← Clinical Trials
Metastatic Castration-resistant Prostate Cancer Phase 3 Time-to-Event Analysis NCT03732820

PROpel: Complete Statistical Analysis of Olaparib Plus Abiraterone in Metastatic Castration-resistant Prostate Cancer

An independent statistical review of the randomized phase 3 PROpel trial evaluating olaparib plus abiraterone acetate versus placebo plus abiraterone acetate as first-line therapy in men with metastatic castration-resistant prostate cancer.

PROpel  ·  Phase 3  ·  Randomized, parallel, quadruple-masked  ·  NCT03732820
Scope of this record

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

PROpel is a randomized phase 3 trial in men with metastatic castration-resistant prostate cancer. The study compares olaparib plus abiraterone acetate with placebo plus abiraterone acetate and uses time-to-event methodology, with radiological progression-free survival (rPFS) as the registered primary endpoint.

895
Enrollment
2 randomized arms
3
Phase
Phase 3
0.66
Primary rPFS HR
95% CI 0.54–0.81
<0.0001
Primary p-value
Two-sided
FeaturePROpel
Trial namePROpel
NCT IDNCT03732820
Brief titleStudy on Olaparib Plus Abiraterone as First-line Therapy in Men With Metastatic Castration-resistant Prostate Cancer
PhasePhase 3
ConditionMetastatic Castration-resistant Prostate Cancer
AllocationRandomized
Design modelParallel
MaskingQuadruple
Primary purposeTreatment
Enrollment895
Lead sponsorAstraZeneca
Sponsor typeIndustry
StatusActive, not recruiting
Start2018-10-31
Primary completion2021-07-30

2. Clinical Question

The central question is whether adding olaparib to abiraterone acetate, compared with placebo plus abiraterone acetate, changes the time to radiological progression or death in men receiving first-line therapy for metastatic castration-resistant prostate cancer.

Population

Men with metastatic castration-resistant prostate cancer, as specified by the trial's brief title and registered condition.

Intervention

Olaparib 300 mg twice daily plus abiraterone 1000 mg once daily.

Comparator

Placebo twice daily plus abiraterone 1000 mg once daily.

Primary question

Does olaparib plus abiraterone improve radiological progression-free survival relative to placebo plus abiraterone?

3. Trial Design

01
Randomize895 enrolled
02
Two armsOlaparib vs placebo
03
Background therapyAbiraterone 1000 mg qd
04
Time-to-eventrPFS and secondary endpoints
05
AnalysisStratified log-rank + Cox HR
Allocation
Randomized
Participants were allocated to 2 parallel treatment arms.
Masking
Quadruple masked
The registry identifies the study as quadruple masked.
Primary purpose
Treatment
The registered primary purpose is treatment.
Primary endpoint type
Time-to-event
The registered primary endpoint is radiological progression-free survival.
INTERVENTION

Olaparib + abiraterone

  • Olaparib 300 mg twice daily
  • Abiraterone 1000 mg once daily
COMPARATOR

Placebo + abiraterone

  • Placebo twice daily
  • Abiraterone 1000 mg once daily

The available trial data identify the study as randomized, parallel, and quadruple masked. They do not provide a randomization ratio, treatment discontinuation rules, crossover provisions, or factorial structure, so those design features are not assumed here.

4. Endpoints

Primary Endpoint

EndpointRegistry definition and time frameAnalysis
Number of Participants With Radiological Progression Free Survival (rPFS) Event by Investigator Assessment An rPFS event is defined as progression determined by Response Evaluation Criteria in Solid Tumours version 1.1 [RECIST 1.1] and/or Prostate Cancer Working Group 3 [PCWG-3] or death (by any cause in the absence of progression), regardless of whether the patient withdraws from randomised therapy or receives another anticancer therapy prior to progression. Per RECIST v1.1, progression is defined.

Assessed from date of randomisation to data cut off (DCO1): 30Jul2021 (Approx. 2 years 9 months).
Stratified log-rank test; hazard ratio and 95% CI calculated using a Cox proportional-hazards model.
Endpoint structure: rPFS combines radiological progression and death into a single time-to-event outcome. A participant can therefore experience the endpoint either through documented progression or through death in the absence of progression, according to the registered definition.

Secondary Endpoints With Posted Statistical Analyses

Secondary endpointTime frameAnalysis population
Number of Participants With Overall Survival (OS) EventDCO3: 12Oct2022 (Approx. 3 years 11 months)Full Analysis Set (FAS)
Number of Participants With Time to First Subsequent Anticancer Therapy or Death (TFST) EventDCO3: 12Oct2022 (Approx. 3 years 11 months)Full Analysis Set (FAS)
Number of Participants With Time to Pain Progression (TTPP) EventDCO3: 12Oct2022 (Approx. 3 years 11 months)Full Analysis Set (FAS)
Number of Participants With Opiate UseDCO3: 12Oct2022 (Approx. 3 years 11 months)Randomised patients who are not on opiates at baseline
Number of Participants With First Symptomatic Skeletal Related Event (SSRE)DCO3: 12Oct2022 (Approx. 3 years 11 months)Full Analysis Set (FAS)
Number of Participants With Second Progression or Death (PFS2) EventDCO3: 12Oct2022 (Approx. 3 years 11 months)Full Analysis Set (FAS)

The registry data identify 9 outcome measures posted and 7 statistical analyses posted. The ClinicalTrials.gov record contains one primary statistical analysis and six secondary statistical analyses.

5. Statistical Methodology

Time-to-event analysis

The primary endpoint and all six registry-reported secondary statistical analyses are time-to-event analyses. Rather than simply counting how many participants experience an event, time-to-event analysis incorporates the timing of the event and accommodates participants whose event time is not observed during the relevant follow-up period through censoring.

Stratified log-rank test

The reported comparison uses a log-rank test stratified by Metastases and Docetaxel at metastatic hormone-sensitive prostate cancer stage, using a pre-specified pooling strategy. The analysis notes state that these stratification factors are the same as those used in the stratified log-rank test.

Reported comparison framework
Stratified log-rank test  →  time-to-event comparison between randomized groups

The log-rank component tests whether the observed event-time experience differs between the two treatment groups while incorporating the specified stratification structure.

Cox proportional-hazards model

The registry states that the hazard ratio and confidence interval were calculated using a Cox proportional-hazards model. For the analyses posted on ClinicalTrials.gov, a hazard ratio below 1 favors olaparib plus abiraterone, according to the registry analysis notes.

Hazard ratio
HR = estimated relative hazard of the event: olaparib + abiraterone vs placebo + abiraterone

A hazard ratio of 0.66, for example, corresponds to an estimated hazard 34% lower in the olaparib-plus-abiraterone group under the fitted model. It is not a probability, a percentage of patients cured, or an absolute difference in event rates.

Full Analysis Set

The primary rPFS analysis and the registry-reported OS, TFST, TTPP, SSRE, and PFS2 analyses use the Full Analysis Set (FAS). The opiate-use analysis instead uses randomized patients who are not on opiates at baseline.

Superiority hypothesis

The registered hypothesis type for the analyses posted on ClinicalTrials.gov is superiority. This means the statistical question is framed around whether the randomized treatment groups differ in the direction favoring the intervention, rather than whether the intervention is merely not unacceptably worse than the comparator.

6. Statistical Methods Explained

Why was a log-rank test used?

Because the primary outcome is time-to-event, the analysis must account for both whether an event occurs and when it occurs. The log-rank test is designed to compare event-time distributions between treatment groups while incorporating censored observations. In PROpel, the reported test is stratified according to the prespecified pooling factors described in the registry analysis notes.

What does an rPFS hazard ratio of 0.66 mean?

The estimated HR of 0.66 means that the modeled instantaneous hazard of an rPFS event was 0.66 times that in the comparator group. Equivalently, the estimated hazard was 34% lower under the fitted model. This is a relative time-to-event measure; it does not mean that 34% of participants avoided progression or death.

Why is the confidence interval important?

The 95% confidence interval of 0.54 to 0.81 describes statistical uncertainty around the estimated hazard ratio. It gives information about the precision of the estimate that a point estimate alone cannot provide. The interval is about uncertainty in the estimated treatment effect under the analysis framework, not a range containing the effects experienced by individual patients.

What does the p-value add?

The p-value addresses evidence against the null hypothesis under the specified statistical test. It does not measure the size of the treatment effect. For rPFS, the reported p-value is <0.0001, while the magnitude of the estimated relative effect is conveyed by the HR of 0.66 and its 95% CI of 0.54 to 0.81.

Why use stratification?

Stratification allows the time-to-event comparison to account for prespecified factors that can be associated with prognosis or treatment balance. In PROpel, the registry-reported analysis notes identify Metastases and Docetaxel at metastatic hormone-sensitive prostate cancer stage as the stratification factors, with a pre-specified pooling strategy.

Why is the Cox model reported alongside the log-rank test?

The log-rank test supplies the reported hypothesis-testing framework for the time-to-event comparison, while the Cox proportional-hazards model provides the hazard ratio and its confidence interval. These quantities answer related but different questions: the test addresses evidence for a difference, whereas the HR quantifies the estimated relative hazard.

7. Results: Primary Endpoint — Radiological Progression-Free Survival

The primary endpoint was the number of participants with an rPFS event by investigator assessment, assessed from randomisation to DCO1 on 30Jul2021, approximately 2 years 9 months. The primary analysis used the Full Analysis Set.

Radiological progression-free survival

HR 0.66

95% CI: 0.54–0.81   ·   P < 0.0001

Olaparib 300 mg bd + abiraterone 1000 mg qd vs placebo bd + abiraterone 1000 mg qd

Primary result componentReported value
EndpointNumber of Participants With Radiological Progression Free Survival (rPFS) Event by Investigator Assessment
Analysis populationFull Analysis Set (FAS)
MethodLog Rank; stratified analysis
Effect measureHazard Ratio (HR)
Estimate0.66
95% CI0.54–0.81
Confidence intervalTwo-sided, 95%
P-value<0.0001
HypothesisSuperiority
Data cutoff30Jul2021 (DCO1)
Clinical Biostats interpretation

The HR of 0.66 indicates an estimated 34% lower instantaneous hazard of an rPFS event for olaparib plus abiraterone relative to placebo plus abiraterone under the fitted Cox model. The endpoint includes radiological progression determined by RECIST 1.1 and/or PCWG-3 and death in the absence of progression, according to the registry definition.

The HR does not mean that 34% of participants were protected from progression or death, nor does it provide an absolute difference in the probability of remaining progression-free at a particular time.

The 95% CI of 0.54–0.81 provides the precision information surrounding the HR estimate. It does not describe individual-patient variability or the range of outcomes that a particular participant should expect.

The p-value of <0.0001 addresses the statistical evidence under the reported log-rank testing framework. It does not measure the magnitude or clinical importance of the treatment effect; that information comes from the HR and its confidence interval.

Because the HR was calculated with a Cox proportional-hazards model, interpretation relies on the model's proportional-hazards framework. The ClinicalTrials.gov record does not report a formal assessment of that assumption, so no independent conclusion about its adequacy is made here.

Analysis population: The primary rPFS analysis was performed in the Full Analysis Set. This is important when interpreting the estimate because the treatment comparison remains tied to the randomized trial population rather than being restricted to participants who remained on treatment.

8. Results: Secondary Time-to-Event Endpoints

The ClinicalTrials.gov record reports six secondary statistical analyses. Each uses a log-rank test, with stratification by Metastases and Docetaxel at metastatic hormone-sensitive prostate cancer stage using a pre-specified pooling strategy, and each reports a Cox-model hazard ratio with a two-sided 95% confidence interval.

Secondary endpointHR95% CIP-value
Overall Survival (OS) Event0.810.67–1.000.0544
Time to First Subsequent Anticancer Therapy or Death (TFST) Event0.760.64–0.900.0025
Time to Pain Progression (TTPP) Event1.060.75–1.500.7456
Opiate Use1.210.82–1.790.3099
First Symptomatic Skeletal Related Event (SSRE)0.820.55–1.220.3212
Second Progression or Death (PFS2) Event0.760.59–0.990.0534

Overall Survival

Hazard ratio for OS event

0.81

95% CI: 0.67–1.00   ·   P = 0.0544

Assessed from randomisation to DCO3: 12Oct2022 (Approx. 3 years 11 months)

Clinical Biostats interpretation

The estimated HR of 0.81 corresponds to an estimated 19% lower instantaneous hazard of an OS event with olaparib plus abiraterone under the fitted model. The 95% CI extends from 0.67 to 1.00, so the interval reaches the null value of 1.00.

The p-value of 0.0544 is close to 0.05, but a p-value should not be treated as a measure of effect magnitude. The effect estimate and confidence interval provide the more direct description of the estimated relative treatment effect and its uncertainty.

The OS analysis was performed in the Full Analysis Set and used the same stratified log-rank and Cox-model framework described above. The ClinicalTrials.gov record does not provide median OS, event counts, or time-specific survival probabilities, so those quantities are not presented here.

Time to First Subsequent Anticancer Therapy or Death

Hazard ratio for TFST event

0.76

95% CI: 0.64–0.90   ·   P = 0.0025

Assessed from randomisation to DCO3: 12Oct2022 (Approx. 3 years 11 months)

Clinical Biostats interpretation

The HR of 0.76 corresponds to an estimated 24% lower instantaneous hazard of the TFST event for olaparib plus abiraterone under the fitted model. The confidence interval of 0.64–0.90 describes uncertainty around that relative estimate.

TFST is distinct from OS and rPFS. It incorporates the occurrence of a first subsequent anticancer therapy or death, so it captures a later treatment-pathway event rather than only the initial radiological progression endpoint.

The p-value of 0.0025 provides statistical evidence under the reported test, but it does not quantify the size of the treatment effect. The HR and confidence interval are required to understand magnitude and precision.

Time to Pain Progression

Hazard ratio for TTPP event

1.06

95% CI: 0.75–1.50   ·   P = 0.7456

Assessed from randomisation to DCO3: 12Oct2022 (Approx. 3 years 11 months)

Clinical Biostats interpretation

The estimated HR of 1.06 corresponds to an estimated 6% higher instantaneous hazard of a TTPP event for olaparib plus abiraterone under the fitted model. The confidence interval is 0.75–1.50, spanning both values below and above 1.00.

The p-value of 0.7456 does not establish evidence of a treatment-group difference under the reported test. It also does not prove that the two treatments have identical effects. The confidence interval is important because it describes the uncertainty surrounding the estimate.

Opiate Use

Hazard ratio for opiate-use event

1.21

95% CI: 0.82–1.79   ·   P = 0.3099

Analysis population: randomized patients who are not on opiates at baseline

Clinical Biostats interpretation

The HR of 1.21 represents an estimated 21% higher instantaneous hazard of the reported opiate-use event in the olaparib-plus-abiraterone group under the fitted model. The 95% CI of 0.82–1.79 includes 1.00, indicating substantial uncertainty around the direction and magnitude of the relative effect.

This endpoint uses a different analysis population from the other registry-reported secondary endpoints: randomized patients who were not on opiates at baseline. That distinction matters because the analysis is conditional on baseline opiate status rather than simply the overall Full Analysis Set.

First Symptomatic Skeletal Related Event

Hazard ratio for SSRE

0.82

95% CI: 0.55–1.22   ·   P = 0.3212

Assessed from randomisation to DCO3: 12Oct2022 (Approx. 3 years 11 months)

Clinical Biostats interpretation

The estimated HR of 0.82 corresponds to an estimated 18% lower instantaneous hazard of a first symptomatic skeletal related event under the fitted model. However, the 95% CI of 0.55–1.22 crosses 1.00, so the point estimate should not be interpreted without its uncertainty interval.

The p-value of 0.3212 does not measure treatment effect size. The HR and confidence interval are the quantities that describe the estimated direction, magnitude, and precision of the treatment comparison.

Second Progression or Death

Hazard ratio for PFS2 event

0.76

95% CI: 0.59–0.99   ·   P = 0.0534

Assessed from randomisation to DCO3: 12Oct2022 (Approx. 3 years 11 months)

Clinical Biostats interpretation

The HR of 0.76 corresponds to an estimated 24% lower instantaneous hazard of a second progression or death under the fitted model. The 95% CI of 0.59–0.99 indicates the uncertainty around that estimate and remains below 1.00 at its upper bound.

The p-value of 0.0534 is close to 0.05. It should not be interpreted as an effect-size measure, and the PFS2 result should be distinguished from the registered primary rPFS endpoint because PFS2 is a secondary endpoint in the registry-reported analysis set.

9. Results in Statistical Context

The ClinicalTrials.gov record shows that the estimated hazard ratios are not identical across endpoints. That is expected because each endpoint defines a different event process and, in one case, a different analysis population.

EndpointEstimated HRDirection under registry convention95% CIP-value
Primary rPFS0.66Favors olaparib + abiraterone0.54–0.81<0.0001
OS0.81Favors olaparib + abiraterone0.67–1.000.0544
TFST0.76Favors olaparib + abiraterone0.64–0.900.0025
TTPP1.06Favors placebo + abiraterone under registry convention0.75–1.500.7456
Opiate Use1.21Favors placebo + abiraterone under registry convention0.82–1.790.3099
SSRE0.82Favors olaparib + abiraterone0.55–1.220.3212
PFS20.76Favors olaparib + abiraterone0.59–0.990.0534

This table should be read horizontally rather than as a ranking of endpoints. A hazard ratio describes the relative event hazard for one specific endpoint, and the confidence interval and analysis population determine how much uncertainty surrounds that estimate.

Do not equate p-values with clinical importance. A smaller p-value does not necessarily correspond to a larger treatment effect. For example, the TFST analysis reports an HR of 0.76 with P = 0.0025, while the primary rPFS analysis reports an HR of 0.66 with P < 0.0001. The endpoints, event definitions, follow-up, and statistical quantities are different, so their p-values should not be used to create an efficacy ranking.

10. Stratification and the Analysis Model

The registry-reported analysis notes specify a stratified log-rank test based on Metastases and Docetaxel at metastatic hormone-sensitive prostate cancer stage, using a pre-specified pooling strategy.

ComponentPROpel analysis
Primary endpointrPFS event by investigator assessment
Testing methodLog-rank test
StratificationMetastases and Docetaxel at metastatic hormone-sensitive prostate cancer stage
PoolingPre-specified pooling strategy
Effect measureHazard ratio
Model for HR and CICox proportional-hazards model
Direction of HRHR < 1 favors olaparib + abiraterone

Using the same stratification factors in the log-rank and Cox analyses provides consistency between the hypothesis-testing and effect-estimation components of the analysis. The registry also identifies intention-to-treat analysis and stratified analysis among the concepts represented in the primary analysis text.

11. Censoring and Time-to-Event Interpretation

A time-to-event analysis must distinguish between participants who experience the event during follow-up and participants whose event time is not observed during the analysis period. The latter participants contribute information up to the point at which their follow-up is censored.

For PROpel's rPFS endpoint, the registry definition also specifies that an event can remain applicable regardless of whether a patient withdraws from randomized therapy or receives another anticancer therapy prior to progression. This is an important feature of the registered endpoint definition because the event definition is not simply "progression while still receiving randomized treatment."

Conceptual survival function
S(t) = P(T > t)

For a time-to-event endpoint, the survival function represents the probability of remaining event-free beyond time t. Kaplan-Meier estimation is a standard nonparametric way to estimate this function in the presence of right censoring.

The statistical analyses posted on ClinicalTrials.gov do not report Kaplan-Meier estimates, median event times, numbers at risk, or reconstructed survival curves. Those quantities are therefore not added to this page.

12. Safety

The ClinicalTrials.gov record provides serious adverse-event counts by arm. These figures are reported as affected participants over the corresponding at-risk population.

Serious adverse eventsAffected / at risk
Olaparib 300 mg bd + Abiraterone 1000 mg161/398
Placebo bd + Abiraterone 1000 mg qd126/396

The ClinicalTrials.gov record does not provide a formal statistical comparison, confidence interval, or p-value for serious adverse events. Accordingly, the safety data are presented descriptively rather than as an inferential comparison.

Why this matters: serious adverse events are safety outcomes, while rPFS, OS, TFST, TTPP, opiate use, SSRE, and PFS2 are time-to-event efficacy or clinical-event endpoints. Their analysis frameworks and interpretation should not be conflated.

13. Trial Timeline

31 Oct 2018

Trial start

The registry lists 2018-10-31 as the study start date.

30 Jul 2021

Primary completion and DCO1

The registry lists 2021-07-30 as primary completion and uses the same date as DCO1 for the primary rPFS analysis, described as approximately 2 years 9 months from the relevant starting point.

12 Oct 2022

DCO3

The registry-reported secondary analyses use 12Oct2022 as DCO3, described as approximately 3 years 11 months.

Current registry status

Active, not recruiting

The ClinicalTrials.gov record identifies the study status as ACTIVE_NOT_RECRUITING.

14. Limitations

15. Why This Trial Matters Statistically

PROpel is a useful teaching case because the ClinicalTrials.gov record illustrate how a single randomized trial can generate a family of related time-to-event analyses without making the endpoints statistically interchangeable.

Statistical conceptHow it appears in PROpel
RandomizationThe trial uses randomized allocation in a parallel phase 3 design.
BlindingThe registry identifies the study as quadruple masked.
Time-to-event endpointsThe primary endpoint and registry-reported secondary statistical analyses are time-to-event outcomes.
Log-rank testThe reported statistical method for the primary and registry-reported secondary comparisons.
Stratified analysisThe log-rank tests are stratified by Metastases and Docetaxel at metastatic hormone-sensitive prostate cancer stage using a pre-specified pooling strategy.
Hazard ratioThe primary analysis reports HR 0.66, while secondary analyses range from 0.76 to 1.21 in the ClinicalTrials.gov record.
Confidence intervalsEach registry-reported effect estimate includes a two-sided 95% CI.
P-valuesEach registry-reported formal analysis reports a p-value, allowing the evidence from the specified test to be separated from effect magnitude.
Full Analysis SetThe primary rPFS and most secondary analyses use the FAS.
Endpoint-specific populationsThe opiate-use analysis is restricted to randomized patients who are not on opiates at baseline.

The most important statistical lesson is that the HR, confidence interval, p-value, endpoint definition, follow-up period, and analysis population must be interpreted together. Looking only at the p-value or only at the HR can obscure important differences in what each analysis actually measures.

16. A Practical Framework for Reading PROpel Results

1. Identify the event

First determine exactly what counts as the event. For rPFS, the registered definition includes radiological progression and death in the absence of progression.

2. Identify the time origin

The registry-reported time-to-event analyses assess outcomes from the date of randomisation to specified data cutoffs.

3. Identify the population

Most registry-reported analyses use the Full Analysis Set; the opiate-use analysis uses a more specific randomized population.

4. Read the HR

The HR describes the relative event hazard under the Cox model. HR < 1 favors olaparib plus abiraterone according to the registry convention.

5. Read the confidence interval

The interval describes uncertainty around the estimated HR and should be considered alongside the point estimate.

6. Read the p-value last

The p-value describes evidence under the reported test. It does not quantify the size of the treatment effect or its clinical importance.

17. Statistical Concepts in This Trial

Learn more about the methods used in this trial:

18. Related Statistical Calculators

19. Sources

Continue through the Clinical Biostats statistical pathway

Use the trial's endpoints and methods as a practical starting point for learning survival analysis, hazard ratios, confidence intervals, randomization, and related clinical-trial methods.

20. Record Summary

PROpel is a randomized, parallel, quadruple-masked phase 3 trial with 895 enrolled participants and a registered primary time-to-event endpoint of radiological progression-free survival. The registry-reported primary analysis used a stratified log-rank test in the Full Analysis Set and reported a hazard ratio of 0.66 with a two-sided 95% CI of 0.54–0.81 and a p-value of <0.0001. The registry-reported secondary analyses extend the same time-to-event framework to overall survival, time to first subsequent anticancer therapy or death, time to pain progression, opiate use, first symptomatic skeletal related event, and second progression or death.

The statistical interpretation depends on more than the primary p-value. The event definition, data cutoff, analysis population, stratification factors, Cox-model hazard ratio, confidence interval, and distinction between primary and secondary endpoints all contribute to understanding what the reported results actually establish.

Clinical Biostats methodology: This page separates reported numerical results from statistical interpretation and does not add trial estimates, event counts, follow-up measures, or design features that are not contained in the ClinicalTrials.gov record.