TA962 · STA

olaparib for maintenance treatment of newly diagnosed BRCA mutation-positive advanced ovarian, fallopian tube or peritoneal cancer after response to first-line platinum-based chemotherapy

Recommended for Cancer Drugs FundTechnology Appraisal Committee AJuly 2019

Recommended only if the conditions in the managed access agreement for olaparib are followed. FIGO stages 3 and 4 only. BRCA mutation-positive (germline and/or somatic). Treatment continued until radiological disease progression, unacceptable toxicity, or for up to 2 years if no radiological evidence of disease.

Source documents

Intervention

olaparib (Lynparza)
PARP inhibitor · PARP inhibitor · oral

Conditions

advanced ovarian canceroncology · newly_diagnosed
fallopian tube canceroncology · newly_diagnosed
primary peritoneal canceroncology · newly_diagnosed
brca mutation-positive advanced ovarian canceroncology · metastatic
brca mutation-positive fallopian tube canceroncology · metastatic
brca mutation-positive peritoneal canceroncology · metastatic

Comparators

NameType Established Committee preferred
placeboplacebo
routine surveillancestandard of careYes

Clinical trials

TrialDesignPhasePivotal
SOLO-1RCTIIIYes
Study 19RCT
GOG-172RCT
JGOG-3016RCT

Economic model

partitioned survival (company)
Time horizon: 20 years (implied from text stating 20% more alive at 20 years)
Cycle length: not stated

ICER

Below £20,000 (olaparib vs routine surveillance) · very_high uncertainty

Methodological decisions (13)

cost assumption

Price per tablet of olaparib is the same regardless of dose, so cost of treatment per day for reduced dose is same as full dose. Economic modelling should use cost of whole tablets rather than average cost per milligram.

Committee: use cost of whole tablets

ICER impact: negligible

cost assumption

Price per tablet is same regardless of dose, so reduced-dose cost equals full-dose cost. After technical engagement, modelling revised to use cost of whole tablets rather than average cost per milligram.

Company: Initially used average cost per milligram; revised to whole tablets after TE

Committee: Use whole tablet costs

ICER impact: increases

cure assumption

Company's model assumes approximately 20% of people are cured from ovarian cancer. Clinical experts explained this is plausible (Study 19 shows 10% disease-free after 10 years, termed 'super responders' more likely from complete responses), but no OS benefit yet demonstrated in SOLO-1.

Company: 20% cure assumption is plausible and should be included

Committee: Cure is possible and estimate is plausible, but uncertainties remain without OS benefit demonstration in SOLO-1

ICER impact: increases

discount rate

Olaparib for first-line maintenance treatment does not meet criteria for 1.5% discount rates. Reference case 3.5% discount rates should be applied in cost-effectiveness analyses.

Committee: 3.5% discount rates

ICER impact: increases

model structure

Company presented a partitioned survival model with 3 health states (progression-free, progressed disease and death). ERG commented on model structure limitations.

Company: 3-state partitioned survival model

ERG: raised limitations with model structure

Committee: 3-state partitioned survival model acceptable for decision making

ICER impact: negligible

model structure

Company submitted 4-health state partitioned survival model with PFS-2 state. ERG considered 3-state model oversimplified and suggested sequenced model would be more appropriate, but acknowledged lack of data to populate multiple therapy lines.

Company: 4-health state partitioned survival model with PFS-2 is appropriate

ERG: Sequenced model would be more appropriate but lacks data; 3-state model was oversimplified

Committee: 4-state model is acceptable for decision making, though main limitations (immature OS data, complexity of pathway) would persist with any model structure

ICER impact: uncertain_direction

population generalisability

Committee noted 82% of SOLO-1 had complete response to platinum chemotherapy, but estimates NHS percentage at ~70% of responders. Clinical experts indicated complete response patients likely to have better long-term outcomes. Additionally, Edinburgh OvCa Database lacks information on response type, limiting comparability.

Company: 82% complete response rate in SOLO-1 is representative

Committee: NHS proportion of complete responders estimated at 70%, lower than trial's 82%. People with residual disease expected to have worse outcomes.

ICER impact: increases

stopping rule

Treatment with olaparib should be stopped after 2 years unless there is evidence of residual disease. Committee considered approximately 15% of patients would continue beyond 2 years due to residual disease, though this estimate is uncertain and may be optimistic.

Committee: approximately 15% continue beyond 2 years

ICER impact: increases

stopping rule

Uncertainty about percentage of patients eligible to continue olaparib beyond 2 years (if residual disease and meet continuation criteria). Committee estimated ~15% vs company's 10%, with large impact on treatment costs.

Company: Approximately 10% of patients eligible to continue beyond 2 years

Committee: Approximately 15% could continue in UK clinical practice, though this estimate may be optimistic due to worse UK outcomes

ICER impact: decreases

survival extrapolation

Company used PFS-2 as surrogate for OS in modelling, predicting 20% survival difference at 20 years. However, SOLO-1 shows OS curve convergence at ~40 months, unlike expectations. Committee noted this is a major weakness; model does not reflect trial's interim OS results.

Company: PFS-2 is appropriate surrogate; immature OS data would give unrealistic routine surveillance OS estimates. Model predicts OS gains will emerge with longer follow-up, consistent with Study 19 pattern.

ERG: Not explicitly stated but concerned model uses extrapolated PFS-2, not trial OS data

Committee: OS modelling is very uncertain and may overestimate survival gain; not possible to resolve until more mature OS data available from SOLO-1

ICER impact: increases

treatment effect duration

Committee discussed uncertainty about how long olaparib treatment benefit will extend, particularly given convergence of OS curves at early SOLO-1 data cuts. Linked to immature OS evidence.

Company: Treatment benefit expected to persist based on Study 19 pattern and achieved PFS gains

Committee: Duration uncertain; cannot assume early curve convergence will reverse with longer follow-up

ICER impact: uncertain_direction

treatment effect waning

Whether progression-free survival benefit from olaparib will translate into overall-survival benefit. Study 19 shows similar pattern of convergence after 37 months, then OS benefit observed after 78 months. Relationship may be 1:1 (Sundar et al.) or 1:2 or more (GOG-172, JGOG-3016).

Committee: expected that treatment with olaparib will extend life but extent is uncertain

ICER impact: uncertain_direction

treatment sequencing

Uncertainty about use of PARP inhibitors after second-line platinum-based chemotherapy and about retreatment with PARP inhibitor. SOLO-1 had subsequent PARP inhibitor use in both arms; committee noted uncertainty about whether trial reflects UK practice.

Company: Subsequent PARP inhibitor use in SOLO-1 is reasonable reflection of UK clinical practice

Committee: There is uncertainty around use of PARP inhibitors after second-line therapy; this has large impact on cost-effectiveness

ICER impact: uncertain_direction

Evidence gaps

immature overall survivalOverall survival was a secondary endpoint in SOLO-1. At 21% maturity, the median was not reached in either arm. No statistically significant benefit observed. Immature data make cost-effectiveness estimates very uncertain.
surrogate not validatedThe relationship between second progression-free survival (PFS-2) and overall survival is not established. While PFS-2 may be a reasonable surrogate, clear relationship has not been established in the absence of mature OS data.
no direct comparisonNo evidence for retreatment with a PARP inhibitor. It is unknown whether tumour sensitivity to PARP inhibitors would be retained after subsequent chemotherapy.
immature overall survivalOverall-survival data from SOLO-1 are immature. After 78 months follow-up an OS benefit was observed but the extent to which PFS benefit translates to OS benefit is unknown. The company's model uses PFS-2 as a surrogate for OS, not actual OS data.
short follow upAt the latest data cut, convergence of OS curves observed after about 40 months follow-up. However, Study 19 showed convergence at early cuts with OS gains observed after several years. It is unknown whether SOLO-1 will mirror this pattern.

Commercial arrangement

managed access agreement · confidential · critical for recommendation

Special considerations

Cancer Drugs Fund eligible Innovation acknowledged

Cross-references

TA381precedent — NICE Technology Appraisal guidance 381 recommends olaparib capsules for BRCA mutation-positive advanced ovarian cancer after 3 or more lines of platinum-based chemotherapy