-world use of bevacizumab,[100] and consistent the standard-of-care specified in NICE and BGCS guidelines.[17, 18]

• Bevacizumab: The dosage of bevacizumab used in PAOLA-1 was aligned to the EMA Marketing Authorisation (i.e. 15mg/kg Q3W, for up to 15 months).[5] Although this dosage is different to the 7.5mg/kg Q3W for up to 12 months regimen that is currently used in England, this is unlikely to impact on the overall results given the similar efficacy of the two bevacizumab doses (as described in detail in Section B.1.3.3). The cost impact associated with treating patients with a higher bevacizumab dose, as well as more women receiving bevacizumab in combination with chemotherapy in order to be eligible to receive olaparib and bevacizumab maintenance therapy is presented in Section 3.11, and show the use olaparib + bevacizumab (15mg/kg) is cost-effective compared to both routine surveillance as well as the CDF recommended use of bevacizumab (7.5mg/kg). As discussed later, these estimates are conservative due to the simplifying assumptions made about the clinical benefit of routine surveillance and bevacizumab 7.5 mg/kg.

In this context, it is important to highlight that in the longer-term, bevacizumab will most-likely be used in routine commissioning at its EMA licensed dose and in the full population covered by its marketing authorisation, following on from Avastin[®] LoE in July 2020 and substantial price reductions due to multiple biosimilar entries.

Company evidence submission template for olaparib with bevacizumab for maintenance treatment of advanced ovarian, fallopian tube or primary peritoneal cancer [ID1652] © AstraZeneca (2020). All rights reserved

The primary endpoint of investigator-assessed PFS in the PAOLA-1 study is:

• Clinically-relevant,

• The Gynaecological Cancer Intergroup (GCIG) preferred endpoint for clinical trials conducted in this disease setting, and

• Directly referenced in the final scope / decision problem for this appraisal.[101]

PFS data from PAOLA-1 are also supported by clinically-relevant secondary endpoints of TFST, PFS2, and TSST, and OS, all of which show a meaningful benefit of olaparib added to bevacizumab maintenance treatment (versus placebo + bevacizumab) in the HRD-positive group.

Although the evidence base available from PAOLA-1 is robust and comprehensive for decisionmaking, the following factors are worth highlighting:

• Use of PARPi in subsequent lines of therapy: In relation to PFS2, TSST, and OS data, it is also worth noting that xxxxx of patients in the placebo + bevacizumab arm received treatment with a PARPi as a first subsequent therapy post discontinuation from the study treatment, versus xxxx of patients in the olaparib + bevacizumab arm. Greater use of PARPi therapies amongst patients in the placebo + bevacizumab will underestimate the true PFS2, TSST, and OS benefit achieved from the addition of olaparib to bevacizumab maintenance treatment, but is reflective of real-world treatment decisions and outcomes:

  • The use of PARPi for the maintenance treatment of relapsed, platinum-sensitive advanced ovarian cancer is the standard-of-care in England (with three different treatments already recommended by NICE in this setting [TA528, TA611, TA620]).[50, ] 55, 56

Note: the use of PARPi therapy in second- and subsequent-lines of treatment is not permitted in women who have already received prior PARPi treatment (i.e. as maintenance treatment after first-line chemotherapy; per NICE recommendation for TA528, TA611)[33, 50, 55]

• Data maturity: while the analyses from the 22 March 2019 DCO, upon which this submission is based, provide a robust and compelling body of evidence for decision-making, with consistent and significant clinical benefit demonstrated for key efficacy endpoints of PFS, TFST, PFS2, TSST, and OS , the majority of the dataset is still immature (as shown in Table 7). Median duration of TFST, PFS2, TSST, and OS have not been reached after >2 years of follow-up. The PAOLA-1 study is still ongoing for final analysis of PFS2 and OS – these analyses will provide important insights into the full magnitude of clinical benefit of olaparib, when added to bevacizumab (beyond what has already been demonstrated). Final analysis of OS is expected in xxxx.