Single Technology Appraisal

Dabrafenib with trametinib for treating advanced BRAF V600 mutationpositive non-small-cell lung cancer [ID3851]

Committee Papers

NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE

SINGLE TECHNOLOGY APPRAISAL

Dabrafenib with trametinib for treating advanced BRAF V600 mutationpositive non-small-cell lung cancer [ID3851]

Contents:

The following documents are made available to stakeholders:

Access the final scope and final stakeholder list on the NICE website.

1. Company submission from Novartis

2. Clarification questions and company responses

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

• a. Roy Castle Lung Cancer Foundation

4. External Assessment Report prepared by Centre for Reviews and Dissemination and Centre for Health Economics – York

5. External Assessment Report – factual accuracy check

6. Technical engagement response from company

7. Technical engagement responses and statements from experts:

• a. Alastair Greystoke. Senior Lecturer, Clinical expert nominated by Novartis

• b. Roy Castle Lung Cancer Foundation patient organisation interview responses

8. External Assessment Group critique of company response to technical engagement prepared by Centre for Reviews and Dissemination and Centre for Health Economics – York

• 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

Dabrafenib and trametinib for advanced non-small cell lung cancer with a BRAF V600 mutation

Company evidence submission

Document B

August 2022

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Instructions for companies

This is the template for submission of evidence to the National Institute for Health and Care Excellence (NICE) as part of the single technology appraisal (STA) process. Please note that the information requirements for submissions are summarised in this template; full details of the requirements for pharmaceuticals and devices are in the user guide.

This submission must not be longer than 150 pages, excluding appendices and the pages covered by this template. If it is too long it will not be accepted.

Companies making evidence submissions to NICE should also refer to the NICE guide to the methods of technology appraisal and the NICE guide to the processes of technology appraisal.

In this template any information that should be provided in an appendix is listed in a box.

Highlighting in the template (excluding the contents list)

Square brackets and grey highlighting are used in this template to indicate text that should be replaced with your own text or deleted. These are set up as form fields, so to replace the prompt text in [grey highlighting] with your own text, click anywhere within the highlighted text and type. Your text will overwrite the highlighted section.

To delete grey highlighted text, click anywhere within the text and press DELETE.

Grey highlighted text in the footer does not work as an automatic form field, but serves the same purpose – as prompt text to show where you need to fill in relevant details. Replace the text highlighted in [grey] in the header and footer with appropriate text. (To change the header and footer, double click over the header or footer text. Double click back in the main body text when you have finished.)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Contents

Contents ........................................................................................................................................... 3 Tables and figures............................................................................................................................ 5 Abbreviations ................................................................................................................................... 7 B.1 Decision problem, description of the technology and clinical care pathway ............................. 9 B.1.1 Decision problem ................................................................................................................ 9 B.1.2 Description of the technology being evaluated ................................................................ 18 B.1.3 Health condition and position of the technology in the treatment pathway ..................... 21 B.1.3.1 Overview of the disease ............................................................................................ 22 B.1.3.2 Epidemiology of BRAF V600 mutation in NSCLC .................................................... 22 B.1.3.3 Diagnosis of NSCLC and molecular testing ............................................................. 23 B.1.3.4 Burden of disease ..................................................................................................... 23 B.1.3.5 Clinical pathway of care ............................................................................................ 24 B.1.4 Equality considerations .................................................................................................... 27 B.2 Clinical effectiveness ............................................................................................................... 28 B.2.1 Identification and selection of relevant studies ................................................................ 29 B.2.2 List of relevant clinical effectiveness evidence (previously untreated patients) .............. 32 B.2.2.1 Dabrafenib and trametinib ......................................................................................... 32 B.2.2.2 Pembrolizumab plus chemotherapy ......................................................................... 33 B.2.3 List of relevant clinical effectiveness evidence (previously treated patients) .................. 33 B.2.4 Summary of methodology of the relevant clinical effectiveness evidence ...................... 34 B.2.4.1 Dabrafenib and trametinib evidence: BRF113928 trial ............................................. 34 B.2.4.2 Study objectives ........................................................................................................ 35 B.2.4.3 Baseline characteristics ............................................................................................ 36 B.2.5 Statistical analysis and definitions of analysis sets .......................................................... 38 B.2.6 Quality assessment .......................................................................................................... 38 B.2.7 Clinical effectiveness results ............................................................................................ 38 B.2.7.1 Overall response rate ................................................................................................ 38 B.2.7.2 Duration of response ................................................................................................. 39 B.2.7.3 Progression-free survival .......................................................................................... 40 B.2.7.4 Overall survival .......................................................................................................... 41 B.2.8 Subgroup analysis ............................................................................................................ 41 B.2.9 Indirect comparisons ........................................................................................................ 42 B.2.9.1 Summary of relevant comparative effectiveness evidence ...................................... 42 B.2.9.2 Melosky et al. (2021) ................................................................................................. 43 B.2.9.3 FLATIRON RWE study (2022) .................................................................................. 43 B.2.10 Adverse reactions: BRF113928 trial .............................................................................. 52 B.2.10.1 Overview of adverse events .................................................................................... 52 B.2.10.2 Adverse events leading to treatment discontinuation ............................................. 53 B.2.10.3 Adverse events by preferred term for all patients ................................................... 53 B.2.10.4 Serious adverse events .......................................................................................... 55 B.2.10.5 Deaths ..................................................................................................................... 56 B.2.10.6 Safety summary ...................................................................................................... 57 B.2.11 Ongoing studies ............................................................................................................. 58 B.2.12 Interpretation of clinical effectiveness and safety evidence ........................................... 58 B.2.12.1 Principle findings from the clinical evidence base .................................................. 58 B.2.12.2 Strengths and limitations of the clinical evidence base .......................................... 58 B.3 Cost effectiveness ................................................................................................................... 60

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3.1 Published cost-effectiveness studies ............................................................................... 61 B.3.2 Economic analysis ............................................................................................................ 61 B.3.2.1 Model structure ......................................................................................................... 62 B.3.2.2 Key features of the economic analysis ..................................................................... 63 B.3.2.3 Patient population ..................................................................................................... 64 B.3.2.4 Intervention technology and comparators................................................................. 64 B.3.3 Clinical parameters and variables .................................................................................... 66 B.3.3.1 Summary of clinical data used in the model ............................................................. 66 B.3.3.2 Implementation of survival data ................................................................................ 67 B.3.3.3 Adverse events ......................................................................................................... 75 B.3.4 Measurement and valuation health effects ...................................................................... 77 B.3.4.1 Health-related quality-of-life data from clinical trials ................................................. 77 B.3.4.2 Mapping ..................................................................................................................... 77 B.3.4.3 Health-related quality-of-life studies ......................................................................... 77 B.3.4.4 Adverse reactions ..................................................................................................... 77 B.3.4.5 Treatment administration disutility ............................................................................ 81 B.3.4.6 Age-adjusted disutility values .................................................................................... 82 B.3.4.7 Health-related quality-of-life data used in the cost-effectiveness analysis ............... 82 B.3.5 Cost and healthcare resource use identification, measurement and valuation ............... 83 B.3.5.1 Intervention and comparators’ costs and resource use ............................................ 83 B.3.5.2 Drug administration costs ......................................................................................... 90 B.3.5.3 Health state unit costs and resource use.................................................................. 94 B.3.5.4 Adverse reaction unit costs and resource use .......................................................... 95 B.3.5.5 Subsequent therapies cost and resource use .......................................................... 97 B.3.5.6 End-of-life costs and resource use ......................................................................... 100 B.3.6 Severity ........................................................................................................................... 100 B.3.7 Uncertainty ..................................................................................................................... 101 B.3.8 Summary of base case economic analysis inputs and assumptions ............................ 101 B.3.8.1 Summary of base case economic analysis inputs .................................................. 101 B.3.8.2 Assumptions ............................................................................................................ 107 B.3.9 Base case economic analysis results ............................................................................ 114 B.3.10 Exploring uncertainty .................................................................................................... 115 B.3.10.1 Probabilistic sensitivity analysis ............................................................................ 115 B.3.10.2 Deterministic sensitivity analysis .......................................................................... 118 B.3.10.3 Scenario analyses ................................................................................................. 119 B.3.11 Subgroup analysis ........................................................................................................ 123 B.3.12 Benefits not captured in the QALY calculation ............................................................ 123 B.3.13 Validation ...................................................................................................................... 124 B.3.13.1 Technical validation ............................................................................................... 124 B.3.13.2 Clinical validation .................................................................................................. 124 B.3.14 Interpretations and conclusions of economic evidence ............................................... 124 References ................................................................................................................................... 126

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Tables and figures

Table 1: Comparators suggested within the NICE final scope ...................................................... 10 Table 2: The decision problem ...................................................................................................... 12 Table 3: Technology being appraised............................................................................................ 18 Table 4: Overview of studies included in the clinical SLR (for previously untreated patients) ...... 30 Table 5: Relevant clinical trial effectiveness evidence for dabrafenib and trametinib ................... 32 Table 6: Outcomes measured in BRF113928 ............................................................................... 35 Table 7: Eligibility criteria for BRF113928 ..................................................................................... 36 Table 8: BRF113928 patient baseline characteristics for Cohort C .............................................. 36 Table 9: BRF113928 summary of ORR by IA for Cohort C........................................................... 38 Table 10: BRF113928 summary of DOR by IA for Cohort C ........................................................ 39 Table 11: BRF113928 summary of PFS by IA for Cohort C ......................................................... 40 Table 12: BRF113928 summary of OS for Cohort C ..................................................................... 41 Table 13: Summary of results of Melosky et al. (2021)[73] .............................................................. 43 Table 14: Baseline characteristics for patients in the FLATIRON RWE study (2022) (BRAF V600E mutation) ............................................................................................................................ 45 Table 15: PFS for patients receiving dabrafenib and trametinib (BRF113928 Cohort C) and patients receiving pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022) (BRAF V600E mutation) ................................................................................................................ 46 Table 16: OS for patients receiving dabrafenib and trametinib (BRF113928 Cohort C) and patients receiving pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022) (BRAF V600E mutation) ................................................................................................................ 48 Table 17: BRF113928 summary of AEs (combined safety population) ........................................ 52 Table 18: BRF113928 summary of AEs (≥1% in total) leading to treatment discontinuation by maximum grade for all patients (combined safety population) ...................................................... 53 Table 19: BRF113928 summary of AEs by preferred term for all patients (combined safety population) ..................................................................................................................................... 54 Table 20: BRF113928 summary of SAEs for all patients (combined safety population) .............. 55 Table 21: BRF113928 summary of deaths for all patients (combined safety population) ............ 57 Table 22: Key features of the economic analysis .......................................................................... 63 Table 23: Patient characteristics model inputs .............................................................................. 64 Table 24: Summary of the modelled pembrolizumab plus chemotherapy dosing regimen .......... 65 Table 25: Comparison of pembrolizumab plus chemotherapy PFS and OS from extrapolations of the FLATIRON RWE study (2022) weighted data versus KEYNOTE-189 ................................... 66 Table 26: Summary of the clinical parameters used in the economic model ................................ 68 Table 27: Summary of goodness-of-fit data for dabrafenib and trametinib PFS (BRF113928 Cohort C); standard parametric models ........................................................................................ 69 Table 28: Summary of goodness-of-fit data for dabrafenib and trametinib OS (BRF113928 Cohort C); standard parametric models .................................................................................................... 70 Table 29: Summary of goodness-of-fit data for dabrafenib and trametinib ToT (BRF113928 Cohort C); standard parametric models ........................................................................................ 72 Table 30: Summary of goodness-of-fit data for pembrolizumab plus chemotherapy ToT (BRAF V600E [weighted] population – FLATIRON RWE study [2022]); standard parametric models .... 74 Table 31: Summary of AEs included in the base case economic analysis ................................... 76 Table 32: Summary of AE disutility estimates included in the base case economic analysis ...... 78 Table 33: Summary of utility values used in the base case economic analysis............................ 83 Table 34: Drug unit and pack costs (dabrafenib and trametinib) .................................................. 86 Table 35: Drug unit and pack/vial costs (pembrolizumab plus chemotherapy) ............................ 86

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 36: Summary of drug acquisition costs for the treatments included in the base case economic analysis .......................................................................................................................... 86 Table 37: Unit costs of drug administration used in the base case economic analysis ................ 90 Table 38. Summary of drug administration costs for pembrolizumab plus chemotherapy included in the base case economic analysis (comparators) ...................................................................... 92 Table 39: Summary of the resource use cost per cycle ................................................................ 94 Table 40. Resource use unit costs ................................................................................................ 94 Table 41. Resource use frequencies ............................................................................................. 95 Table 42. Unit costs of the AEs included in the economic model ................................................. 95 Table 43: Subsequent treatments modelled upon entry to the progressed disease state ............ 98 Table 44: Calculated costs for each subsequent treatment .......................................................... 99 Table 45. Terminal care costs upon entering death state ........................................................... 100 Table 46: Summary features of QALY shortfall analysis ............................................................. 100 Table 47: Summary of health state benefits and utility values for QALY shortfall analysis ........ 100 Table 48: Summary of QALY shortfall analysis ........................................................................... 101 Table 49: Summary of base case economic analysis inputs ...................................................... 101 Table 50: Summary of assumptions adopted within the base case economic analysis ............. 108 Table 51: Base case deterministic cost-effectiveness results (comparator at list price)............. 114 Table 52: Base case probabilistic cost-effectiveness results (comparator at list price) .............. 115 Table 53: Summary of deterministic scenario analyses and results (comparator at list price)[a] . 119 Figure 1: Mechanism of action of dabrafenib and trametinib ........................................................ 18 Figure 2: Current treatment pathway for previously untreated advanced NSCLC in UK clinical practice ........................................................................................................................................... 25 Figure 3: Patient flow in Cohorts A, B and C in BRF113928 ......................................................... 35 Figure 4: BRF113928 KM curve for DOR by IA for Cohort C ........................................................ 39 Figure 5: BRF113928 KM curve for PFS by IA for Cohort C ......................................................... 40 Figure 6: BRF113928 KM curve for OS in Cohort C ..................................................................... 41 Figure 7: PFS KM curves – dabrafenib and trametinib (BRF113928 Cohort C) versus pembrolizumab plus chemotherapy (FLATIRON RWE study [2022], BRAF V600E mutation) .... 47 Figure 8: OS KM curves – dabrafenib and trametinib (BRF113928 Cohort C) versus pembrolizumab plus chemotherapy (FLATIRON RWE study [2022], BRAF V600E mutation) .... 50 Figure 9: Partitioned-survival model with three health states ........................................................ 62 Figure 10: Dabrafenib and trametinib PFS extrapolations up to ten years (BRF113928 Cohort C) ........................................................................................................................................................ 69 Figure 11: Dabrafenib and trametinib OS extrapolations up to ten years (BRF113928 Cohort C) ........................................................................................................................................................ 71 Figure 12: Dabrafenib and trametinib ToT extrapolations up to ten years (BRF113928 Cohort C) ........................................................................................................................................................ 73 Figure 13: Pembrolizumab plus chemotherapy ToT extrapolations up to three years (BRAF V600E [weighted] population – FLATIRON RWE Study [2022])* ................................................. 74 Figure 14: Base case PSA scatter plot (comparator at list price) ................................................ 116 Figure 15: Base case PSA cost-effectiveness acceptability curve (comparator at list price) ..... 116 Figure 16: PSA stability plot (comparator at list price) ................................................................ 117 Figure 17: Deterministic sensitivity analysis results (comparator at list price) ............................ 118

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

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

B.1.1 Decision problem

This submission demonstrates the cost-effectiveness of dabrafenib and trametinib as a treatment for patients with advanced non-small cell lung cancer (NSCLC) with a BRAF V600 mutation.

Patients and clinicians in England currently have access to dabrafenib and trametinib via National Health Service (NHS) England interim COVID-19 guidance, due to it being an oral treatment, removing the requirement for hospital visits, as well as reducing the risk of immunosuppression associated with standard of care treatments. This submission seeks to allow routine access to dabrafenib and trametinib at a cost-effective price, to ensure NHS England receive value for money for this clinically valued combination treatment.

The submission focusses on the full marketing authorisation for dabrafenib and trametinib in this indication which is: dabrafenib and trametinib for the treatment of adult patients with advanced NSCLC with a BRAF V600 mutation .[1]

This population is primarily those patients with previously untreated advanced NSCLC with a BRAF V600 mutation. This is based on data provided by NHS England, which suggests that **% of the initiations of dabrafenib and trametinib since August 2020 were in patients with previously untreated advanced NSCLC harbouring a BRAF V600 mutation, based on data from **** **** **

********* *** ****** ***** **** **** **** ********* ************[2]

This is aligned with the typical treatment pathway for NSCLC in England[3] , where actionable NSCLC mutations, such as BRAF V600, are routinely identified by Genomic Hubs,[4] and typically, patients identified with an actionable mutation will be offered a targeted treatment option.

Despite improvements being reported in turnaround times for routine testing of biomarkers in advanced NSCLC patients, insights from a market research survey conducted with pathology centres, as well as clinical opinion given to Novartis, noted that delays in receiving testing results remain in some parts of the country. Some centres noted an average turnaround time of ** ** ** days for next generation sequencing (NGS) results to be returned (versus the recommended turnaround time of 10 working days).[4, 5]

As a result of these delays, the initial treatment for a small (but clinically important) minority of patients with previously untreated advanced NSCLC with a BRAF V600 mutation may continue to be pembrolizumab with chemotherapy. This would typically occur when a patient’s mutation status is not yet known to the treating physician, but delaying treatment initiation would be detrimental to the patient.

As such, some patients with previously treated advanced NSCLC would be eligible for dabrafenib and trametinib. However, this population will likely diminish over time, as the turnaround times for testing continue to improve in line with the implementation of the Genomic Hubs strategy and a patient’s mutation status is increasingly known at the time of the first treatment decision. This previously treated patient population of NSCLC exhibiting a BRAF V600 mutation presents a small (but clinically important) group within the total advanced NSCLC cohort. Based on the

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

availability of suitably robust evidence, cost-effectiveness analyses are presented only for the previously untreated population, although clinical data for previously treated patients are also provided in Appendix M.

In patients with previously untreated advanced NSCLC with a BRAF V600 mutation, pembrolizumab plus chemotherapy is considered to represent the standard of care, and so represents the principal comparator for this appraisal.

The National Institute for Health and Care Excellence (NICE) final scope lists all possible treatment options for advanced NSCLC, including treatments available by histology (nonsquamous and squamous), programmed death-ligand 1 (PD-L1) expression, and line of treatment. Of these, pembrolizumab plus chemotherapy (specifically, pembrolizumab plus pemetrexed and platinum-based chemotherapy [carboplatin or cisplatin]), reflects the current standard of care for patients with previously untreated advanced NSCLC with a BRAF V600 mutation in the NICE treatment pathway and UK clinical practice.[3]

As discussed with the NICE and External Assessment Group (EAG) teams during the Decision Problem stage, this is aligned with recent NICE appraisals assessing treatments for an actionable mutation in patients with previously untreated advanced NSCLC, where the NICE Committee and Cancer Drugs Fund (CDF) clinical lead noted that immunotherapy with chemotherapy represents the standard of care in the UK for patients with previously untreated advanced NSCLC (NICE TA789 [tepotinib for treating advanced NSCLC with MET gene alterations], NICE TA781 [sotorasib for previously treated KRAS G12C mutation-positive advanced NSCLC]).[6, 7]

Novartis sought to verify the UK standard of care for patients with previously untreated advanced NSCLC with a BRAF V600 mutation via an advisory board with leading UK NSCLC clinical experts.[8] The clinicians explained that, in cases where dabrafenib and trametinib are not available, or in cases where BRAF status is unknown at the time of first treatment, the majority of patients would receive pembrolizumab plus chemotherapy.[3] Pembrolizumab plus chemotherapy is preferred to pembrolizumab monotherapy irrespective of PD-L1 expression, given the aggressive nature of the disease and thus prognostic importance of the BRAF V600 mutation.

Patients with previously treated advanced NSCLC with a BRAF V600 mutation would typically receive pembrolizumab plus chemotherapy as their first treatment (if dabrafenib and trametinib was not available), for the reasons outlined above. As such, chemotherapy (consisting of either docetaxel monotherapy, or docetaxel plus nintedanib) would represent the most relevant treatment option for these patients, if dabrafenib and trametinib were not available.

Table 1 provides full details of the comparators suggested by NICE in the final scope for this appraisal, with an explanation for their inclusion or exclusion in the current submission. The decision problem for this submission in summarised in Table 2.

Table 1: Comparators suggested within the NICE final scope

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations: NICE: National Institute for Health and Care Excellence; NSCLC: non-small cell lung cancer; PDL1: programmed death-ligand 1. Source: Clinical Advisory Board Summary (14[th] July 2022).[8]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 2: The decision problem

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 12 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 13 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 14 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 15 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 16 of 133

Abbreviations: HRQoL: health-related quality of life; NHS: National Health Service; NICE: National Institute for Health and Care Excellence; NSCLC: non-small cell lung cancer; OS: overall survival; PD-L1: programmed death-ligand 1; PFS: progression-free survival; UK: United Kingdom.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 17 of 133

B.1.2 Description of the technology being evaluated

A description of the technology being appraised is summarised in Table 3.

Table 3: Technology being appraised

UK approved name and Dabrafenib (Tafinlar®) and Trametinib (Mekinist®) brand name Mechanism of In NSCLC with a BRAF V600 mutation, the mutated form of BRAF plays a role action in the development of the cancer by allowing uncontrolled division of tumour cells. The mutated BRAF protein also switches on another protein, called MEK, which is also involved in stimulating cell division, and consequently promotes the development of the cancer by allowing uncontrolled division of the tumour cells. Most cancer cells with a BRAF V600 mutation display a great reliance on MEK activity for growth and survival, and thus are very sensitive to both selective BRAF and MEK inhibitors.[9]

Dabrafenib is an oral, selective inhibitor of BRAF; trametinib is an oral, selective inhibitor of MEK. By blocking the action of BRAF and MEK simultaneously, dabrafenib and trametinib help to slow down the growth and spread of the cancer (Figure 1). Pre-clinical studies have shown dabrafenib and trametinib to be highly efficacious BRAF and MEK inhibitors, respectively, validating their potential clinical benefits when used to treat patients with advanced NSCLC with a BRAF V600 mutation.[10]

Figure 1: Mechanism of action of dabrafenib and trametinib

==> picture [356 x 274] intentionally omitted <==

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

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

treatment pathway

Advanced NSCLC with a BRAF V600 mutation

• Dabrafenib and trametinib, inhibitors of the BRAF V600 mutation pathway, are a treatment combination for patients with advanced NSCLC with a BRAF V600 mutation.[11, 12] The BRAF V600 mutation is rare, found in approximately 1–3% of NSCLC cases,[18 ] equating to a patient population of approximately 66 to 100 patients in England each year.

• Advanced NSCLC is an incurable and debilitating disease, and patients face an extremely poor prognosis. Stage III and IV NSCLC are associated with five-year survival rates of 12.6% and 2.9%, respectively.[19]

Current clinical pathway of care for patients with advanced NSCLC

• In the UK, patients diagnosed with advanced NSCLC with an oncogenic driver mutation receive treatments which are specifically designed to target their mutation.[3] Patients with a BRAF V600 mutation are only able to access a targeted treatment, dabrafenib and trametinib, on a temporary basis via interim COVID-19 guidance.[20]

• Clinical experts noted that, where dabrafenib and trametinib are not available, patients with previously untreated advanced NSCLC with a BRAF V600 mutation currently receive pembrolizumab plus chemotherapy.[21, 22] Pembrolizumab plus chemotherapy is preferred to pembrolizumab monotherapy irrespective of PD-L1 expression, given the aggressive nature of the disease and thus prognostic importance of the BRAF V600 mutation. As such, pembrolizumab plus chemotherapy is the principal relevant comparator in this appraisal.

Dabrafenib and trametinib

• In response to the COVID-19 pandemic, dabrafenib and trametinib were made available to provide patients with an oral treatment option, to reduce the burden on the NHS due to patients attending hospital to receive IV treatment, as well as to reduce the risk of immunosuppression associated with standard of care treatments.[20]

• Dabrafenib and trametinib primarily represents a treatment for patients with previously untreated advanced NSCLC with a BRAF V600 mutation. Interim COVID-19 usage showed that **% of the patients who received dabrafenib and trametinib had previously untreated advanced NSCLC.[2] This is also aligned with the treatment pathway for other oncogenic driver mutations, where patients receive targeted treatment as early as possible.[3]

• Dabrafenib and trametinib will also represent an important treatment for patients with previously treated advanced NSCLC with a BRAF V600 mutation. However, this population represents a small (but clinically important) minority of patients eligible for dabrafenib and trametinib, and is expected to diminish over time, as turnaround times for testing continue to improve, and a patient’s mutation status is increasingly known at the time of initial treatment.

• • As oral treatments, dabrafenib and trametinib are associated with a reduced burden of administration versus pembrolizumab plus chemotherapy, which must be administered in hospital via intravenous (IV) infusion.[23] The availability of an oral treatment option has a positive impact on alleviating capacity issues within the NHS, while oral alternatives to IV therapies also represent an important preference for NSCLC patients.[24]

• Routine commissioning of dabrafenib and trametinib would establish a targeted treatment pathway for patients with advanced NSCLC with a BRAF V600 mutation, in line with the pathways for advanced NSCLC with other driver mutations.[3]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.1.3.1 Overview of the disease

Lung cancer

According to the National Lung Cancer Audit annual report, 39,653 cases of lung cancer were reported in England and Wales in 2018.[25] There are two main categories of lung cancer: smallcell lung cancer and non-small cell lung cancer (NSCLC). NSCLC accounts for the large majority of lung cancers in England and Wales (88%).[25]

Based on histology, NSCLC is typically divided into squamous cell carcinomas and nonsquamous carcinomas (which can be subdivided into adenocarcinoma and large cell carcinoma, as well as other rare types).[26] The approximate distributions of each NSCLC subtype, based on data from the Royal College of Physicians (RCP), National Lung Cancer Audit 2020, indicate that non-squamous NSCLC represents the predominant histology.[27] Of these, adenocarcinomas represent the largest subtype (adenocarcinoma 66%; large cell 2%; squamous 23%; other 8%).[28] Whilst the BRAF V600 mutation occurs predominantly in adenocarcinomas, dabrafenib and trametinib represents a treatment option for all advanced NSCLC patients with a BRAF V600 mutation, irrespective of histologic subtype, as per the licensed indication.[11, 12]

The stage at which lung cancer is diagnosed plays an important role in patient survival outcomes.[29] Lung cancer is often diagnosed at an advanced stage, due to a low index of suspicion surrounding the symptoms of NSCLC, or the presence of symptoms only at an advanced stage of disease.[30] In England, 22% and 49% of all lung cancer patients are diagnosed at stage III and IV, respectively.[25] Dabrafenib and trametinib are licensed for all advanced NSCLC patients, and therefore represents a treatment option for patients with advanced NSCLC with a BRAF V600 mutation.[11, 12]

B.1.3.2 Epidemiology of BRAF V600 mutation in NSCLC

The BRAF protein, encoded by the BRAF gene, forms an important part of the MAPK signalling pathway, which plays a critical role in the proliferation and survival of normal cells (Figure 1).[31] Oncogenic mutations to the BRAF gene (most commonly V600) result in constitutive activation of the MAPK pathway, leading to disease progression.[9] Most cancer cells with a BRAF V600 mutation display a reliance on MEK activity for growth and survival and thus are sensitive to selective BRAF and MEK inhibitors.[9, 32] This represents an important actionable component of the MAPK pathway and a therapeutic target which underscores the rationale for combination therapy with a BRAF inhibitor and a MEK inhibitor. By inhibiting BRAF and MEK, respectively, dabrafenib and trametinib target this oncogenic growth and survival pathway.

BRAF V600 refers to BRAF mutations in which valine (V) is substituted at amino acid 600.[33 ] BRAF V600 mutations are rare and seen in approximately 1–3% of all cases of NSCLC, predominantly in adenocarcinomas.[18 ] This equates to a patient population of approximately 66 to 100 patients in England each year.

The most common activating BRAF mutation is V600E (valine [V] substituted by glutamic acid [E]), representing approximately 50% or more of the BRAF mutations detected in NSCLC.[34-36] Dabrafenib and trametinib are licensed for adult patients with advanced NSCLC with a BRAF V600 mutation. Non-V600 BRAF mutations can also be found in NSCLC, but this is outside the scope of the licence and thus not considered in this submission. It should be noted that, of relevance to this submission, BRAF mutations are generally considered to be mutually exclusive to other oncogenic driver mutations.[37]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Finally, it should also be noted that the treatment pathway for patients with previously untreated advanced NSCLC is stratified by PD-L1 expression in UK clinical practice, to determine eligibility and likely response to treatment with immunotherapy.[3] While patients with advanced NSCLC with a BRAF V600 mutation may display PD-L1 expression (previous studies have shown that PD-L1 positivity in BRAF patients can range between 40–80%),[38-40] several retrospective analyses did not find a clear correlation between PD-L1 and BRAF mutations.[41-45] Furthermore, dabrafenib and trametinib functions as a targeted therapy, where its mode of action is independent of the of the PD-L1/PD-1 immune checkpoint. As such, dabrafenib and trametinib will represent a treatment option for all patients with a BRAF V600 mutation, regardless of PD-L1 expression.

B.1.3.3 Diagnosis of NSCLC and molecular testing

NICE Guideline 122 (NG122) outlines the diagnosis and management of lung cancer.[3] The approach outlined in the guideline covers recommendations for diagnosis, staging, rapid assessment of suspected cases, appropriate diagnostic tools such as sputum cytology, bronchoscopy, computed tomography (CT) and magnetic resonance imaging (MRI) scans and X- rays.

As part of a broader programme to improve lung cancer management, the NHS England Lung Clinical Expert Group established the National Optimal Lung Cancer Pathway in August 2017 to streamline the pathway to improve outcomes in lung cancer by encouraging best practice and reducing variation or delays in diagnosis, staging and treatment.[46]

The pathway allows for a maximum of 49 days for a patient to be diagnosed and treated. In cases where a molecular diagnosis could influence treatment, biopsy results providing information on actionable mutations should be available within 10 working days from the point of referral. To facilitate access to molecular diagnostics, NHS England have established seven centrally funded Genomic Laboratory Hubs, with a central testing directory of molecular biomarkers that require routine testing.[4]

The BRAF gene, alongside other oncogenic driver mutations, such as EGFR, ALK and KRAS p.G12C, is tested through an NGS panel as recommended in the National Genomics Testing Directory 2022 for advanced or metastatic solid tumours.[4] While the panel is usually for nonsquamous NSCLC, there may be scenarios where clinicians wish to test other subtypes (i.e., squamous) of NSCLC. UK clinical experts noted that while the Genomic Hubs are generally utilised, some parts of the country may rely on reflex or other forms of in-house testing to speed up the turnaround time for biomarker testing results.[8]

Given that BRAF V600 testing is confirmed and listed on the National Testing Directory, no additional tests, or associated costs, beyond those that are used in the routine diagnostic work up and management of patients with NSCLC, are anticipated to be required to determine eligibility for dabrafenib with trametinib.[4]

B.1.3.4 Burden of disease

Despite the recent UK initiatives to improve the diagnosis and management of lung cancer, the prognosis for patients diagnosed with lung cancer at an advanced stage remains poor.[29] According to data from Cancer Research UK from 2013 to 2017 (adapted from the Office for National Statistics), the one and five-year survival rates for patients diagnosed with Stage III lung cancer are 48.7% and 12.6%, respectively.[19] The prognosis for patients diagnosed with Stage IV

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

lung cancer is particularly poor: the one-year survival rate is 19.3%, whilst the five-year survival rate is 2.9%.[19] Considering nearly half of patients with lung cancer in England are diagnosed with Stage IV disease (49%, Section B.1.3.1), these extremely low survival rates underscore the need for better treatments for patients diagnosed with advanced lung cancer.

Furthermore, the impact of COVID-19 on lung cancer diagnosis and treatment has been the subject of several studies and reports.[47-50] The shifting of healthcare and hospital resources to treating COVID-19 patients led to delays in diagnosis of patients with lung cancer and changes in the treatment strategies of lung cancer patients.[49, 50] Delays in diagnosis attributed to COVID-19 may result in an additional 5% increase in the number of deaths of advanced NSCLC patients up to five years after diagnosis.[48]

The prognostic importance of the BRAF V600E mutation in NSCLC has not been widely studied, but some studies have found that patients harbouring the V600E mutation have shorter disease free survival and overall survival (OS) when compared to wildtype NSCLC.[36, 51] Marchetti et al. (2011) noted that the V600E mutation was associated with a more aggressive tumour histology, characterised by micropapillary features.[36] UK clinicians noted that patients harbouring the BRAF V600 mutation have a more aggressive disease compared to wild-type NSCLC.[8]

Patients with advanced NSCLC have a worse health-related quality of life (HRQoL) compared to both the general population and patients with other advanced cancers.[52] Various studies have shown that patients with advanced NSCLC experience diminished HRQoL compared to patients with less severe NSCLC.[53, 54] Advanced NSCLC is associated with severe symptoms, such as fatigue, loss of appetite, shortness of breath, as well as severe pain, depression and anxiety.[55-57]

NSCLC is also associated with a considerable burden of care. A descriptive, longitudinal study investigated the burden, skills preparedness, psychological distress, and quality of life for caregivers of patients with NSCLC.[58] The caregivers within the study experienced high selfperceived demands of caregiving responsibilities.[58] Over time, this burden increased significantly and in turn, caregiver preparedness to deal with the multiple domains of the care-giving role decreased.[58] Caregivers for patients with NSCLC also reported an increase in psychological distress and deterioration in psychological well-being and overall quality of life.[58]

B.1.3.5 Clinical pathway of care

Overview of the treatment pathway

Testing for a BRAF mutation in NSCLC is now part of routine UK clinical practice as of August 2020 (Section B.1.3.3). NGS panels, typically performed at Genomic Hubs, are used to test for the BRAF mutation, alongside other actionable mutations (ALK, EGFR, EML4-ALK, KRAS, MET, MET 14 exon skipping, NTRK, RET and ROS1).[4] This has been confirmed in a Novartis market research survey, which reported that **% of surveyed pathology laboratories stated that BRAF testing has moved to the Genomic Hubs.[5]

The treatment pathway for patients with advanced NSCLC in England and Wales is separated by targeted and non-targeted treatment options as per the NICE clinical guideline for lung cancer (NG122, Figure 2).[3]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 24 of 133

Figure 2: Current treatment pathway for previously untreated advanced NSCLC in UK clinical practice

==> picture [711 x 290] intentionally omitted <==

Footnotes:[a] Clinical opinion noted that for patients with previously untreated advanced NSCLC with a BRAF V600 mutation, pembrolizumab plus chemotherapy is preferred to pembrolizumab monotherapy, irrespective of PD-L1 expression, given the aggressive nature of the disease and thus prognostic importance of the BRAF V600 mutation (Section B.1.1). Abbreviations : ALK: anaplastic lymphoma kinase; EGFR: epidermal growth factor receptor; NSCLC: non-small cell lung cancer; PD-L1: programmed death-ligand 1; TA: Technology Appraisal; UK: United Kingdom.

Source: Based on NICE NG122[3] and clinical expert opinion.[8]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 25 of 133

As detailed in Section B.1.1, UK NSCLC experts confirmed to Novartis that, for patients with previously untreated advanced NSCLC without an oncogenic driver mutation, pembrolizumab plus chemotherapy is the most relevant treatment option. Clinical experts noted that pembrolizumab plus chemotherapy is preferred to pembrolizumab monotherapy irrespective of PD-L1 expression, given the aggressive nature of the disease and thus prognostic importance of the BRAF V600 mutation.

As such, pembrolizumab plus chemotherapy represents the only relevant comparator to dabrafenib and trametinib in this appraisal (detailed in Table 1 of Section B.1.1).

Limitations with current treatment options for advanced NSCLC patients with BRAF mutation

Pembrolizumab-based treatment regimens have been extensively studied in the wider population of patients with advanced NSCLC.[21, 22, 59, 60] However, the available data for the effectiveness of pembrolizumab-based treatment regimens for patients with advanced NSCLC harbouring a BRAF V600 mutation is limited (clinical systematic literature review [SLR] results are detailed in Appendix D.2).

Treatment with pembrolizumab can be associated with adverse events (AEs) that can lead to complications, such as immunotoxicity, infusion-related reactions, or discomfort associated with administration. These AEs can have a detrimental and potentially long term impact to a patient’s quality of life.[61, 62] Furthermore, pembrolizumab is administered via intravenous (IV) infusions and therefore requires nursing staff and in-patient hospital visits, which potentially represent an additional burden to patients and the healthcare system; an increased risk of IV cannulationrelated infections also exists for patients treated in hospitals. [63]

The availability of an oral treatment option, such as dabrafenib and trametinib, would likely have a positive impact on alleviating capacity issues within the NHS. Receiving treatment with an oral therapy as an alternative to IV therapies is further highlighted as an important treatment preference for NSCLC patients; a recent patient preference survey reported that oral treatment was preferred over IV therapies by the majority (60%) of patients who were surveyed. [24]

Patient advocates have also reported a preference for receiving a targeted treatment option that targets a patient’s specific gene mutation, potentially considering the improved survival outcomes observed for patients with other gene mutations. As part of the HTA for dabrafenib and trametinib in Canada, the registered clinicians and patient advocacy group providing input for the submission all emphasised the preference for patients with an identified driver mutation to be treated with targeted therapy upfront as first-line therapy, as opposed to non-targeted treatments.[64]

For these reasons, dabrafenib and trametinib was made available on an interim basis in response to the COVID-19 pandemic, to reduce the burden on the NHS due to patients attending hospital to receive IV treatment, as well as to reduce the risk of immunosuppression associated with standard of care treatments.[20] This highlights the enthusiasm of the clinical community to provide access to dabrafenib and trametinib for patients with advanced NSCLC with a BRAF V600 mutation.[20]

Data provided by NHS England via Blueteq forms suggest that *** of the initiations of dabrafenib and trametinib since August 2020 were in previously untreated advanced NSCLC patients, with the majority of the remaining initiations being in previously treated patients (where most patients received immunotherapy with chemotherapy at the time of first treatment decision [***%]).[2]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Clinical experts indicated that patients with previously treated NSCLC who received dabrafenib and trametinib likely did so because their BRAF mutation status was not known at the time of first treatment decision, meaning they first received an alternative treatment.[8] Clinicians recognise the BRAF V600 mutation as an actionable driver mutation in NSCLC which is routinely identified in UK clinical practice, and are treating this cohort of patients in line with upfront usage of targeted treatments for previously untreated advanced NSCLC patients with EGFR, ALK or other oncogenic driver mutations.[3, 8]

Positioning of dabrafenib and trametinib within the current clinical pathway of care

A summary of the proposed positioning of dabrafenib plus trametinib in the treatment pathway for patients with advanced NSCLC with a BRAF V600 mutation is presented in Figure 2.

As detailed previously, dabrafenib and trametinib will primarily represent a treatment option for patients with previously untreated advanced NSCLC with a BRAF V600 mutation, in line with usage since August 2020, and the position taken in the treatment pathway for targeted treatments for other NSCLC driver mutations.[2, 65-69]

It is the view of Novartis that the population patients with previously treated advanced NSCLC with a BRAF V600 mutation eligible for dabrafenib and trametinib is likely to diminish in size over time. This decrease is expected as turnaround times for testing improve, and a patient’s mutation status is increasingly known at the time of first treatment initiation. Therefore, the previously treated patient population is expected to form a small (but clinically important) minority of patients eligible to receive dabrafenib with trametinib, constituting patients that have faced a delay in receiving their routine biomarker testing (See Table 2).

B.1.4 Equality considerations

The inclusion of the BRAF V600 gene in the 2021/2022 National Genomic Testing Directory indicates BRAF V600 mutation testing is now routine in clinical practice.[4] However, Novartis understand that while the average testing turnaround time is between ***** days[5] , the uptake and turnaround times for genetic testing are variable across regions of the UK, depending on individual Genomic Laboratory Hubs and testing pathways.[8] As such, equality considerations relating to BRAF V600 mutation testing may be relevant to this appraisal.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2 Clinical effectiveness

Clinical effectiveness summary

BRF113928 trial

• The pivotal trial for dabrafenib and trametinib is BRF113928 (NCT01336634), a phase II, multicentre, open-label, single-arm trial that enrolled 36 patients with previously untreated advanced NSCLC with a BRAF V600E mutation (Cohort C), and 57 patients with previously treated advanced NSCLC with a BRAF V600E mutation (Cohort B).[70-72]

• • Patients in Cohort C showed a clinically meaningful response to dabrafenib and trametinib, with an overall response rate (ORR) of 63.9% (95% CI: 46.2%, 79.2%). Responses were durable, with a median duration of response (DOR) of 10.2 months (95% CI: 8.3, 15.2).

• • At the time of the final data cut-off, once the trial was completed with a minimum of five years follow-up per patient, patients treated with dabrafenib and trametinib experienced a median PFS of 10.8 months (95% CI: 7.0, 14.5) and median OS of 17.3 months (95% CI: 12.3, 40.2), with estimated survival at Month 60 of **% (95% CI: **%, **%).

• • Similar results were observed in Cohort B (presented in Appendix M).

• Comparative efficacy evidence

• Robust data in the published literature for pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600 mutation is limited.

• • Novartis have therefore conducted a series of real-world evidence (RWE) studies to provide evidence for pembrolizumab plus chemotherapy for patients with previously untreated advanced NSCLC with a BRAF V600 mutation, including Melosky et al. (2021)[73] and the FLATIRON RWE study (2022) (unpublished data).

• • Melosky et al. (2021) was a RWE study that identified real-world patients with previously untreated advanced NSCLC with a BRAF V600 mutation, receiving either dabrafenib and trametinib, or pembrolizumab plus chemotherapy.[73] The weighted analysis did not report any statistically significant differences between the two treatments.

• • As Melosky et al. (2021) used RWE for dabrafenib and trametinib rather than clinical trial data (which is considered more robust, with the BRF113928 trial having a minimum of five years’ follow-up),[73] Novartis conducted a subsequent RWE study and associated weighted analysis: the FLATIRON RWE study (2022). The study (which is currently unpublished), aimed to compare the outcomes for dabrafenib and trametinib in the BRF113928 trial (Cohort C) to real-world treatment with pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600E mutation.

• • The FLATIRON RWE study (2022) weighted analysis did not report any significant differences with respect to PFS or OS between the two treatments. The study was associated with limitations, such as small sample sizes and differences in follow-up.

• • Given the results from both Melosky et al. (2021) and the FLATIRON RWE study (2022), an assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy was considered to be a reasonable conclusion. UK clinical experts also indicated that an assumption of clinical equivalence was the most reasonable conclusion that could be drawn from the data.[8]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.1 Identification and selection of relevant studies

An SLR was conducted to identify clinical evidence of the efficacy and safety of dabrafenib and trametinib and pembrolizumab plus chemotherapy, as well as other treatments for patients with advanced NSCLC with a BRAF V600 mutation. The SLR was initially conducted to be broad in scope and included studies in patients with advanced NSCLC with any BRAF mutation, at any line of therapy, and receiving a range of treatments for NSCLC. All studies identified were subsequently reviewed for relevance: this section focusses on evidence for either dabrafenib and trametinib or pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600 mutation. For a full list of studies in the previously untreated and previously treated advanced NSCLC with a BRAF mutation patient populations, see Appendix D.

In total, 28 studies met the broad inclusion criteria of the SLR, of which 9 studies were conducted in patients with previously untreated advanced NSCLC and a BRAF mutation, receiving either dabrafenib and trametinib, or pembrolizumab plus chemotherapy. A summary of these 9 studies is provided in Table 4 below, alongside an assessment of their relevance to this submission.

Dabrafenib and trametinib

Of the 9 potentially relevant studies, 5 studies included data for dabrafenib and trametinib: the pivotal BRF113928 trial,[70-72, 74-76] and 4 observational studies.[77-82] Given the availability of clinical trial data with over 5 years’ follow-up for dabrafenib and trametinib from the BRF113928 trial, the observational studies for dabrafenib and trametinib were not considered to represent robust sources of data and were not considered further. The one exception was Melosky et al. (2021),[73] which provided real-world evidence for patients with previously untreated advanced NSCLC with a BRAF V600 mutation receiving dabrafenib and trametinib and pembrolizumab plus chemotherapy (as well as other treatments). As such, Melosky et al. (2021) provides potentially relevant comparative efficacy evidence between dabrafenib and trametinib and pembrolizumab plus chemotherapy in this submission, and is detailed in Section B.2.2.2 and Section B.2.9.2.

Pembrolizumab plus chemotherapy

To ascertain whether any of the 9 studies identified within the clinical SLR would be suitable for consideration as a source of comparative efficacy evidence for pembrolizumab plus chemotherapy, a feasibility assessment was conducted as follows. Each study was reviewed and only considered further if it met the following criteria:

• Studies with a sample size ≥10 patients, to avoid limitations from very small sample sizes

• Studies reporting PFS and OS Kaplan-Meier data, to allow inclusion in an economic model

• Studies reporting baseline characteristics within the relevant patient population (i.e., patients with a BRAF V600 or V600E mutation), so that the relevant patient populations can be analysed for comparability with the BRF113928 trial

• Studies reporting outcomes separately for patients in the previously untreated advanced NSCLC setting versus those in the previously treated setting

The results of this feasibility assessment are presented in Table 4. Only two studies were identified that provided potentially relevant comparative evidence for pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600 mutation: two RWE studies conducted by Novartis, Kanakamedala et al. (2020)[83] and Melosky al. (2021).[73] These studies are discussed in more detail in Section B.2.2.2.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 4: Overview of studies included in the clinical SLR (for previously untreated patients)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 30 of 133

Footnotes :[a] Specific regimens not reported.[b] PFS and OS KM data available to Novartis.

Abbreviations: KM: Kaplan-Meier; NR: not reported; OS: overall survival; PD1: programmed death 1; PD-L1; programmed death-ligand 1; PFS: progression-free survival; TKI: tyrosine kinase inhibitor.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 31 of 133

B.2.2 List of relevant clinical effectiveness evidence (previously untreated patients)

B.2.2.1 Dabrafenib and trametinib

The primary clinical evidence for dabrafenib and trametinib is the BRF113928 trial, which was used in support of the marketing authorisation for dabrafenib and trametinib in this indication (Table 5). This was a single-arm trial including adult patients with advanced NSCLC and a BRAF V600E mutation. The BRF113928 trial was initiated in August 2011 and is now complete with a minimum of 5 years of follow-up per patient, with the last patient visit in January 2021.

It should be noted that the trial included patients with a BRAF V600E mutation, while the licence for dabrafenib with trametinib is for patients with a BRAF V600 mutation. BRAF V600E mutations comprise approximately 96% of all BRAF V600 mutations, meaning that the trial evidence can be considered generalisable to the wider licensed population for dabrafenib and trametinib.[32, 39, 87]

Table 5: Relevant clinical trial effectiveness evidence for dabrafenib and trametinib

Abbreviations: CR: complete response; DOR: duration of response; IA: investigator assessment; NA: not applicable; NSCLC: non-small cell lung cancer; ORR: overall response rate; OS: overall survival; PFS: progression-free survival; PR: partial response; RCT: randomised controlled trial; RWE: real-world evidence Source : BRF113928 Clinical Study Report.[17]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.2.2 Pembrolizumab plus chemotherapy

As detailed above in Table 4, the clinical SLR identified two studies which provided evidence for pembrolizumab plus chemotherapy in previously untreated advanced NSCLC patients with a BRAF V600 mutation which were considered robust enough for further consideration as potential sources of comparative effectiveness evidence. Both studies were publications on distinct RWE studies conducted by Novartis. In addition, a more recent RWE study has been conducted by Novartis that is not yet published. These three studies are detailed below:

Melosky et al. (2021) :[73 ] A publication of a RWE study conducted by Novartis, that compared real-world dabrafenib and trametinib versus real-world pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600 mutation.

Kanakamedala et al. (2020) :[83 ] An earlier, distinct, external control analysis of the BRF113928 trial, which compared data for dabrafenib and trametinib from the BRF113928 trial with real-world pembrolizumab plus chemotherapy. However, in an effort to increase the sample size, this analysis focused on a BRAF-mutant population, including both BRAF V600 patients and other subtypes of the BRAF mutation.

FLATIRON RWE study (2022, unpublished) : Since the publication of Kanakamedala et al. (2020), an updated data cut for the BRF113928 trial has been released. Novartis have therefore conducted an updated analysis, in line with the methodology used in Kanakamedala et al. (2020).[83 ] This study is termed the FLATIRON RWE study (2022), and is currently unpublished. The FLATIRON RWE study (2022) includes patients with advanced NSCLC and a BRAF V600E mutation. Kanakamedala et al. (2020) is therefore outdated, and not considered relevant to this submission.

Melosky et al. (2021) and the FLATIRON RWE study (2022) therefore represent the two relevant sources of comparative efficacy evidence for patients with previously untreated advanced NSCLC with a BRAF V600 mutation. These studies are discussed further in Section B.2.9.

Melosky et al. (2021) also compared real-world dabrafenib and trametinib to real-world pembrolizumab monotherapy, the results of which are presented in Appendix D.3.1. The FLATIRON RWE study (2022) additionally compared dabrafenib and trametinib to real-world pembrolizumab monotherapy, the results of which are presented in Appendix D.3.2.

As detailed in Section B.1.1, pembrolizumab monotherapy is not considered to represent a relevant comparator in this appraisal, but these results are presented for completeness.

B.2.3 List of relevant clinical effectiveness evidence (previously

treated patients)

The BRF113928 trial included evidence for patients with previously treated advanced NSCLC with a BRAF V600E mutation receiving dabrafenib and trametinib (Cohort B), while the FLATIRON RWE study (2022) included a comparison between Cohort B of the BRF113928 trial versus patients with previously treated advanced NSCLC receiving chemotherapy in real-world US clinical practice.

As detailed previously in Section B.1, dabrafenib and trametinib primarily represents a treatment option for patients with previously untreated advanced NSCLC with a BRAF V600 mutation. The

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

previously treated advanced NSCLC population represents a small (but clinically important) patient population, which is expected to diminish over time as the turnaround times for testing continue to improve, and a patient’s mutation status is increasingly known at the time of first treatment decision.

As a result, clinical effectiveness evidence for dabrafenib and trametinib for patients with previously treated advanced NSCLC with a BRAF V600 mutation from Cohort B of the BRF113928 trial is presented is presented in Appendix M, for completeness.

The sample sizes resulting from the comparative efficacy analysis in this population, with just *** patients with previously treated advanced NSCLC with a BRAF V600E mutation receiving chemotherapy in the FLATIRON RWE study (2022) were also not considered large enough to constitute robust comparative efficacy evidence, and **** of these patients received either docetaxel monotherapy or docetaxel plus nintedanib, the two most relevant chemotherapy regimens in UK clinical practice.

Therefore, economic analyses for patients with previously treated advanced NSCLC and a BRAF V600 mutation are not presented within this submission. However, alongside clinical results for Cohort B from the BRF113928 trial presented in Appendix M, comparative efficacy results from the FLATIRON RWE study (2022) between dabrafenib and trametinib and chemotherapy in patients with previously treated advanced NSCLC and a BRAF V600 mutation are presented in Appendix D.3.2.5 for completeness.

B.2.4 Summary of methodology of the relevant clinical

effectiveness evidence

B.2.4.1 Dabrafenib and trametinib evidence: BRF113928 trial

The BRF113928 trial was conducted at 71 study sites in 11 countries across the Asia Pacific region, Europe and North America. There were five study sites in the UK, all of which were in England.

In total the trial enrolled 177 adult patients with confirmed Stage IV NSCLC with a BRAF V600E mutation and consisted of three sequentially enrolled patient cohorts A, B and C. The flow of the 177 enrolled patients through the three trial cohorts is depicted in Figure 3.

Patients in Cohort A received dabrafenib monotherapy 150 mg twice daily. Dabrafenib monotherapy is not licensed for the treatment of advanced NSCLC, and therefore Cohort A is not discussed further in this submission.

Patients in Cohort B had previously treated NSCLC and received dabrafenib and trametinib. As detailed previously, these results are presented for completeness in Appendix M. Patients in Cohort B are also included alongside patients in Cohort C in a combined population used for the analysis of safety outcomes, as safety outcomes for Cohort C alone are not available (Section B.2.10).

Patients in Cohort C had previously untreated NSCLC and received the combination of dabrafenib 150 mg twice daily plus trametinib 2 mg once daily, with results presented in full below.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Figure 3: Patient flow in Cohorts A, B and C in BRF113928

==> picture [453 x 189] intentionally omitted <==

Abbreviations: NSCLC: Non-small cell lung cancer Source: BRF113928 Clinical Study Report;[89] BRF113928 Clinical Study Report – Updated Analysis.[17]

B.2.4.2 Study objectives

The outcomes measured in the BRF113928 trial are presented in Table 6. The primary objective was to assess overall response rate (ORR). Secondary objectives were to assess duration of response (DOR), progression-free survival (PFS), and OS, and to further characterise the safety, tolerability and pharmacokinetics (PK) of dabrafenib as a single agent (Cohort A) and in combination with trametinib (Cohorts B and C).

Table 6: Outcomes measured in BRF113928

Abbreviations: AE: adverse event; CR: complete response; DOR: duration of response; ECG: electrocardiogram; ORR: overall response rate; OS: overall survival; PFS: progression-free survival; RECIST: Response Evaluation Criteria in Solid Tumours. Source: BRF113928 Clinical Study Report.[17]

Eligibility criteria

Adult (18 years and older) male and female patients with stage IV NSCLC with a BRAF V600E mutation were enrolled into the study. A full list of inclusion and exclusion criteria is presented in Table 7.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 7: Eligibility criteria for BRF113928

Inclusion criteria

• Histologically or cytologically-confirmed diagnosis of Stage IV (according to AJCC Staging 7th Edition) NSCLC determined to be BRAF V600E mutation-positive in a CLIA-certified laboratory

• Patients in Cohort C were required to have not received prior systemic anti-cancer therapies for metastatic disease (i.e., dabrafenib and trametinib was the first-line treatment for metastatic disease)

• Measurable disease according to Response Evaluation Criteria in Solid Tumours (RECIST) V1.1[88]

• • Women of childbearing potential had to have a negative serum pregnancy test within 14 days before the first dose of study drug and must agree to use effective contraception during the study

• Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 to 2[90]

• • Previously tested for presence of EGFR and ALK mutations in lung cancer tissue confirmed in a CLIAcertified laboratory or equivalent. Patients with concomitant EGFR or ALK mutations were eligible if they had previously received EGFR or ALK inhibitor(s), respectively

• Life expectancy of more than three months

• Exclusion criteria • Previous treatment with a BRAF inhibitor or a MEK inhibitor prior to the start of study treatment

• Anti-cancer therapy including chemotherapy, radiation therapy, immunotherapy, biologic therapy or major surgery within 14 days prior to the start of study treatment

• Use of any investigational anti-cancer drug within 14 days or 5-half-lives (minimum 14 days), prior to the start of study medication

• • Presence of active gastrointestinal disease or other condition that could interfere significantly with the absorption of drugs

• • Known Hepatitis B Virus or Hepatitis C Virus infection. Patients with laboratory evidence of the cleared virus infection could be enrolled

• History of another malignancy < 3 years prior to starting study treatment or any malignancy with confirmed activating RAS mutation (although some exceptions were permitted)

• Patients with brain metastases were excluded if their brain metastases were:

  • Symptomatic or treated (surgery, radiation therapy) but not clinically and radiographically stable three weeks after local therapy (as assessed by contrast enhanced MRI or CT) or were asymptomatic and untreated but > 1 cm in the longest dimension

• A history or evidence of cardiovascular risk

• Pregnant, or actively breastfeeding females

Abbreviations: AJCC: American Joint Committee on Cancer; ALK: anaplastic lymphoma kinase; CLIA: clinical laboratory improvement amendments; CT: computed tomography; ECOG PS: Eastern Cooperative Oncology Group performance status EGFR: epidermal growth factor receptor; MEK: mitogen-activated protein kinase kinase; MRI: magnetic resonance imaging; NSCLC: non-small cell lung cancer; RECIST: Response Evaluation Criteria in Solid Tumours.

Source: BRF113928 Clinical Study Report.[17]

B.2.4.3 Baseline characteristics

The baseline characteristics of the patients in Cohort C are outlined in Table 8. The median age of enrolled patients was 67 years (range: 44, 91). There were more female than male patients (61% versus 39%, respectively). Most patients (83%) were white. The median time since diagnosis reflects a first-line metastatic disease population (2.05 months). More than half (58%) of patients were former smokers, 14% were current smokers, and 28% never smoked. The median duration of smoking was 30 years. The generalisability of the BRF113928 trial to UK clinical practice is discussed in Section B.2.9.3.2.

Table 8: BRF113928 patient baseline characteristics for Cohort C

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Footnotes :[a ] Predominantly adenocarcinoma.[b] Predominantly squamous cell carcinoma. Abbreviations : ECOG PS: European Cooperative Oncology Group Performance Status; NSCLC: non-small cell lung cancer; SD: standard deviation. Source : BRF113928 Clinical Study Report;[91] Planchard et al. (2021).[72]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.5 Statistical analysis and definitions of analysis sets

Details of the statistical methods and analysis sets used in BRF113928 are provided in Appendix M.1.1.

B.2.6 Quality assessment

Details of the quality assessment of the BRF113928 is provided in Appendix M.1.2.

B.2.7 Clinical effectiveness results

The BRF113928 trial was recently completed, with the last patient visit in 2021. Clinical effectiveness results are presented based on the most recent and complete data cut of the trial (24[th] February 2021), which includes a minimum of five years’ worth of follow-up data for each patient. This extended follow-up, relative to many other trials for patients with advanced cancers, should be considered a key strength of the analyses. Patients had a median follow-up of 16.3 months, which ranged from 0.4 months to 80 months.

B.2.7.1 Overall response rate

As of the final 24[th] February 2021 data cut-off, a clinically meaningful response rate was observed following treatment with dabrafenib and trametinib, as shown in Table 9.

The proportion of patients with confirmed ORR by IA was 63.9% (95% CI: 46.2, 79.2), including two (6%) patients who had a CR and 21 (58%) patients who had a PR (Table 9). In addition, four (11%) patients had SD, resulting in a DCR of 75.0% (95% CI: 57.8, 87.9). No data for ORR by independent review committee (IRC) evaluation were available for this data cut-off.

Table 9: BRF113928 summary of ORR by IA for Cohort C

Abbreviations : CI: confidence interval; CR: complete response; DCR: disease control rate; IA: investigator assessment; NE: not evaluable; ORR: overall response rate; PR: partial response; SD: stable disease. Source : BRF113928 Clinical Study Report - Table 11-1.[91] ; Planchard et al . (2021).[72]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.7.2 Duration of response

A durable response was observed by IA following treatment with dabrafenib and trametinib at the time of the 24[th] February 2021 data cut-off. In total, ** (**%) patients with a confirmed response had subsequently experienced disease progression or death, with an estimated median DOR of 10.2 months (95% CI: 8.3, 15.2) (Table 10; Figure 4).

Table 10: BRF113928 summary of DOR by IA for Cohort C

Abbreviations : CI: confidence interval; DOR: duration of response; IA: investigator assessment. Source : BRF113928 Clinical Study Report - Table 11-7;[91] Planchard et al . (2021).[72]

Figure 4: BRF113928 KM curve for DOR by IA for Cohort C

==> picture [482 x 246] intentionally omitted <==

Dotted lines indicate 95% CIs Abbreviations: CI: confidence interval; DOR: duration of response; IA: investigator assessment; KM: KaplanMeier.

Source: Planchard et al . (2021).[72]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.7.3 Progression-free survival

The IA-assessed PFS at the 24[th] February 2021 data cut-off are shown in Table 11 and Figure 5. Median PFS in patients receiving first-line dabrafenib and trametinib was 10.8 months (95% CI: 7.0, 14.5), with estimated PFS of **% after 12 months (95% CI: **%, **%).

Table 11: BRF113928 summary of PFS by IA for Cohort C

Abbreviations : CI: confidence interval; IA: investigator assessment; PFS: progression-free survival. Source: Planchard et al . (2021);[72] BRF113928 Clinical Study Report: Table 11-4.[91]

Figure 5: BRF113928 KM curve for PFS by IA for Cohort C

==> picture [452 x 233] intentionally omitted <==

Dotted lines indicate 95% CIs

Abbreviations: CI: confidence interval; IA: investigator assessment; KM: Kaplan-Meier; PFS: progression-free survival

Source: Planchard et al . (2021).[72]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.7.4 Overall survival

OS data from the 24[th] February 2021 data cut-off are presented in Figure 6. The estimated median OS of 17.3 months (95% CI; 12.3, 40.2) appears conservative, with an estimated survival at 60 months of **% (95% CI; **, **), indicating that dabrafenib and trametinib may have potential long-term survival benefits not reflected by the median OS.

Table 12: BRF113928 summary of OS for Cohort C

Abbreviations: CI: confidence interval; OS: overall survival. Source: Planchard et al . (2021);[72] BRF113928 Clinical Study Report - Table 11-10.[91]

Figure 6: BRF113928 KM curve for OS in Cohort C

==> picture [445 x 236] intentionally omitted <==

Dotted lines indicate 95% CIs Abbreviations: CI: confidence interval; KM: Kaplan-Meier; OS: overall survival. Source: Planchard et al . (2021).[72]

B.2.8 Subgroup analysis

No subgroup analyses of the BRF113928 trial were undertaken.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.9 Indirect comparisons

B.2.9.1 Summary of relevant comparative effectiveness evidence

Given the single-arm nature of the BRF113928 trial, randomised comparative evidence between dabrafenib and trametinib and pembrolizumab plus chemotherapy is not available. Given the limited published data in the literature for pembrolizumab plus chemotherapy for patients with previously untreated advanced NSCLC and a BRAF V600 mutation (Section B.2.1), Novartis considered a series of RWE studies to represent the most robust source of evidence for comparative assessment (Section B.2.2.2).

Since routine testing for BRAF mutations in advanced NSCLC patients in the UK was only recently established (2021/2022), it would not be possible to identify patients with a BRAF V600 mutation receiving pembrolizumab plus chemotherapy from UK clinical data sets. Similarly, across Europe, testing for BRAF mutations in advanced NSCLC was only recently recommended by the European Society of Medical Oncology in 2018, following the European marketing authorisation of dabrafenib with trametinib in 2017.[92] However, testing rates and corresponding reimbursement of testing services remains variable across countries, so no further investigation was taken to assess registries in Europe, as identifying patients receiving pembrolizumab plus chemotherapy with a BRAF V600 mutation was considered unlikely.[93]

In contrast, dabrafenib and trametinib has been available in the United States since 2017, so it was assumed that BRAF testing would be available as part routine care. Furthermore, pembrolizumab plus chemotherapy has also been available in the United States since 2018. Novartis therefore conducted a series of RWE studies utilising the FLATIRON Health database to provide evidence for patients with advanced NSCLC with a BRAF mutation treated with current standard of care such as pembrolizumab plus chemotherapy.[73, 83] The FLATIRON Health enhanced data mart (EDM) constitutes a well-documented population of patients observed contemporaneously through electronic health records (EHR) ascertained in the FLATIRON Health network, which comprises over 280 community oncology practices and academic medical centres in the US.[94-97] The FLATIRON database has previously been used to inform comparative efficacy estimates for previous NICE appraisals in NSCLC.[98, 99]

As detailed in Section B.2.2, two RWE studies represent the comparative effectiveness evidence in this submission. Melosky et al. (2021)[73] compares real-world dabrafenib and trametinib with real-world pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC harbouring a BRAF V600 mutation. A weighted analysis has been conducted using these data and is presented in Section B.2.9.2

The FLATIRON RWE study (2022), which is an external control analysis of the BRF113928 trial compares data for dabrafenib and trametinib from the BRF113928 trial versus real-world pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600E mutation. The subsequent weighted analysis using these data is considered to represent the primary comparative effectiveness evidence in this submission, as this includes the more robust BRF113928 trial data for dabrafenib and trametinib (versus the real-world data for dabrafenib and trametinib used in Melosky et al. [2021]). Full details are provided in Sections B.2.9.3 to B.2.9.3.3 below.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.9.2 Melosky et al. (2021)

Melosky et al. (2021)[73] was a non-interventional, retrospective observational study comparing real-world outcomes among patients diagnosed with previously untreated advanced NSCLC with a BRAF V600 mutation from the FLATIRON database.

A propensity score methodology was used to closely weight the two populations given differences in prognostic variables at baseline. A summary of the results of this weighted analysis are shown in Table 13.

After weighting, dabrafenib and trametinib resulted in numerically longer OS and similar PFS when compared with pembrolizumab plus chemotherapy (Table 13). However, no significant differences were detected with respect to OS or PFS between the two treatments. UK clinicians indicated that a reasonable conclusion was one of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy, citing the similar trajectory of the KaplanMeier curves for both treatments, and the low number of patients at risk after Month 12 (Figure 8 in Appendix D.3.1.4).

Full details of the Melosky et al. (2021) study[73] and associated weighted analysis are provided in Appendix D.3.1.

Table 13: Summary of results of Melosky et al. (2021)[73]

Abbreviations: CI: confidence interval; ESS: effective sample size; HR: hazard ratio; NR: not reached; OS: overall survival; PFS: progression-free survival. Source: Melosky et al. (2021).[73]

B.2.9.3 FLATIRON RWE study (2022)

B.2.9.3.1. Methodology

In the FLATIRON RWE study (2022) Novartis compared dabrafenib and trametinib data from Cohort C of the BRF113918 trial, to an external control arm of real-world pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600E mutation. The FLATIRON RWE study (2022) and associated weighted analysis represent the pivotal source of comparative efficacy for this submission.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Overview

The FLATIRON RWE study (2022) included:

• Adult patients who received first-line treatment for advanced NSCLC

• Patients with a BRAF V600E mutation in lung cancer tissue

• Patients with an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0– 2

• Patients who had previously received treatment with a BRAF-inhibitor were excluded

Full details of the inclusion/exclusion criteria of the FLATIRON RWE study (2022), and the associated attrition table, are provided in Appendix D.3.2.1.

In total, ** patients with previously untreated advanced NSCLC with a BRAF V600E mutation who received first-line treatment with pembrolizumab plus chemotherapy were identified.

In line with the Kanakamedala et al. (2020) study, a separate sensitivity analysis was also conducted as part of the FLATIRON RWE study (2022), where the BRAF V600E inclusion criterion was relaxed to include a wider population of patients with any BRAF mutation (termed the BRAF V600E+ population). While many of the patients in the BRAF V600E+ population are presumed to have a BRAF V600 mutation based on published epidemiology, the data captured in the FLATIRON database does not contain complete data on the subtype of BRAF mutation for all patients. The results for the BRAF V600E+ patient population are presented for completeness in D.3.2.3, but are not considered relevant for this submission.

Statistical methods

It was anticipated that differences in prognostic variables at baseline would exist between patients receiving pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022) and patients in Cohort C of the BRF113928 trial. Weighting by odds and the inverse probability of treatment weighting (IPTW) was used to account for differences between the trial and real-world cohort. Propensity scoring is recommended by NICE Decision Support Unit guidance (Technical Support Document [TSD] 17) as an approach to minimise the risk of bias when making inferences on treatment effect using observational data.[100]

Propensity scoring was estimated using a logistic regression that modelled treatment assignment as a function of the following baseline characteristics:

• Age at index

• Sex

• ECOG PS

• History of smoking

• Race

Clinical experts consulted as part of this appraisal noted that all these covariates were appropriate.[8] Further details on the prognostic weighting methodology can be found in Appendix D.3.2.5, and assessment of proportional hazards can be found in Appendix N.1.

B.2.9.3.2. Results

Baseline characteristics

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Baseline characteristics for the RWE patients who received pembrolizumab plus chemotherapy are shown in Table 14 (unweighted and weighted), and are compared to those in the BRF113928 trial, Cohort C. In the weighted analysis, the baseline characteristics for the adjusted

pembrolizumab plus chemotherapy cohort were similar to those in BRF113928 Cohort C with the exception of the sex covariate where there was a slightly lower percentage of females in the realworld cohort.

Table 14: Baseline characteristics for patients in the FLATIRON RWE study (2022) (BRAF V600E mutation)

Abbreviations: ECOG PS: Eastern Cooperative Oncology Group Performance Status; IQR: interquartile range; SD: standard deviation

Generalisability to UK clinical practice

The baseline characteristics of patients in the BRF113928 trial and the FLATIRON RWE study (2022) appear broadly similar to those seen in UK clinical practice, based on the UK Lung Cancer Audit (2022) dataset,[27] which reports data for a cohort of 28,337 patients with NSCLC in the UK, including approximately 44% with advanced NSCLC. For patients with NSCLC, the UK Lung Cancer Audit (2022 data), reported a mean age of 72.8 years of age, 52% were male, and 46% had a history of smoking at time of diagnosis for a NSCLC patient (though it should be noted that a substantial further percentage of patients did not have their smoking status reported in the audit data).

It should be noted that given the rarity of the BRAF V600 mutation, no large UK studies reporting baseline characteristics of patients with advanced NSCLC with a BRAF V600 mutation are available to compare to those of patients in the BRF113928 trial. Whilst the association of certain characteristics with the BRAF V600 mutation has been suggested, there remains no clear

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

consensus as to the difference in characteristics in patients harbouring a BRAF V600 mutation, as compared to wild-type disease.

Clinicians consulted as part of an advisory board concluded that no clear phenotypes have been identified for the BRAF V600 mutation, and as such considered the baseline characteristics in the BRF113928 trial and the FLATIRON RWE study (2022) to be broadly aligned with the population of patients in UK clinical practice harbouring a BRAF V600 mutation.[8]

Progression-free survival

Table 15 presents a summary of the PFS comparison between dabrafenib and trametinib (BRF113928 trial, Cohort C) versus pembrolizumab plus chemotherapy (FLATIRON RWE study [2022]). Kaplan-Meier curves for PFS in the unweighted and weighted comparisons are presented in Figure 7.

The mean follow-up reported for dabrafenib with trametinib was **** months versus **** months for pembrolizumab plus chemotherapy. At the time of the respective data cut-offs, ****% of patients receiving dabrafenib and trametinib had experienced a real-world progression event, versus ***** of patients receiving pembrolizumab plus chemotherapy (unweighted). This difference is driven by the limited follow-up in the FLATIRON RWE study (2022) and increased censoring compared to the BRF113928 trial. The HR point estimates between dabrafenib and trametinib and pembrolizumab plus chemotherapy were associated with wide 95% CIs, which included 1 in the weighted analyses.

Beyond the absence of significant differences, it is difficult to draw any robust conclusions from this comparison, as discussed further in Section B.2.9.3.3 to B.2.9.3.5 below.

Table 15: PFS for patients receiving dabrafenib and trametinib (BRF113928 Cohort C) and patients receiving pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022) (BRAF V600E mutation)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations: CI: confidence interval; ESS: effective sample size; HR: hazard ratio; NA: not applicable; PFS: progression-free survival.

Figure 7: PFS KM curves – dabrafenib and trametinib (BRF113928 Cohort C) versus pembrolizumab plus chemotherapy (FLATIRON RWE study [2022], BRAF V600E mutation)

==> picture [399 x 310] intentionally omitted <==

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

==> picture [398 x 309] intentionally omitted <==

Footnotes: “PD(L)1 + chemo” represents pembrolizumab plus chemotherapy. Abbreviations : KM: Kaplan-Meier; PFS: progression-free survival; PD-L1: programmed-death ligand 1

Overall survival

A summary of the OS comparison between dabrafenib and trametinib (Cohort C) versus pembrolizumab plus chemotherapy (FLATIRON RWE Study [2022]) is presented in Table 16, and Kaplan-Meier curves for OS in the unweighted and weighted comparisons are presented in Figure 8. Approximately **% of patients receiving dabrafenib and trametinib experienced an OS event, compared to **% in the real-world cohort (differences driven by varied follow-up durations and censoring between the two studies). The HRs between dabrafenib and trametinib and pembrolizumab plus chemotherapy were associated with wide 95% CIs, which included 1, across both the unweighted and weighted analyses.

Beyond the absence of significant differences, it is difficult to draw any robust conclusions from this comparison, as discussed further in Section B.2.9.3.3 to B.2.9.3.5 below.

Table 16: OS for patients receiving dabrafenib and trametinib (BRF113928 Cohort C) and patients receiving pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022) (BRAF V600E mutation)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations: CI: confidence interval; ESS: effective sample size; NA: not applicable; OS: overall survival.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Figure 8: OS KM curves – dabrafenib and trametinib (BRF113928 Cohort C) versus pembrolizumab plus chemotherapy (FLATIRON RWE study [2022], BRAF V600E mutation)

==> picture [394 x 305] intentionally omitted <==

==> picture [395 x 306] intentionally omitted <==

Footnotes: “PD(L)1 + chemo” represents pembrolizumab plus chemotherapy. Abbreviations : KM: Kaplan-Meier; OS: overall survival; PD-L1: Programmed-Death Ligand 1.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.2.9.3.3. Clinical interpretation of the FLATIRON RWE study (2022)

UK clinical experts noted that drawing robust conclusions from the results of the updated external control analysis where real-world pembrolizumab plus chemotherapy (from the FLATIRON RWE study [2022]) was compared with dabrafenib and trametinib (Cohort C, BRF113928 trial) was not possible, given the important differences in follow-up between treatments, and the small patient numbers across both datasets.[8]

The results of the FLATIRON RWE (2022) study were presented to four UK clinical experts with experience of using dabrafenib and trametinib in UK clinical practice for patients with previously untreated advanced NSCLC harbouring a BRAF V600 mutation.[8] The clinicians were presented with unweighted and weighted PFS and OS results, and noted the small patient numbers in the weighted analysis.[8]

In the weighted analysis for PFS, the clinicians noted that until approximately Month 9, the Kaplan-Meier curves closely follow each other, suggesting similar outcomes.[8] This is broadly in line with time to discontinuation estimates for the two treatments (**** months for dabrafenib and trametinib and *** months for pembrolizumab plus chemotherapy, Section B.3.3.2.4). In the weighted analysis for OS, the clinicians noted that from Month 0 to Month 12, the results indicated that the Kaplan-Meier curves for both treatments were similar and overlap with one another, after which the estimates become more unstable.

The clinical experts noted that a true difference between the two treatments could not be assessed due to a lack of follow-up in the pembrolizumab plus chemotherapy cohort, and based on the early observations of the weighted analysis, clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy is a reasonable conclusion.[8]

This conclusion is supported by the clinical interpretation of the Melosky et al. (2021) results,[73] detailed in Section B.2.9.2, which provide further support that an assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy is reasonable.

B.2.9.3.4. Uncertainties in the FLATIRON RWE study (2022) weighted analysis

There are several uncertainties in the weighted analysis that are considered below:

• The BRF113928 is a complete and mature data set, with a minimum of five-years’ worth of follow-up data available for all patients (mean follow-up is **** months), reducing the uncertainty associated with the long-term PFS and OS outcomes for dabrafenib and trametinib. In contrast, the mean duration of follow-up in the FLATIRON RWE study (2022) for pembrolizumab plus chemotherapy was **** months. The differences in follow-up time could also lead to the observed differences in event rates (% of real-world progression events) between the trial and real-world cohorts

• In total, **% of patients in the pembrolizumab plus chemotherapy cohort initiated treatment in 2021, and were only followed up for 12 months or less (with a data cut-off date of 31[st] January 2022). As a result of these patients in particular, the censoring rate was much higher in the real-world cohort during the first 12 months, compared to the BRF113928 trial. This limited follow-up, and the resulting censoring, could lead to overestimation of the HR point estimates in favour of pembrolizumab plus chemotherapy

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

• The overall effective sample size was ** for pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022) after weighting, (compared to ** in the unweighted analysis). The reduction in sample size means that greater weights are being placed on the outcomes for a select few patients, meaning that methods were used to minimise bias that led to lower effective sample size

• A small number of patients in the FLATIRON RWE study 2022 (N=*) had an unknown ECOG PS; it was therefore assumed that these patients had an ECOG PS equal to 1

• There were marked differences in how PFS was defined in the FLATIRON RWE study (2022) versus the BRF113928 trial. For instance, in the BRF113928 trial, PFS was evaluated based RECIST v1.1 criteria to determine disease progression, based on specific assessment timepoints prospectively outlined in the study protocol, and this adds to the uncertainty.[88]

B.2.9.3.5. Conclusions

An assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy was assumed based on the clinical interpretation of the FLATIRON RWE 2022 results.

B.2.10 Adverse reactions: BRF113928 trial

The assessment of safety for dabrafenib and trametinib was a secondary objective of the BRF113928 trial. No safety data were collected for the relevant comparators in the FLATIRON RWE (2022) study.

Safety data from the latest data cut-off of the BRF113928 trial are presented within this section. All patients who received dabrafenib and trametinib in the BRF113928 trial, in either the previously untreated (Cohort C) or previously treated (Cohort B), were combined for the analysis of safety outcomes; safety outcomes for Cohort C alone are not available.

B.2.10.1 Overview of adverse events

An overview of the AEs experienced by all patients in the combined safety population of the BRF113928 trial (N=93) at the 24[th] February 2021 data cut-off is presented in Table 17. Overall, the safety profile of dabrafenib and trametinib in patients with previously untreated advanced NSCLC with a BRAF V600 mutation can be considered manageable.

Table 17: BRF113928 summary of AEs (combined safety population)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations : AE: adverse event; SAE: serious adverse event. Source : BRF113928 Clinical Study Report - Table 12-5;[91] Planchard et al . (2021).[72]

B.2.10.2 Adverse events leading to treatment discontinuation

An overview of AEs leading to treatment discontinuation for all patients in the combined safety population in the BRF113928 trial (N=93) at the 24[th] February 2021 data cut-off are presented in Table 18.

In total, ** patients (***) had experienced at least one AE leading to permanent discontinuation of study treatment. Overall, ***** ***** patients experienced a grade 3/4 AE that led to treatment discontinuation, and **** patients (**) experienced a grade 5 AE that led to treatment discontinuation.

Table 18: BRF113928 summary of AEs (≥1% in total) leading to treatment discontinuation by maximum grade for all patients (combined safety population)

Abbreviations : AE: adverse event. Source : BRF113928 Clinical Study Report - Table 12-14.[91]

Dose reductions

Overall, in the combined safety population (N=93), ** patients () experienced an AE resulting in dose reduction of treatment; of those, ** patients () experienced a grade 3/4 AE resulting in dose reduction. The most frequently occurring AEs leading to dose reduction were ******* which occurred in ** patients (***), followed by *********, ****** and ******** in **** patients (**) each.[91]

Dose interruptions

Overall, in the combined safety population (N=93), ** patients () experienced an AE resulting in dose interruption of treatment; of those, ** patients () had a grade 3/4 AE, and ***** patients () had a grade 5 AE that led to dose interruption. The most frequently occurring AEs leading to dose interruption were ******* which occurred in ** patients (), followed by ******** in ** patients (), and ****** in * patients (*).[91]

B.2.10.3 Adverse events by preferred term for all patients

A breakdown of AEs by preferred term for all patients in the combined safety population BRF113928 trial (N=93) at the 24[th] February 2021 data cut-off are presented in Table 19.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Overall, 92 patients (99%) experienced at least one AE, the majority of events were classed as grade 3/4 (66%). Only eight patients in the total combined safety population experienced an AE of this maximum grade severity (grade 5).

The most frequently occurring AE was pyrexia, which was experienced in 52/93 (56%) of patients; most of these events were grade 2 (26%) and grade 1 (24%) in severity. Beyond pyrexia, the most frequently occurring AEs were nausea (47/93 [51%]), vomiting (38/93 [41%]) and dry skin (36/93 [39%]). Of these, the majority of events were either grade 1 or grade 2.

Table 19: BRF113928 summary of AEs by preferred term for all patients (combined safety population)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations : AE: adverse event. Source : Planchard et al . (2021).[72]

B.2.10.4 Serious adverse events

The overview of SAEs occurring in all patients in the combined safety population of the BRF113928 trial (N=93) at the 24[th] February 2021 data cut-off are presented Table 20.

Overall, ** patients (***) experienced at least one SAE, where the ******** of events were pyrexia ** *** ***** ** *** ( patients; ). patients ( ) experienced a grade 5 SAE, including patients who experienced respiratory stress.

Table 20: BRF113928 summary of SAEs for all patients (combined safety population)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations : SAE: serious adverse event. Source : BRF113928 Clinical Study Report - Table 12-13.[91]

B.2.10.5 Deaths

An overview of the deaths that occurred within the combined safety population in the BRF113928 trial (N=93) at the 24[th] February 2021 data cut-off are presented in Table 21.

A total of ** () patients died whilst participating in the trial; however, **** of these patients died due to an SAE that was likely to be related to study treatment. The ******** of deaths () were due to the study indication (i.e., advanced NSCLC) and ** patients (***) died due to reasons other than the study indication (stroke, retroperitoneal bleed, pulmonary infection, respiratory and cardiac arrest, clozapine intoxication, myocardial infarction, euthanasia, subarachnoid

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

haemorrhage).

Table 21: BRF113928 summary of deaths for all patients (combined safety population)

Abbreviations : SAE: serious adverse event. Source : BRF113928 Clinical Study Report - Table 12-11.[91]

B.2.10.6 Safety summary

In summary, the safety profile of dabrafenib and trametinib was consistent with that reported for patients treated with dabrafenib and trametinib in melanoma, with no new safety signals identified.[72]

In line with the safety profile for dabrafenib and trametinib in melanoma, the most frequently observed AE was pyrexia (56%), which led to a dose reduction in 11 patients (12%) and treatment withdrawal in two patients (2%).[72] Within the trial, antipyretics were given to control fever, with oral corticosteroids recommended in instances in which antipyretics were insufficient.

For pyrexia with a temperature greater than or equal to 38.5°C or a complication (any temperature with dehydration, hypotension, or renal insufficiency), dabrafenib was paused and trametinib continued until resolution of pyrexia, and then dabrafenib was restarted at a lower dose. For less severe pyrexia (temperature less than 38.5°C with no associated symptoms), dabrafenib was withheld until resolution of pyrexia and then resumed at the same dose.

It should be noted that an updated and simplified protocol (compared to the above) for the management of pyrexia following treatment with dabrafenib and trametinib has been developed since the BRF113928 trial. This new management protocol allows both dabrafenib and trametinib to be interrupted if a patient’s temperature is ≥38.0°C. In case of recurrence, treatment can also be interrupted at the first symptom of pyrexia; both treatments can be restarted at the same dose level if patients are symptom free for ≥24 hours. This new algorithm appears to reduce the incidence of severe pyrexia outcomes, enabling patients to manage pyrexia at home, and helping an increased percentage of patients to remain on treatment compared to the BRF113928 trial. This protocol is now within the SmPCs for dabrafenib and trametinib, and represents guidance

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

for all pyrexia management, regardless of indication.[101, 102]

Finally, most grade 3/4 AEs were managed through dose modification, thereby mitigating the risk of unacceptable toxicity. The overall safety profile of dabrafenib and trametinib is therefore considered to be manageable with appropriate clinical protocols, supported by a wealth of experience amongst the melanoma clinical community.

B.2.11 Ongoing studies

No other studies of dabrafenib and trametinib in NSCLC are ongoing.

B.2.12 Interpretation of clinical effectiveness and safety evidence

B.2.12.1 Principle findings from the clinical evidence base

The comparison of the BRF113928 trial and the FLATIRON RWE study (2022) did not detect any significant differences with respect to PFS and OS between dabrafenib and trametinib and pembrolizumab plus chemotherapy. As such, and as detailed in Section B.2.9, UK clinical experts indicated that an assumption of clinical equivalence between the two treatments would be a reasonable conclusion, based on the early observations between the two treatments; longer-term estimates of pembrolizumab plus chemotherapy efficacy are associated with uncertainty, due to limited follow-up and high levels of censoring.

This conclusion was further supported by the clinical interpretation of the Melosky et al. (2021) data,[73] which compared real-world dabrafenib and trametinib with real-world pembrolizumab plus chemotherapy; again, a similar set of limitations were encountered with respect to limited followup and small patient numbers.

B.2.12.2 Strengths and limitations of the clinical evidence base

BRF113928 trial

Strengths

BRF113928 represents the only clinical trial investigating the efficacy and safety of dabrafenib and trametinib, and therefore the best clinical trial evidence for its efficacy and safety. The trial is completed, with a minimum of 5 years of follow-up per patient, and therefore represents a mature dataset. As such, this should be considered a key strength, and limits the uncertainty surrounding the long-term outcomes for dabrafenib and trametinib.

The trial assessed the survival outcomes of most relevance to the oncology setting, namely OS and PFS, with ORR and DOR providing additional key evidence for the anti-tumour activity of dabrafenib and trametinib. This response data benefited from being assessed by IA throughout the follow-up period, in addition to IRC in the 28[th] April 2017. IRC ORR data were in strong concordance with the IA assessment, providing greater confidence in the response data.

Limitations

Inclusion of a comparator or placebo arm was not considered for the BRF113928 trial as an RCT would require large sample sizes and an extended period of time to be adequately powered, something that is less feasible with a rare genetic mutation with a high unmet need, as is the

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

case for patients with advanced NSCLC with a BRAF V600 mutation.

FLATIRON RWE study (2022)

Strengths

There is very limited evidence for pembrolizumab plus chemotherapy in patients with advanced NSCLC with a BRAF V600 mutation in the published literature, and the quality of the evidence is hindered as there is a lack of follow-up for these patients.

The FLATIRON RWE study (2022) mitigates the limitations of the published literature, providing evidence for pembrolizumab plus chemotherapy in the patient population of relevance to this submission. The availability of individual patient data from the FLATIRON RWE study (2022) is a further advantage, facilitating the conduct of a more robust weighted analysis using propensity score weighting.

Limitations

The FLATIRON RWE study (2022) only identified ** patients with previously untreated advanced NSCLC with a BRAF V600E mutation treated with pembrolizumab plus chemotherapy. One key factor in this is likely to be that patients have been able to access dabrafenib and trametinib in US clinical practice since 2017, and given the increasingly preferential use of targeted therapies for patients with oncogenic driver mutations, it is likely that only a minority of patients in the US with previously untreated advanced NSCLC would receive pembrolizumab plus chemotherapy.

Additionally, given the recent introduction of pembrolizumab plus chemotherapy, the majority of the patients in the FLATIRON RWE study (2022) receiving pembrolizumab plus chemotherapy were only recruited in 2019 or later, introducing further uncertainty in the long-term estimates of survival, given the limited follow-up for this study.

Summary

The clinical effectiveness evidence presented within this submission is associated with uncertainty, due to the small sample sizes and limited follow-up for the relevant comparator pembrolizumab plus chemotherapy that impact the ability to draw robust conclusions for the updated external control analysis.

The results presented in the sections above aim to mitigate these limitations, but uncertainty remains given the available data sets for patients with advanced NSCLC harbouring a BRAF V600 mutation. Nevertheless, the clinical interpretation of the results of the updated external control analysis (FLATIRON RWE study [2022]), support a reasonable assumption that dabrafenib and trametinib are likely to be clinically equivalent to pembrolizumab plus chemotherapy in terms of PFS and OS in patients with previously untreated advanced NSCLC with a BRAF V600 mutation.[8]

As such, and as detailed further in Section B.3, it was considered most appropriate to consider an assumption of equal efficacy between dabrafenib and trametinib versus pembrolizumab plus chemotherapy in the base case economic analysis, given the limitations associated with the FLATIRON RWE study (2022).

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3 Cost effectiveness

Cost effectiveness summary

• A de novo cost-effectiveness model was developed in Microsoft Excel to estimate the costeffectiveness of dabrafenib and trametinib versus pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600 mutation. The model was a partitioned survival model, with three health states of PFS, PD and death. • Clinical efficacy data (PFS and OS) for dabrafenib and trametinib were based on the pivotal BRF113928 trial. • As detailed in Section B.2.9.2 and Section B.2.9.3, comparisons between dabrafenib and trametinib and pembrolizumab plus chemotherapy in Melosky et al . (2021),[73] and between the BRF113928 trial and the FLATIRON RWE study (2022), did not detect any significant differences between the two treatments with respect to PFS or OS. Given this, and that using data from the FLATIRON RWE study (2022) led to clinically implausible estimates, PFS and OS for pembrolizumab plus chemotherapy were set to be equivalent to dabrafenib and trametinib in the base case economic analysis. This assumption of clinical equivalence was validated by UK clinical experts.[8]

• In the absence of utility data collected in the BRF113928 trial, health state utility values, AE disutility and duration estimates were sourced from relevant published literature. Costs included drug acquisition and drug administration costs, subsequent treatment costs, AE costs, monitoring and follow-up costs, and end-of-life costs.

• In the base case economic analysis, dabrafenib and trametinib (at current PAS price) was associated with cost savings of *******, and ***** more quality-adjusted life years (QALYs) versus pembrolizumab plus chemotherapy (at list price), meaning that dabrafenib and trametinib dominated pembrolizumab plus chemotherapy.

• Deterministic and probabilistic sensitivity analyses (DSA/PSA) demonstrated the base case results were robust to uncertainty. With dabrafenib and trametinib at PAS price, versus pembrolizumab plus chemotherapy (at list price), the PSA showed that dabrafenib and trametinib had a ***% probability of being cost-effective versus pembrolizumab plus chemotherapy at both the £20,000 and £30,000 willingness-to-pay (WTP) thresholds. The main drivers of the base case economic analysis were differences in quality of life between the two treatments. • Scenario analyses were conducted to assess the appropriateness of the base case economic analysis assumptions, and across all scenarios, dabrafenib and trametinib was associated with a positive QALY gain and was dominant versus pembrolizumab plus chemotherapy.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

• The NHS is under significant and increasing capacity constraints brought on by COVID-19 pandemic, and the administrative burden associated with the management of immunotherapy IV regimens. Dabrafenib and trametinib represents an appropriate use of NHS resources and provides patients with the opportunity to receive an orally available, targeted treatment early on within their treatment course, in line with other advanced NSCLC patients with oncogenic driver mutations.

B.3.1 Published cost-effectiveness studies

A series of economic SLRs were conducted on 10[th] May 2021 to identify relevant economic evidence in advanced NSCLC with a BRAF mutation by searching for published evidence on cost-effectiveness analyses, utility data and cost and resource use data as per NICE guidance.

The methodology of all three economic SLRs is presented in Appendix G.1 (with results of the economic evaluations SLR presented in Appendix G.3, the utility data SLR presented in Appendix H.3, and the cost and resource use data SLR presented in Appendix I.3). No relevant records relating to advanced NSCLC with a BRAF mutation were identified in the SLR.

B.3.2 Economic analysis

Since no prior published economic evaluations in patients with advanced NSCLC with a BRAF mutation were identified in the economic SLR, a de novo economic model was developed to assess the cost-effectiveness of dabrafenib with trametinib versus pembrolizumab plus chemotherapy for patients with previously untreated advanced NSCLC with a BRAF V600 mutation.

As discussed in Section B.2, an assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy in terms of PFS and OS was assumed to be reasonable given the uncertainties associated with the updated external control analysis of the BRF113928 trial (FLATIRON RWE study [2022]). This assumption was therefore adopted within the base case economic analysis and was validated by UK clinical experts.[8] As described in Section B.3.3.1, alternative approaches using the FLATIRON RWE study (2022) data directly within the economic model were not plausible, when comparing the predicted survival outcomes for pembrolizumab plus chemotherapy with external sources.

Based on the assumption of equivalence for PFS and OS, the economic model does not calculate any differences in survival outcomes (life years [LYs]) between the two treatments. Any difference in QALYs is driven by the disutility associated with AEs and the administration of treatments via IV infusion. Time on treatment (ToT) and drug acquisition costs are still calculated separately for the two treatments, and cost differences are also calculated with respect to drug administration costs, subsequent treatment costs, and AE costs, as detailed in Section B.3.5.

The remaining sections below outline the approach for selecting the PFS and OS extrapolations for dabrafenib with trametinib based on data from the BRF113928 trial. Appendix N details the extrapolations for pembrolizumab plus chemotherapy (based on the weighted analysis of the FLATIRON RWE study [2022]) for completeness.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3.2.1 Model structure

In line with the NICE reference case, the economic analysis was conducted from the perspective of the NHS and Personal Social Services (PSS) in the UK and included direct medical costs over a lifetime horizon.[103] The economic model constructed was a partitioned survival model (PSM) consisting of 3 health states (PFS, progressed disease [PD], and death).

A PSM was chosen primarily because of the maturity of the data available for dabrafenib with trametinib, where the BRF113928 trial provides a minimum of 5-years of follow-up data for each patient. At the time of trial completion, the data were mature; % of patients within the BRF113928 trial experienced an event (progression or death) and the remaining patients (%) were censored with follow-up ended, at the time of the final data cut-off (24[th] February 2021) (see Table 11).

Where mature data are available, extrapolations of PFS and OS KM curves are associated with less uncertainty and therefore render the PSM a more appropriate approach to consider. Secondly, advanced NSCLC is a progressive disease that, given patients are in the later stages of disease, typically follows a natural course of disease progression and ultimately death. As such it was not considered necessary for other additional health states (and therefore additional complexity) to be modelled, and so a PSM with three health states was considered appropriate. The use of a PSM model structure is consistent with all of the recent NICE appraisals in advanced NSCLC.[6, 7, 22, 98, 99, 104-108] As such, a PSM model was considered the most appropriate model structure for this submission.

Within the model, a cohort of patients on each treatment enter the model in the PFS health state and either remain in this health state, transition to the PD state and then die, or die without entering the PD state. KM OS and PFS data for dabrafenib and trametinib from the BRF113928 trial were used to determine the occupancy of the three health states within the economic model. These data were extrapolated beyond the recorded study period using statistical models, displayed with smooth curves, as displayed in diagrammatic form in Figure 9.

Figure 9: Partitioned-survival model with three health states

==> picture [373 x 184] intentionally omitted <==

Abbreviations: OS: overall survival; PFS: progression-free survival.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3.2.2 Key features of the economic analysis

An overview of the key features of the base case economic analysis is provided below in Table 22, and described in more detail in the following sections.

Table 22: Key features of the economic analysis

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations : BNF: British National Formulary; eMIT: electronic Market Information Tool; HRQoL: healthrelated quality of life; NICE: National Institute of Health and Care Excellence NHS: National Health Service; NSCLC: non-small cell lung cancer; PSS: Personal Social Services; PSSRU: Personal Social Services Research Unit; UK: United Kingdom.

B.3.2.3 Patient population

The patient population included in the base case economic analysis consisted of patients with previously untreated advanced NSCLC with a BRAF V600 mutation, based on cohort C of the BRF113928 trial.

As discussed in Section B.1.1, subgroup economic analyses by line of therapy, tumour histology or PD-L1 expression were not conducted. Further details on why an economic analysis within the previously treated patient population was not possible are presented in Table 22 above.

As detailed in Section B.1.1, Table 2, subgroup analyses by tumour histology or PD-L1 expression would not be feasible, as these would be informed by prohibitively small patient numbers. Furthermore, dabrafenib and trametinib achieves clinical benefit via a mechanism of action independent of PD-L1 expression, and is expected to show similar clinical benefit regardless of histology. This approach is aligned with previous NICE appraisals for targeted treatments in advanced NSCLC, including NICE TA789 (tepotinib for treating advanced NSCLC with MET gene alterations) and TA760 (selpercatinib for previously treated RET fusion-positive advanced NSCLC), where the NICE Committee did not make recommendations restricted by histology group, based on factors including the lack of data for patients with squamous NSCLC, and the fact that squamous NSCLC represents a generally small patient population.[6, 98]

The patient characteristics informing the model are presented in Table 23, based on Cohort C of the BRF113928 trial.

Table 23: Patient characteristics model inputs

Footnotes :[a] Calculated using the Mosteller formula with data from BRF113928 trial (height and weight available only for combined Cohorts B and C; mean height: ***** cm, mean weight: ***** kg). Abbreviations: BSA: body surface area. Source: BRF113928 Clinical Study Report.[17]

B.3.2.4 Intervention technology and comparators

As discussed in Section B.1, the principal comparator to dabrafenib and trametinib in this submission is pembrolizumab plus chemotherapy. The base case economic analysis therefore compared dabrafenib and trametinib to pembrolizumab plus chemotherapy (specifically, pembrolizumab plus pemetrexed and platinum-based chemotherapy [carboplatin or cisplatin]).

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Details of the dosing regimens modelled for both the intervention and comparator are presented below:

Dabrafenib and trametinib

• Dabrafenib 150 mg twice daily[11]

• Trametinib 2 mg once daily[12]

Pembrolizumab plus chemotherapy

A summary of the dosing regimen modelled for pembrolizumab plus chemotherapy is presented in Table 24.

Pembrolizumab plus chemotherapy was assumed to comprise pembrolizumab plus pemetrexed and either carboplatin or cisplatin and, in line with the licensed indication, was modelled for a total duration of two years (35 three-weekly treatment cycles) (see Table 24). The proportions of patients assumed to receive either carboplatin or cisplatin as part of this regimen were 84.4% and 15.6%, respectively, based on the assumed split between carboplatin and cisplatin adopted in NICE TA789.[6]

Of the patients still on treatment after four treatment cycles (12 model cycles), ****% were modelled to receive maintenance treatment with pemetrexed alongside pembrolizumab, while the remaining ****% were modelled to receive pembrolizumab alone. These proportions were based on the proportions of patients in the pembrolizumab plus chemotherapy arm of the FLATIRON RWE study (2022) who received maintenance treatment with pemetrexed or not following discontinuation of carboplatin or cisplatin.

The dosing of pembrolizumab was based on clinical expert feedback that confirmed that during the COVID-19 pandemic, the dosing of pembrolizumab when administered as a monotherapy, changed from 3-weekly dosing to 6-weekly dosing.[8] As such, after the first four treatment cycles (12 model cycles), pembrolizumab was assumed to be administered at a three-weekly dosing interval if patients were also receiving maintenance pemetrexed, but at a six-weekly dosing interval for patients who were not receiving maintenance pemetrexed.

Pembrolizumab and maintenance pemetrexed were assumed to be given for a maximum of 2 years to account for the stopping rule imposed within the recommendation for pembrolizumab in combination with platinum-based chemotherapy in NICE TA683 (pembrolizumab with pemetrexed and platinum chemotherapy for untreated, metastatic, non-squamous NSCLC), which stipulates that all treatments should be stopped following two years of treatment, or earlier if the disease progresses.[21]

Table 24: Summary of the modelled pembrolizumab plus chemotherapy dosing regimen

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations : AUC: area under the curve.

B.3.3 Clinical parameters and variables

B.3.3.1 Summary of clinical data used in the model

In the base case economic analysis, PFS, OS and ToT data for dabrafenib and trametinib were derived from Cohort C of the BRF113928 trial.

In line with NICE DSU TSD 14, distributions were fitted independently to the KM curves for PFS and OS for dabrafenib and trametinib using data from Cohort C of the BRF1139128 trial (presented in Sections B.3.3.2.2 and B.3.3.2.3 below).

For completeness, the same was done for the real-world pembrolizumab plus chemotherapy cohort using the weighted data from the FLATIRON RWE study (2022) (presented in Appendix N). The predicted pembrolizumab plus chemotherapy PFS and OS outcomes were assessed for external validity by comparing them to the survival estimates in KEYNOTE-189, the pivotal RCT comparing pembrolizumab plus chemotherapy with platinum-based chemotherapy (see Table 25).[110] UK clinical experts noted that this trial did not select for patients harbouring the BRAF V600 mutation, but agreed it was an appropriate source to utilise for external validation.[8]

UK clinical experts noted that predicted PFS and OS for pembrolizumab plus chemotherapy from extrapolations of the FLATIRON RWE study (2022) weighted data were overestimated and clinically implausible compared to the published KEYNOTE-189 trial data (Table 25).[8, 110] KEYNOTE-189 reported a median OS of 22.0 months for pembrolizumab plus chemotherapy, with a 2-year OS rate of 45.7%. In contrast, the predicted 2-year OS rates for pembrolizumab plus chemotherapy were all higher than **%. The extrapolations predicted a median OS ranging from **** months (for the exponential model) to **** months (for the lognormal model)

Table 25: Comparison of pembrolizumab plus chemotherapy PFS and OS from extrapolations of the FLATIRON RWE study (2022) weighted data versus KEYNOTE-189

Abbreviations : OS: overall survival; PFS: progression-free survival; RWE: real-world evidence.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Given these uncertainties, the clinical experts agreed that it would be reasonable to assume that the PFS and OS estimates for pembrolizumab plus chemotherapy would be equivalent to the predicted survival outcomes for dabrafenib and trametinib using data from the BRF113928 trial, in line with the conclusions made in Section B.2 of this submission.[8] An assumption of clinical equivalency between dabrafenib and trametinib and pembrolizumab plus chemotherapy was also assumed in the previously conducted Canadian Agency for Drugs and Technologies in Health (CADTH) appraisal for dabrafenib and trametinib in this indication in Canada. [64]

Therefore, in the base case economic analysis it was assumed that PFS and OS for pembrolizumab plus chemotherapy were equivalent to PFS and OS for dabrafenib and trametinib, based on Cohort C of the BRF113928 trial.

The weighted ToT data for pembrolizumab plus chemotherapy from the FLATIRON RWE study (2022), was used in the base case economic analysis, as the ToT data were consistent with ToT values for pembrolizumab plus chemotherapy reported in the published literature (as detailed in Section B.3.3.2.4 below).

B.3.3.2 Implementation of survival data

The proportion of patients in the PFS, PD and death health states at each cycle in the model were defined by PFS and OS curves (Section B.3.2.1). As the follow-up period of the BRF113928 trial and the FLATIRON RWE study (2022) were shorter than the model time horizon (~30 years), extrapolations of the observed PFS, OS and ToT data were required.

In accordance with the NICE DSU TSD 14 guidance on survival analyses, a range of standard parametric distributions (exponential, Weibull, log-logistic, lognormal, Gompertz and generalised gamma) were explored.[111] Alternative model choices such as spline models were not required, as the standard parametric extrapolations appeared to provide a good fit to the observed data in all cases.

The goodness-of-fit criteria (including the Akaike information criterion [AIC] and the Bayesian information criterion [BIC]) were then estimated for each parametric function.

In determining the choice of survival model for the base case for dabrafenib and trametinib, consideration was given to the following, according to the recommendations provided in NICE DSU TSD 14:[111]

• AIC and BIC goodness-of-fit statistics (i.e., statistical fit) in order to assess how well the statistical models fitted to the observed data

• Visual inspection of the extrapolated curves versus the observed Kaplan-Meier curves

• Clinical plausibility for both short-term and long-term estimates of survival based on discussion with UK clinical experts and published data

Additionally, in order to ensure that any OS extrapolations did not provide implausible estimates of mortality, all mortality rates used in the model were bound by the age- and gender-specific natural mortality of the general population as a minimum (calculated using the 2018-2020 UK life tables).[112] Adjustments were made in the model traces to prevent logical inconsistencies:

• PFS could not exceed OS

• ToT could not exceed PFS

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3.3.2.1. Summary

A summary of the base case clinical efficacy data for PFS, OS and ToT for the comparison of dabrafenib and trametinib versus pembrolizumab plus chemotherapy in patients with previously untreated advanced NSCLC with a BRAF V600 mutation is presented in Table 26. A range of scenario analyses, detailed in Section B.3.10.3, were conducted to explore alternative efficacy parameters for both dabrafenib and trametinib and pembrolizumab plus chemotherapy.

Table 26: Summary of the clinical parameters used in the economic model

Abbreviations : AE: adverse events; OS: overall survival; PFS: progression-free survival; RWE: real-world evidence; ToT: time on treatment; UK: United Kingdom.

B.3.3.2.2. Progression-free survival

Dabrafenib and trametinib

To model PFS, the standard parametric distributions were fitted to the PFS individual patient data (IPD) for patients receiving dabrafenib and trametinib in Cohort C of the BRF113928 trial. The AIC and BIC values for each of the extrapolations are summarised in Table 27. Extrapolations of PFS using each model up to ten years are presented for all functions in Figure 10, to aid investigation of the visual fit of the distributions to the observed study data, and to aid investigation of the clinical plausibility of the long-term extrapolations.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 27: Summary of goodness-of-fit data for dabrafenib and trametinib PFS (BRF113928 Cohort C); standard parametric models

Footnotes:[a] A small AIC or BIC value represents a better goodness of fit. Abbreviations : AIC: Akaike information criterion; BIC: Bayesian information criterion; ITT: intention-to-treat; PFS: progression-free survival.

Figure 10: Dabrafenib and trametinib PFS extrapolations up to ten years (BRF113928 Cohort C)

==> picture [431 x 239] intentionally omitted <==

Abbreviations : KM: Kaplan-Meier; PFS: progression-free survival.

NICE DSU TSD 14 indicates that, in cases where survival data are relatively complete, the use of AIC/BIC tests may be of most use in determining the most appropriate curve selection, as the extrapolated proportion of the curve contributes little to the overall mean area under the curve.[111]

As such, given the maturity of the data from the BRF113928 trial, the best statistically fitting loglogistic curve was considered to represent the most appropriate distribution to extrapolate the PFS KM curve in the base case analysis. UK clinical experts noted that approximately **% of patients remained on treatment for 2.5 years in the BRF113928 trial, and noted that it would be plausible for patients who experienced few AEs to be on treatment for an extended amount of time. Based on this, the log-logistic curve would be considered clinically plausible, citing that very few patients would not have progressed after 5 years, in line with the results observed in the BRF113928 trial (Section B.2.7.3). As clinical equivalence was assumed in the base case analysis, the same curve was adopted for pembrolizumab with chemotherapy.

Burnham and Anderson (2004) indicate that, for two parametric models with less than ≤2 AIC

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

points between them, there is substantial support for the fact that the two models have the same merits.[113] Similarly, Raftery et al. (1995) indicate that there is weak evidence of any differences between two parametric models with less than ≤2 BIC points between them.[114] Larger differences provide increasing evidence for significant differences between any two models. As such, the lognormal curve (within *** AIC and BIC points of the log-logistic extrapolation) was explored within a scenario analysis (Section B.3.10.3). The next best fitting curve (the Generalised gamma, within *** AIC points and *** BIC points) was also explored for completeness. The remaining models were not considered in scenario analyses, as the AIC and BIC data indicate that they would be expected to provide a significantly worse fit to the observed PFS data, and, given the relative completeness of the PFS data from the BRF113928 trial, were not considered relevant.

B.3.3.2.3. Overall survival

Dabrafenib and trametinib

To model OS, the standard parametric distributions were fitted to the OS IPD for patients receiving dabrafenib plus trametinib in Cohort C of the BRF113928 trial. The AIC and BIC values for each of the extrapolations are summarised in Table 28. Extrapolations of OS using each model up to 15 years are presented for all functions in Figure 11, to aid the investigation of the visual fit of the distributions to the observed study data, and to aid investigation of the clinical plausibility of the long-term extrapolations.

Table 28: Summary of goodness-of-fit data for dabrafenib and trametinib OS (BRF113928 Cohort C); standard parametric models

Footnotes:[a] A small AIC or BIC value represents a better goodness of fit. Abbreviations : AIC: Akaike information criterion; BIC: Bayesian information criterion; ITT: intention-to-treat; OS: overall survival.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Figure 11: Dabrafenib and trametinib OS extrapolations up to ten years (BRF113928 Cohort C)

==> picture [450 x 289] intentionally omitted <==

Abbreviations : KM: Kaplan-Meier; OS: overall survival.

All of the OS extrapolations were seen to have a similar fit to the observed data, with only a *** difference in AIC points, and a *** difference in BIC points, between the best-fitting curve (the log-logistic) and the sixth-best fitting curve (the Weibull).

In the BRF113928 trial, the OS KM data estimated **% (95% CI: *** **) of patients were still alive after 4 years, and **% after 5 years (95% CI: *** **). Clinical experts were shown the above extrapolations of the KM data and asked to confirm which were clinically plausible.[8] The clinical experts noted the maturity of the BRF113928 trial data in this patient population as an advantage, but indicated that selection of the most appropriate curve was challenging, given the limited experience of dabrafenib and trametinib in this patient population in clinical practice.[8]

Clinicians also highlighted the aggressive nature of the BRAF mutation and expected survival to be less than % at 10 years.[8] The clinicians ultimately agreed that the Weibull extrapolation was the most plausible, as it predicted the lowest 10-year OS rate (%).[8] As such, the Weibull curve was selected to extrapolate OS in the base case analysis. As clinical equivalence is assumed in the base case analysis, the same curve was adopted for pembrolizumab with chemotherapy.

The exponential curve (10-year OS of 4.5%) was also considered potentially plausible, and explored in a scenario analysis (Section B.3.10.3). The other curves, which predicted a range of between ***% to ****% of patients would be alive at 10 years, were considered less clinically plausible and not explored further.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3.3.2.4. Time on treatment

Dabrafenib and trametinib

To model ToT, standard parametric distributions were fitted to the ToT IPD for patients receiving dabrafenib and trametinib in Cohort C of the BRF113928 trial. The ToT KM data from the BRF113928 trial were mature and provided a complete dataset.

Given this, it is possible to use the ToT KM curve directly. However, it was considered more appropriate to fit a parametric curve to the observed data, to smooth out the stepwise KM data and assume a more constant rate of discontinuation over time. The use of the KM data directly was explored in a scenario analysis.

The AIC and BIC values for each of the dabrafenib and trametinib ToT extrapolations are summarised in Table 29. Extrapolations of ToT using each model up to ten years are presented for all functions in Figure 12. No stopping rules were applied for dabrafenib and trametinib as per UK clinical practice; a relative dose intensity (RDI) was applied when calculating the costs, as detailed in Section B.3.5.1.

Table 29: Summary of goodness-of-fit data for dabrafenib and trametinib ToT (BRF113928 Cohort C); standard parametric models

Footnotes:[a] A small AIC or BIC value represents a better goodness of fit.

Abbreviations : AIC: Akaike information criterion; BIC: Bayesian information criterion; ITT: intention-to-treat; ToT: time on treatment.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Figure 12: Dabrafenib and trametinib ToT extrapolations up to ten years (BRF113928 Cohort C)

==> picture [376 x 259] intentionally omitted <==

Abbreviations : KM: Kaplan-Meier; ToT: time on treatment.

In the BRF113928 trial, the median treatment duration was ***** months (range: *** to ** months). To evaluate the long-term plausibility of the chosen extrapolation, UK clinical experts noted that experience of dabrafenib and trametinib in current clinical experience is limited to the interim COVID-19 guidance period.[8] The experts referenced the median treatment durations from the BRF113928 trial as encouraging and noted that it is possible for some patients to have an extended treatment response if they did not experience any AEs. Therefore, the best statistically fitting exponential distribution was considered plausible, and used in the base case analysis.

The next two best statistically-fitting extrapolations, the Weibull and the Gompertz, had similar AIC and BIC estimates to the exponential distribution, and were explored in scenario analyses (Section B.3.10.3). As the dabrafenib and trametinib KM curve was complete, a scenario analysis also explored the use of the KM curve directly to model ToT.

Pembrolizumab plus chemotherapy

The standard parametric distributions were fitted to the weighted ToT IPD for patients receiving pembrolizumab plus chemotherapy in the FLATIRON RWE study (2022). In selecting the most appropriate curve for pembrolizumab plus chemotherapy, it is important to note that a 2-year stopping rule was applied, in line with UK clinical practice, as detailed below.[21]

As such, the parametric model is only required over the first two years of the model time horizon. This means that the use of the KM data directly could represent a plausible option. However, it was considered more appropriate to fit a parametric curve to the observed data, in order to smooth out the stepwise nature of the KM data and assume a more constant rate of discontinuation of treatment over time. A scenario analysis utilising the pembrolizumab plus chemotherapy ToT KM curve directly was also conducted (Section B.3.10.3).

The AIC and BIC values for each of the extrapolations for the pembrolizumab plus chemotherapy

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

BRAF V600E population are summarised in Table 30. Extrapolations of ToT using each model up to 3 years are presented in Figure 13.

Table 30: Summary of goodness-of-fit data for pembrolizumab plus chemotherapy ToT (BRAF V600E [weighted] population – FLATIRON RWE study [2022]); standard parametric models

Footnotes:[a] A small AIC or BIC value represents a better goodness of fit. Abbreviations : AIC: Akaike information criterion; BIC: Bayesian information criterion; ITT: intention-to-treat; ToT time on treatment.

Figure 13: Pembrolizumab plus chemotherapy ToT extrapolations up to three years (BRAF

V600E [weighted] population – FLATIRON RWE Study [2022])*

==> picture [417 x 256] intentionally omitted <==

**Footnotes: *** Extrapolations are presented prior to the application of the stopping rules Abbreviations : KM: Kaplan-Meier; ToT: time on treatment.

The median time to treatment discontinuation in the FLATIRON RWE study (2022) for patients with a V600E mutation receiving pembrolizumab plus chemotherapy was *** months (95% CI: *** **** , ), versus a median duration of treatment of 7.2 months (range: 1 day, 35.4 months) in KEYNOTE-189.[110] As these estimates were closely aligned, the use of the pembrolizumab plus chemotherapy ToT data was considered appropriate.

However, despite the external validity when compared to KEYNOTE-189, the use of the pembrolizumab plus chemotherapy data from the FLATIRON RWE study (2022) was still considered to be associated with some uncertainty, due to the limited follow-up (mean **** months), compared to a mean **** months of follow-up for dabrafenib and trametinib in the

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

BRF113928 trial.

The exponential curve was selected in the base case analysis, as the curve providing the best fit to the observed data. The second and third best statistically fitting curves (lognormal and Gompertz) were explored in scenario analyses. Two further scenarios explored the use of the pembrolizumab plus chemotherapy ToT KM curve directly, and explored setting pembrolizumab plus chemotherapy ToT equal to ToT for dabrafenib and trametinib (but with the relevant stopping rules applied), to align with the assumptions for PFS and OS (Section B.3.10.3).

Stopping rules

Adjustments were made to the ToT extrapolation for pembrolizumab plus chemotherapy to account for the stopping rule imposed in NICE TA683, which stipulates that all treatments should be stopped following two years of treatment, or earlier if the disease progresses.[21] As such, after two years, the number of patients receiving pembrolizumab plus chemotherapy in the economic model was set to zero. Patients were assumed to discontinue carboplatin, cisplatin and pemetrexed earlier than this, based on the dosing regimens for each treatment, detailed in Section B.3.2.4 and B.3.5.1.

B.3.3.3 Adverse events

Given the base case economic analysis assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy in terms of PFS and OS, the robust modelling of differences in the AE profiles of each therapy was considered important. In particular, clinical expert feedback indicated that it would be important for any modelled differences in AEs to reflect the risk of immune-mediated AEs experienced by patients receiving pembrolizumab plus chemotherapy.[8]

As such, in the base case economic analysis, costs and disutility estimates were included for any all-cause Grade ≥3 AEs experienced by ≥1% of patients receiving either dabrafenib and trametinib or pembrolizumab plus chemotherapy.

The proportions of patients modelled to experience these AEs for dabrafenib and trametinib and pembrolizumab plus chemotherapy are shown in Table 31, based on the safety cohort (Cohorts B and C) of the BRF113928 trial for dabrafenib and trametinib (Section B.2.10), and KEYNOTE189 for pembrolizumab plus chemotherapy.[110] Where a Grade ≥3 AE was experienced by ≥1% of patients receiving one therapy, the proportion of patients experiencing that Grade ≥3 AE for the other therapy was included (for example, with pyrexia).

It should be noted that AEs were reported differently between the two treatments. The BRF113928 trial reported AEs reported the incidence of Grade ≥3 AEs in cases where that AE (any Grade) occurred in ≥10% of patients. In comparison, the KEYNOTE-189 trial publication only reports Grade ≥3 AEs where that AE (any Grade) occurred in ≥15% of patients in either treatment group, with the exception of immune-mediated AEs, which were reported in full. While there are likely some differences in AEs that are therefore not captured in the base case, this is unlikely to have any significant impact, since immune-mediated AEs are reported in full.

The disutility and costs associated with the AEs included within the model are detailed in Section B.3.4.4 and Section B.3.5.4, respectively.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 31: Summary of AEs included in the base case economic analysis

Footnotes:[a] Note AEs were not reported separately for Cohort C from the BRF113928 trial thus these data are derived from the combined safety cohort of Cohort B and Cohort C. It is not expected that the AE profile for dabrafenib and trametinib would differ depending on whether patients had previously untreated or previously treated disease. Abbreviations: AEs: adverse events.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

B.3.4 Measurement and valuation health effects

B.3.4.1 Health-related quality-of-life data from clinical trials

This section is not applicable as HRQoL data were not collected within the BRF113928 trial.

B.3.4.2 Mapping

Not applicable.

B.3.4.3 Health-related quality-of-life studies

As detailed in Appendix H.3, no studies were identified in the economic SLRs that reported utility value estimates for patients with NSCLC with a BRAF mutation. As such, utility values for the economic model were derived from previous NICE appraisals in the wider NSCLC population, as detailed in Section B.3.4.7.

B.3.4.4 Adverse reactions

As discussed in Section B.3.3.3, the base case economic analysis included all-cause Grade ≥3 AEs experienced by ≥1% of patients in either the BRF113928 trial for dabrafenib and trametinib or KEYNOTE-189 for pembrolizumab plus chemotherapy. To accurately account for the disutility associated each of these Grade ≥3 AEs, disutility values were sourced from the published literature, alongside the duration that each AE was expected to be experienced for. This represents the most accurate approach to the modelling of AE disutility and ensures that each disutility is applied for the appropriate duration of time.

As one of the most recent appraisals conducted in advanced NSCLC for a targeted therapy, AE disutility estimates were sourced from NICE TA789.[6] The expected duration of each AE, and therefore the duration of the disutility application, was also sourced from NICE TA789, where the majority of AE durations were derived from the pivotal trial for tepotinib (the VISION trial).[6]

Where NICE TA789 did not report an appropriate disutility or duration for an AE of relevance, further NICE NSCLC appraisals or other sources in the published literature were reviewed (Table 32). Beyond that, data from the BRF113928 trial on the duration of AEs were utilised. The BRF113928 trial was not utilised for the duration of all AEs in the first instance as it did not report the duration of all AEs separately, and it did not report the duration for Grade ≥3 AEs only.[17]

In line with NICE TA812 (pralsetinib for treating RET fusion-positive advanced NSCLC), in cases where no disutility data or appropriate durations were able to be sourced from the literature, a disutility and AE duration of 0 was assumed.[7, 104] This was the case for “weight increased” and “weight decreased” and therefore in the base case economic analysis, these AEs were not associated with any disutility. As these AEs were only experienced by 1.08% and 3.23% of patients receiving dabrafenib and trametinib, respectively, it was not considered that this would have a large impact on the base case economic results.

Within the base case economic analysis, each AE disutility was applied as the associated QALY decrement, which was calculated by adjusting the AE disutility (which represents the total QALY loss over one year, if a patient were to experience the AE for one year) by the expected duration of each AE. A summary of the AE disutility estimates and AE durations including in the base case economic analysis is presented in Table 32.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 32: Summary of AE disutility estimates included in the base case economic analysis

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 78 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 79 of 133

Abbreviations: AEs: adverse event; AESI: adverse event of special interest; CDF: Cancer Drugs Fund; NICE: National Institute of Health and Care Excellence; TA: technology appraisal.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 80 of 133

B.3.4.5 Treatment administration disutility

An important difference between dabrafenib and trametinib and pembrolizumab plus chemotherapy relates to the administration of both therapies. Dabrafenib and trametinib are oral therapies and can be taken at home, and therefore represent a more convenient, less painful, and less burdensome method of administration compared to pembrolizumab plus chemotherapy which must be administered via IV infusion and requires a patient to visit a hospital to receive treatment. Indeed, the ability for patients to receive treatment away from a hospital setting was one of the reasons behind dabrafenib and trametinib being provided through interim measures by NHS England during the COVID-19 pandemic.[20]

Pembrolizumab plus chemotherapy requires three separate IV infusions (pembrolizumab, carboplatin/cisplatin, and pemetrexed), which would be expected to last for at least 60 minutes in total (pembrolizumab: 30 minutes; carboplatin: 15–60 minutes or cisplatin: 120 minutes; pemetrexed: 10 minutes). Within the first 12 weeks, these therapies are administered every 3 weeks. As such, it is plausible that there may be a disutility associated with an IV regimen (versus an oral treatment) due to this burden of administration, and this should be captured in the model.

In the base case economic analysis, a disutility of -0.023 was applied to the PFS health state utility value for patients receiving pembrolizumab plus chemotherapy. This disutility was applied for the duration that patients remained on treatment, defined by the pembrolizumab plus chemotherapy ToT curve.

The disutility of -0.023 was based on the study by Matza et al. (2013),[128] which investigated the disutility associated with infusion-treatments for bone metastases. The disutility of -0.023 was reflective of patients receiving a treatment with a 30-minute IV infusion regimen every four weeks, and the question used to elicit the utility values focussed on the mode of administration of the treatment.

Matza et al. (2013) was previously used in NICE TA728 to support a disutility associated with IV infusions.[129] Another recent NICE appraisal in advanced NSCLC (NICE TA781) applied a similar disutility of -0.025 in all cycles that patients were on treatment receiving docetaxel monotherapy via IV infusion.[7] This disutility was derived from a cost-effectiveness study of erlotinib versus docetaxel, and reported utility values of 0.451 versus 0.426 for oral versus IV therapy, respectively, representing a difference of -0.025.[130] These utilities were determined by having a sample of the UK general population fill out a visual analogue scale (VAS) and not EQ-5D. However, despite the limitations of the data, the Committee agreed that a disutility may be plausible for IV infusion, and considered different scenarios in their decision making.[7]

In order to explore the uncertainty surrounding administration-related disutility, two alternative scenarios were considered.

First, as patients receiving cisplatin as part of their pembrolizumab plus chemotherapy regimen are likely to receive IV infusions that would last longer than 30 minutes, an increased disutility of - 0.037 was applied for patients receiving pembrolizumab, pemetrexed and cisplatin (15.4% of the overall pembrolizumab plus chemotherapy cohort), for the first 4 treatment cycles (12 model cycles). This estimate was also taken from Matza et al. (2013), and represents a disutility associated with a 2-hour infusion.[128] As outlined in NICE TA181, cisplatin is associated with hydration requirements that require patients to be brought in the night before treatment, which is

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

likely to be associated with a further reduction in utility.[117] In this scenario, beyond 12 model cycles (and for all cycles for patients receiving pembrolizumab, carboplatin and pemetrexed), the original IV infusion disutility of -0.023 was applied for all patients remaining on treatment.

The second scenario used the disutility of -0.037 to reflect the complexity of the pembrolizumab plus chemotherapy regimen and, in the first 4 treatment cycles (12 model cycles), this was applied to all patients receiving pembrolizumab plus chemotherapy but during the model cycles that they would receive an infusion only. Beyond 12 model cycles, the original disutility of -0.023 was applied only in the cycles that patients would receive an infusion with pembrolizumab (with or without pemetrexed). While Matza et al. (2013) note that the value of -0.023 represents a “health state disutility” for patients receiving treatment with a 30-minute IV infusion regimen every 4 weeks, this assumption was relaxed in this scenario, to consider the impact of the frequency and intensity of IV infusions reducing over time as the number of IV administrations is decreased beyond the initial 4 treatment cycles.

B.3.4.6 Age-adjusted disutility values

In the base case economic analysis, health state utility values were also age-adjusted over the model time horizon using UK population norm values for EQ-5D as reported in the HSE 2014 dataset by the NICE DSU.[131]

B.3.4.7 Health-related quality-of-life data used in the cost-effectiveness analysis

Health state utility values

In the absence of HRQoL data from the BRF113928 trial, it was necessary to derive health state utility value estimates from the published literature in order to inform the health state utility values for PFS and PD in the base case economic analysis. As the economic SLR for utility values did not identify any relevant utility data in patients with advanced NSCLC and a BRAF mutation, utility values were sourced from alternative advanced NSCLC population data utilised in previous NICE appraisals.

Given the base case assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy in terms of PFS and OS, it is important to note that the choice of health state utility value has no impact on the incremental QALYs accrued for dabrafenib and trametinib versus pembrolizumab plus chemotherapy, because patients spend the same length of time in the PFS and PD health states in both treatment arms. As such, while the choice of health state utility value does impact the total QALYs accrued in the model, it has no impact on the base case ICER.

Nevertheless, to identify the most appropriate source of health state utility values, the three most recently published NICE appraisals for a targeted therapy in advanced NSCLC were reviewed: NICE TA789, TA781 and TA812.[6, 7, 104]

NICE TA781 utilised time-to-death utilities based on their pivotal trial which had collected EQ-5D data.[7] As this is a different approach to the use of health state utility values utilised in this economic analysis, the utility values from NICE TA781 were not considered further.

NICE TA789 utilised utility values derived from EQ-5D data collected in their pivotal trial in the

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

base case, with scenario analyses based on the published literature and a number of other prior appraisals.[6] NICE TA812 did not collect EQ-5D data in their pivotal trial and therefore utilised utility values from prior appraisals only: NICE TA654 in the base case, and NICE TA310 and NICE TA643 in scenario analysis.[104, 107, 132, 133]

Given the absence of any clear preferred source of utility estimates, the base case analysis used the utility values from Chouaid et al. ( 2013) as an assumption (Table 33).[54] These utilities were used in a scenario in NICE TA789, and are considered generalisable to UK clinical practice, as the study was a cross-sectional, multi-site study that prospectively measured health states in advanced NSCLC with 263 patients from 25 centres including the UK using EQ-5D and EQVAS.[54]

Abbreviations : PD: progressed disease; PFS: progression-free survival; SD: standard deviation. Source: Chouaid et al. (2013)[54]

Scenario analyses exploring alternative utility estimates were not considered, given that health state utility values do not impact the incremental QALYs between dabrafenib and trametinib and pembrolizumab plus chemotherapy, given the assumption of clinical equivalence.

B.3.5 Cost and healthcare resource use identification,

measurement and valuation

The base case economic analysis was conducted from an NHS and PSS perspective and therefore included only costs that would be incurred by the NHS and PSS. Appropriate sources of unit costs, such as NHS reference costs 2019–20, the British National Formulary (BNF) online, and the electronic Marketing Information Tool (eMIT) (2021), were used for cost inputs in the model.

Specifically, the following costs components were considered in the base case economic analysis: drug acquisition and administration costs for dabrafenib and trametinib, pembrolizumab plus chemotherapy, drug acquisition and administration costs for subsequent therapies, follow-up and monitoring costs (based on health state), AE costs and terminal care costs.

B.3.5.1 Intervention and comparators’ costs and resource use

Intervention

Dabrafenib and trametinib was modelled as dabrafenib 150 mg twice daily and trametinib 2 mg twice daily, in line with their respective SmPCs. [11, 12]

As described in Section B.3.3.2, treatment with dabrafenib and trametinib was modelled in line with the ToT data collected within the BRF113928 trial. As such, patients were assumed to discontinue treatment with dabrafenib and trametinib in line with the treatment discontinuation observed within the BRF113928 trial.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Comparators

Pembrolizumab plus chemotherapy was modelled as a weighted average, based on NICE TA789[6] :

• Pembrolizumab with pemetrexed and carboplatin (84.4%)

• Pembrolizumab with pemetrexed and cisplatin (15.6%)

Full details of the dosing regimens modelled for pembrolizumab plus chemotherapy are listed in Section B.3.2.4.

Based on these dosing regimens, a number of stopping rules were applied to carboplatin, cisplatin and pemetrexed, in addition to the two-year stopping rule applied to pembrolizumab plus chemotherapy.[21] Platinum-based chemotherapy (carboplatin and cisplatin) taken as part of this treatment combination is only given for a maximum number of four three-weekly treatment cycles (Section B.3.2.4). As such, patients in the pembrolizumab plus chemotherapy arm of the model were assumed to discontinue carboplatin or cisplatin in the model after four three-weekly treatment cycles (12 weekly model cycles).

It was also assumed that some patients discontinued pemetrexed after four three-weekly treatment cycles (12 model cycles). Of patients on treatment after 12 model cycles, ****% of patients were assumed to receive maintenance treatment with pemetrexed after this point (Section B.3.2.4).

This is aligned with clinical expert opinion in NICE TA789, which indicated that between 50–60% of patients go onto receive pemetrexed maintenance therapy.[6] This approach was also aligned with the KEYNOTE-189 trial, where patients receive pemetrexed and platinum-based chemotherapy in combination with pembrolizumab every 3 weeks for four cycles, followed by pemetrexed maintenance plus pemetrexed for up to 35 three-weekly cycles.[110] A post-hoc analysis of the KEYNOTE-189 trial showed that the median number of three weekly treatment cycles of pemetrexed was 11, ranging from 5 to 30.[134] Given that all patients received at least one cycle of maintenance pemetrexed in KEYNOTE-189, the base case assumption that only ****% of patients receive maintenance pemetrexed is likely to be conservative.[6]

Drug acquisition costs for all treatments were derived from the dosing regimens as detailed in Table 36, with the pack costs taken from the British National Formulary for branded medicines (BNF 2022) or electronic market information tool for generic medicines (eMIT 2021) as outlined in Table 34 and Table 35. Where treatment dosing was dependent on patient weight and/or BSA, drug acquisition costs were calculated based on average dose requirements for the population under evaluation in the model, using patient baseline characteristics from Cohort C of the BRF113928 trial (Section B.3.2.3).

No vial sharing was assumed in the base case economic analysis and where multiple vial/package sizes were available, the cheapest price per mg was applied. Once the drug acquisition costs for dabrafenib and trametinib and pembrolizumab plus chemotherapy were calculated from the respective ToT curves, treatment-specific relative dose intensities (RDIs) were subsequently applied to derive the final drug acquisition costs. RDIs were based on mean daily doses from the BRF113928 trial (detailed in Appendix F.1) and the KEYNOTE-189 trial, respectively.[72, 110]

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Finally, it should be noted that a confidential patient access scheme exists for pembrolizumab (as well as atezolizumab, nivolumab and nintedanib that are considered within the model as subsequent therapies – see Section B.3.5.5) that is unknown to Novartis. As such, two sets of cost-effectiveness results have been presented within this submission.

The first set of results uses the currently approved PAS prices for dabrafenib (*** *** ********) and trametinib (*** *** ********), and the list prices for all comparator and subsequent treatments. These results are presented in the relevant sections of Section B.3 below.

A second set of results has been conducted, where Novartis have made assumptions regarding the PAS discount for pembrolizumab, as well as all subsequent treatments, in order to provide a more indicative set of results. Novartis have assumed that pembrolizumab is associated with a PAS discount of **%, and for completeness, have also assumed that atezolizumab and nivolumab are associated with PAS discounts of **%, and nintedanib is associated with a PAS *** ************ ***** ******** discount of . For these analyses, dabrafenib is provided at the ** ******* ****** ** * ******** ** ****** *********** ** ******** ** * *** ********). These results are provided

in a confidential appendix, Appendix O, alongside the main submission dossier.

The tables below provide details of the costs used in the first set of results, where dabrafenib is provided at a *** PAS discount, and all comparators and subsequent treatments are provided at list price.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 34: Drug unit and pack costs (dabrafenib and trametinib)

Abbreviations: BNF: British National Formulary; eMIT: electronic Market Information Tool.

Table 36: Summary of drug acquisition costs for the treatments included in the base case economic analysis

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 86 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 87 of 133

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 88 of 133

Drug Dosing regimen Packs/ RDI Cost per Mean Mean Proportion of Source vials dose dose cost per patients per per cycle receiving each dose cycle regimen pemetrexed and carboplatin Footnotes:[a ] For the first 12 model cycles, all patients receive pembrolizumab once every three weeks. After 12 model cycles, ****% of patients receive pembrolizumab (alongside pemetrexed), at a dose of 200 mg, once every three weeks. The remaining ****% of patients receive pembrolizumab, at a dose of 400 mg, once every six weeks. Abbreviations: AUC: area under the curve; RDI: relative dose intensity; SmPC: Summary of Product Characteristics; TA: technology appraisal.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 89 of 133

B.3.5.2 Drug administration costs

Drug administration costs were considered within the model to reflect the fact that pembrolizumab plus chemotherapy is administered by IV infusion and therefore requires patients to attend hospital appointments to receive treatment. Dabrafenib and trametinib is administered orally and was not assumed to be associated with any administration costs. This assumption is aligned with a number of previous oncology NICE appraisals, where oral treatments have not been assumed to incur any administration costs, including NICE TA810 (abemaciclib with endocrine therapy for adjuvant treatment of hormone receptor-positive, HER2-negative, nodepositive early breast cancer at high risk of recurrence).[138]

The unit administration costs associated with pembrolizumab plus chemotherapy were derived from NHS reference costs (2019/2020) and are presented in Table 37.[139] The frequency of administration for the pembrolizumab plus chemotherapy regimens included in the model are outlined in Table 38.

In previous relevant NICE NSCLC submissions (such as NICE TA683[21] ), to differentiate between the administration costs for regimens with carboplatin and cisplatin, the unit costs per cycle for regimens with carboplatin were considered ‘complex chemotherapy, including prolonged infusional treatment – outpatient’ while the costs for regimens with cisplatin were taken as ‘complex chemotherapy, including prolonged infusional treatment – day case and regular day/night’.[21] The justification for this was that cisplatin has hydration requirements that requires patients to be brought in the night before treatment, as outlined in NICE TA181 (Pemetrexed for the first-line treatment of NSCLC).[117]

It should also be noted that one administration cost was applied per model cycle, regardless of the number of therapies being received via IV infusion in that cycle. As such in Table 38 below, some treatments administered via IV infusion as part of pembrolizumab plus chemotherapy are listed as having administration costs of £0.00 in particular cycles – for example, pembrolizumab has an administration cost of £0.00 in Cycles 0–2. This does not mean that pembrolizumab is not associated with an administration cost, but reflects the fact that the administration cost of pembrolizumab is assumed to be captured within the administration costs used for carboplatin/cisplatin within these cycles.

Table 37: Unit costs of drug administration used in the base case economic analysis

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Abbreviations: NHS: National Health Service; IV: intravenous.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 38. Summary of drug administration costs for pembrolizumab plus chemotherapy included in the base case economic analysis (comparators)

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 92 of 133

Footnotes :[a ] For the first 12 model cycles, all patients receive pembrolizumab once every three weeks. After 12 model cycles, ****% of patients receive pembrolizumab (alongside pemetrexed), at a dose of 200 mg, once every three weeks. The remaining ****% of patients receive pembrolizumab, at a dose of 400 mg, once every six weeks. Abbreviations : AUC: area under the curve; IV: intravenous; NSCLC: non-small cell lung cancer; RDI: relative dose intensity; SmPC: summary of product characteristics; TA: technology appraisal.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation. © Novartis Pharmaceuticals UK Ltd (2022). All rights reserved. Page 93 of 133

B.3.5.3 Health state unit costs and resource use

Healthcare resource use costs (for example, the costs of follow-up and monitoring services and consultations) were modelled as per-cycle costs, with separate values for the PFS and PD health states. This aligns with prior NICE submissions in advanced NSCLC.[21, 59] The per-cycle costs (detailed in Table 39) were calculated by multiplying the unit costs of services (detailed in Table 40) by the mean frequency of resource use (detailed in Table 41). Unit costs were based on NHS reference costs (2019/20),[139] and the frequency of resource use was based on NICE TA789.[6]

Separate resource use frequencies, and resulting costs per cycle, were considered in the PFS and PD health states. However, given the base case assumption of clinical equivalence between dabrafenib and trametinib and pembrolizumab plus chemotherapy in terms of PFS and OS, in the base case economic analysis, patients in both arms of the model are assumed to reside in the PFS and PD health states for the same duration of time. As a result, in the base case economic analysis, healthcare resource use costs were assumed to be the same for both dabrafenib and trametinib and pembrolizumab plus chemotherapy (i.e., they effectively cancel out). Nevertheless, it was still considered important to model the costs associated with follow-up and monitoring services and consultations accurately within the model.

Table 39: Summary of the resource use cost per cycle

Abbreviations : PD: progressed disease; PFS: progression-free survival.

Table 40. Resource use unit costs

Abbreviations: ECG: electrocardiogram; HRG: healthcare resource group; NHS: national health service; PSSRU: personal social services research unit.

Company evidence submission template for dabrafenib and trametinib in advanced NSCLC with a BRAF V600 mutation.

© Novartis Pharmaceuticals UK Ltd (2022). All rights reserved.

Table 41. Resource use frequencies