TA886 · STA

Olaparib for adjuvant treatment of BRCA mutation-positive HER2-negative high-risk early breast cancer after chemotherapy

Recommended with restrictionsApril 2023

Only recommended if the company provides it according to the commercial arrangement (commercial access agreement with confidential discount)

Source documents

Intervention

olaparib (Lynparza)
PARP inhibitor · PARP inhibitor · oral

Condition

brca mutation-positive her2-negative high-risk early breast canceroncology · adjuvant

Comparators

NameType Established Committee preferred
routine monitoring for cancer recurrenceno treatmentYes
placeboplacebo
routine monitoringstandard of care

Clinical trials

TrialDesignPhasePivotal
OlympiARCTPhase 3Yes

Economic model

semi markov (company)
Time horizon: lifetime
Cycle length: not specified

ICER

Below £20,000 (olaparib vs routine monitoring) · not specified uncertainty
£20,000–£30,000 (olaparib vs routine monitoring) · moderate uncertainty

Methodological decisions (10)

cost assumption

Company assumed BRCA testing costs were zero, arguing most people with HER2-negative high-risk early breast cancer undergo routine BRCA screening. EAG questioned this for hormone receptor-positive HER2-negative population.

Company: No additional BRCA testing costs as routine BRCA screening already performed for eligible patients under National Genetic Test Directory criteria

ERG: BRCA testing costs should be included for hormone receptor-positive HER2-negative population

Committee: Company position accepted; most patients with HER2-negative early breast cancer at high risk meet current testing criteria

ICER impact: negligible

cure assumption

Risk of recurrence assumption in triple-negative breast cancer population after 5 years

Company: 0% chance of recurrence after 5 years

ERG: 5% risk from year 5 to year 15

Committee: 2% to 3% risk of recurrence between years 5 and 8, while uncertain, was reasonable

ICER impact: uncertain direction

discount rate

Company proposed 1.5% discount rates for costs and outcomes for triple-negative population instead of standard 3.5%, arguing immature long-term data.

Company: Discount rates of 1.5% for costs and outcomes in triple-negative population

ERG: Standard 3.5% discount rates appropriate; immaturity and uncertainty of data means olaparib unlikely to restore full health or provide sustained benefits to justify reduced rate

Committee: Standard 3.5% discount rates apply; olaparib does not meet criteria for reduced rate (75.4% of placebo arm had no invasive DFS event by 4 years)

ICER impact: increases

population generalisability

Committee noted concern about company's utility values being unrealistically high and not consistent with other TA estimates for triple-negative breast cancer. Also questioned whether mapped values from OlympiA were plausible.

Company: Utility values estimated from OlympiA trial data are appropriate

Committee: Company's estimates implausibly high; disease-free value (0.869) only slightly lower than age-matched general population (0.877), unrealistic for patients post-surgery and chemotherapy

ICER impact: decreases

survival extrapolation

Choice of distribution for extrapolating recurrence in hormone receptor-positive HER2-negative breast cancer population beyond trial data

Company: log-normal distribution

ERG: generalised gamma distribution

Committee: Insufficient evidence to determine; noted clinical experts' opinion that little difference between distributions for long-term invasive disease-free survival

ICER impact: small effect on ICERs

survival extrapolation

Choice of distribution for extrapolating survival after early metastatic recurrence

Company: exponential distribution

ERG: Gompertz distribution (due to non-proportional hazards)

Committee: Did not discuss further given small effect on ICER

ICER impact: small effect on ICER

utility source

Source and values for health-related quality of life in health states, particularly disease-free state

Company: EORTC QLQ-C30 data from OlympiA mapped to EQ-5D-3L using Crott and Briggs algorithm; utility values 0.869 (disease-free and non-metastatic), 0.685 (metastatic)

ERG: Verrill et al. 2020 UK study in HER2-positive breast cancer population; utility values 0.732 (disease-free), 0.668 (non-metastatic), 0.603 (metastatic); with age-adjustment: 0.770, 0.702, 0.634

Committee: EAG's age-adjusted utility values from Verrill et al. were most appropriate

ICER impact: increases

utility source

Company used EQ-5D-3L mapped from EORTC QLQ-C30 in OlympiA. EAG used UK study (Verrill et al. 2020) with direct EQ-5D measurement in HER2-positive breast cancer populations.

Company: Health-related quality-of-life data from EORTC QLQ-C30 in OlympiA, mapped to EQ-5D-3L; utility values 0.869 (disease-free and non-metastatic), 0.685 (metastatic)

ERG: External UK study (Verrill et al. 2020) with age-adjusted estimates: 0.770 (disease-free), 0.702 (non-metastatic), 0.634 (metastatic)

Committee: EAG's age-adjusted estimates from Verrill et al. were most appropriate

ICER impact: decreases

utility value choice

Whether to use same utility value for disease-free and non-metastatic breast cancer (local recurrence) health states

Company: Same utility value (0.869) for both disease-free and non-metastatic breast cancer health states based on non-significant difference in trial

ERG: Different utilities warranted

Committee: Different utilities should be used; same utility value was unrealistic given anxiety of local recurrence and probable further surgery

ICER impact: increases

utility value choice

Company assumed non-metastatic breast cancer utility equal to disease-free utility (difference non-significant). Committee concerned this was unrealistic given anxiety associated with local recurrence and possible further surgery.

Company: Non-metastatic breast cancer utility set equal to disease-free utility value (0.869) as difference was non-significant

Committee: Separate utility value for non-metastatic breast cancer reflecting anxiety and probable further surgery

ICER impact: decreases

Evidence gaps

short follow upNo long-term data from OlympiA beyond 4 years to support extrapolation of recurrence and survival; median follow-up 3.5 years at latest data cut-off
immature overall survivalOverall survival benefit at 4 years (3.4% difference, 95% CI -0.1 to 6.8%) not statistically significant; overall survival data immature

Commercial arrangement

commercial access agreement · confidential · critical for recommendation

Special considerations

Cancer Drugs Fund eligible Innovation acknowledged