TA1110 · STA

Abiraterone for treating newly diagnosed high-risk hormone-sensitive metastatic prostate cancer

Not recommendedAugust 2021

Source documents

Intervention

abiraterone (Zytiga)
androgen receptor signalling inhibitor · androgen synthesis inhibitor · oral

Condition

newly diagnosed high-risk hormone-sensitive metastatic prostate canceroncology · metastatic

Comparators

NameType Established Committee preferred
androgen deprivation therapy (adt) alonestandard of careYes
docetaxel plus adt plus prednisoloneactive drugYes
adt alonestandard of careYes
docetaxel in combinationactive drugYesYes
enzalutamide plus adtactive drugYes

Clinical trials

TrialDesignPhasePivotal
LATITUDERCTnot statedYes
STAMPEDERCTnot statedYes

Economic model

partitioned survival (company)
Time horizon: not specified in this section
Cycle length: not stated

ICER

Above £100,000 (abiraterone in combination vs docetaxel in combination) · high uncertainty

Methodological decisions (20)

comparator selection

Committee recognized need to consider separate analysis for population where docetaxel is contraindicated or unsuitable, but acknowledged that defining this population clinically is complex

Committee: For whole population: both ADT alone and docetaxel in combination are appropriate comparators; separate consideration possible for chemotherapy-ineligible group if cost-effective for whole population not demonstrated

comparator selection

Whether ADT alone is an appropriate comparator for people for whom docetaxel is contraindicated or unsuitable versus the general population

Company: presented single modelled comparison with ADT alone for the docetaxel-contraindicated population

Committee: concluded company's modelled comparison with ADT alone was not a valid estimate for people for whom docetaxel is contraindicated or unsuitable but used it in absence of specific data

ICER impact: increases

cost assumption

Modelling of costs and benefits of follow-on treatments in hormone-relapsed phase

Company: Company modelled treatment pathways based on opinion of 4 clinicians regarding NHS market shares; assumed 80% had abiraterone or enzalutamide after progression; modelled costs of multiple follow-on treatments

ERG: not stated

Committee: Committee noted mismatch between modelled treatments and proportions in trials; noted model accounted for costs but potentially not all life-extending benefits; concluded accounting for costs but not benefits could bias ICER upwards

ICER impact: increases

equivalence assumption

Whether abiraterone in combination and docetaxel in combination can be assumed equivalent in overall survival

Company: Company used indirect comparison with hazard ratio from NMA (academic in confidence)

ERG: not stated

Committee: Committee considered useful scenario assuming equal hazard of overall survival (HR 1.00) between abiraterone in combination and docetaxel; also considered scenario using HR 1.13 from STAMPEDE direct comparison metastatic subgroup

ICER impact: uncertain_direction

indirect comparison method

Company developed network meta-analysis to compare abiraterone with docetaxel in combination, incorporating STAMPEDE and additional trials, arguing STAMPEDE subgroup analysis was underpowered and did not reflect licensed population

Company: Network meta-analysis incorporating multiple trials is more appropriate than STAMPEDE subgroup analysis

Committee: Direct data from STAMPEDE is preferred for comparing abiraterone with docetaxel

indirect comparison method

Company developed network meta-analysis for abiraterone vs docetaxel comparison, incorporating LATITUDE, STAMPEDE, CHAARTED, and GETUG-AFU 15. Committee preferred direct comparison from STAMPEDE metastatic subgroup.

Company: Network meta-analysis should be used as it includes multiple trials and contributes additional information on treatment effect

Committee: Direct randomised comparison from STAMPEDE is preferred as less likely to be biased

ICER impact: uncertain_direction

model structure

Original multistate Markov model did not provide plausible estimates. Company revised to partitioned survival model.

Committee: Partitioned survival model is appropriate

ICER impact: negligible

model structure

Choice between Markov vs partitioned survival model structure

Company: Company initially submitted multistate Markov model; revised to partitioned survival model after committee feedback

ERG: not stated

Committee: Committee deemed Markov model did not provide plausible estimates of post-progression or overall survival and did not generate valid cost-effectiveness estimates; agreed partitioned survival model was appropriate

ICER impact: uncertain_direction

population generalisability

LATITUDE and STAMPEDE included only people with adequate haematological function and ECOG/WHO performance status 0-2. No specific data for people for whom docetaxel is contraindicated or unsuitable.

Company: Treatment effect from whole LATITUDE population can be used to represent people for whom docetaxel is contraindicated or unsuitable

Committee: Data specific to docetaxel-ineligible population is preferred, but acknowledges unavailability and uses LATITUDE whole population with uncertainty acknowledged

ICER impact: increases

population generalisability

Whether trial data from LATITUDE (whole population including those able to have docetaxel) can be generalised to people for whom docetaxel is contraindicated or unsuitable

Company: Company considered results of LATITUDE could be generalised to people who cannot or should not have docetaxel

ERG: not stated

Committee: Committee agreed it was not possible to determine the size of effectiveness for people for whom docetaxel is contraindicated or unsuitable; not presented with data on effectiveness in this subgroup

ICER impact: uncertain_direction

population generalisability

Whether LATITUDE trial population data represent people for whom docetaxel is contraindicated or unsuitable. Committee acknowledged considerable uncertainty when using whole LATITUDE population to estimate cost-effectiveness for this subgroup due to lack of specific data

Company: presented modelled comparison with ADT using whole LATITUDE population

ERG: presented scenarios removing chemotherapy follow-on treatment and associated survival benefit to model the docetaxel-contraindicated subgroup

Committee: acknowledged uncertainty and noted that even in company's base case and ERG scenarios, ICERs for this subgroup exceeded £30,000 per QALY

ICER impact: increases

proportional hazards

Whether hazards of progression and death are proportional between arms

Company: Company agreed hazards were not proportional

ERG: not stated

Committee: Committee agreed with company that hazards were not proportional; appropriate to fit curves to each arm separately

ICER impact: negligible

survival extrapolation

Choice of parametric distribution for extrapolating progression-free and overall survival from LATITUDE

Company: Company presented both log-logistic (considered plausible but optimistic) and Weibull (considered plausible but pessimistic) distributions

ERG: ERG highlighted that model assumed treatment effect maintained long-term; may wane in clinical practice; provided scenarios for adjustment

Committee: Weibull distribution broadly appropriate for progression-free survival; log-logistic distribution plausible for overall survival in full trial population. Would have preferred further extrapolations for subgroup of people unable to have docetaxel

ICER impact: uncertain_direction

survival extrapolation

Committee considered scenarios for overall survival hazard ratios for abiraterone in combination compared with docetaxel in combination: direct comparison HR of 1.13 from STAMPEDE metastatic subgroup, company's indirect comparison (confidential), and assumption of equal hazards at various time points

Company: presented indirect comparison with confidential HR

Committee: approved use of direct comparison HR of 1.13 from STAMPEDE metastatic subgroup and equal hazards scenarios to account for potential treatment waning

ICER impact: uncertain_direction

treatment effect waning

Whether treatment effect of abiraterone compared with comparators is maintained long-term or wanes over time

Company: Company model assumed treatment effect maintained over long term

ERG: ERG provided scenarios to adjust for potential treatment waning

Committee: Committee noted longer-term STAMPEDE data (8-year follow-up) supported ongoing benefit; agreed scenarios exploring alternative time points for treatment waning were useful

ICER impact: decreases

treatment sequencing

Use of abiraterone in first-line hormone-sensitive disease limits follow-on treatment options because abiraterone and enzalutamide are commissioned only once in treatment pathway, whereas docetaxel or ADT-first approaches preserve more options

Committee: First-choice treatment affects follow-on treatments; abiraterone in combination limits options compared to ADT alone or docetaxel in combination

treatment sequencing

Trials included follow-on treatments (docetaxel, abiraterone, enzalutamide for hormone-relapsed disease) that would not be available to people who cannot have docetaxel, but this was reflected in trial survival estimates.

Company: LATITUDE results can be generalised to docetaxel-ineligible people

Committee: LATITUDE survival estimates include docetaxel-based follow-on treatments not applicable to docetaxel-ineligible patients; STAMPEDE estimates more relevant to NHS practice

ICER impact: increases

treatment sequencing

Whether treatment pathway in LATITUDE (including docetaxel for hormone-relapsed disease) reflects real-world pathway for people who cannot have docetaxel

Company: Company based modelling on LATITUDE results assuming follow-on docetaxel use

ERG: not stated

Committee: Committee concluded that overall survival estimates from LATITUDE included effect of docetaxel in hormone-relapsed phase which would not apply to people who cannot have docetaxel; preferred a model reflecting actual treatment pathway

ICER impact: increases

treatment sequencing

Modelling of subsequent treatment (chemotherapy as follow-on) and associated survival benefit for people in whom docetaxel is contraindicated or unsuitable

Company: included chemotherapy follow-on treatment in base case

ERG: removed chemotherapy follow-on treatment and survival benefit in scenario analyses

Committee: acknowledged validity concern with company's approach for docetaxel-contraindicated population but used it as best available evidence

ICER impact: uncertain_direction

utility source

Source of health-related quality of life data for utility values in the model

Company: Company used EQ-5D from LATITUDE for ADT and abiraterone in combination; used separate preference study (general public) for docetaxel disutility

ERG: ERG derived docetaxel disutility from EQ-5D data in STAMPEDE and carried out scenario; stated consistent with company approach

Committee: Committee preferred EQ-5D data from STAMPEDE metastatic subgroup (high-risk hormone-sensitive) where data available for all three arms (abiraterone in combination, docetaxel in combination, ADT alone). In absence of company access to STAMPEDE EQ-5D, company's and ERG's approaches were broadly consistent and acceptable

ICER impact: negligible

Evidence gaps

no direct comparisonNo clinical evidence for abiraterone plus ADT compared with ADT alone in the subgroup of people for whom docetaxel is contraindicated or unsuitable
no direct comparisonNo specific data available for people for whom docetaxel is contraindicated or unsuitable. LATITUDE and STAMPEDE only included people with adequate performance status and haematological function who could theoretically receive docetaxel.
short follow upUncertainty about magnitude of long-term survival gain with abiraterone due to potential differences in follow-on treatment patterns between trials and NHS clinical practice

Commercial arrangement

simple discount pas · confidential · critical for recommendation

Special considerations

Innovation acknowledged Equality issues raised

Cross-references

comparator guidance — Reference to NICE technology appraisal on radium-223 dichloride for treating hormone-relapsed prostate cancer with bone metastases regarding approach to defining populations for whom chemotherapy is contraindicated or unsuitable