← Clinical Trials
Metastatic Prostate Cancer Phase 3 Time-to-Event Analysis NCT02489318

TITAN: Complete Statistical Analysis of Apalutamide Plus ADT in mHSPC

An independent statistical review of the randomized phase 3 TITAN trial evaluating apalutamide plus androgen deprivation therapy (ADT) versus ADT in participants with metastatic hormone-sensitive prostate cancer.

TITAN  ·  Phase 3  ·  Randomized, parallel-group, triple-masked  ·  Enrollment 1052
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. The numerical results on this page are restricted to the statistical analyses posted in the ClinicalTrials.gov record.

1. Trial at a Glance

TITAN is a randomized phase 3, parallel-group trial evaluating apalutamide plus androgen deprivation therapy versus androgen deprivation therapy in participants with metastatic hormone-sensitive prostate cancer. The registry reports two primary time-to-event endpoints and four secondary time-to-event analyses, all comparing the randomized groups using log-rank methodology and hazard ratios.

1052
Enrolled
2 treatment arms
3
Masking level
Triple-masked
0.484
rPFS HR
95% CI 0.391–0.600
0.651
OS HR
95% CI 0.534–0.793
FeatureTITAN
Trial nameTITAN
PhasePhase 3
ConditionProstate Cancer
PopulationParticipants with metastatic hormone-sensitive prostate cancer (mHSPC)
DesignRandomized, parallel-group, triple-masked
Primary purposeTreatment
Enrollment1052
Arms2
Primary endpointsRadiographic progression-free survival (rPFS) and overall survival (OS)
Primary endpoint typeTime-to-event
Primary analysis methodLog-rank test
Effect measureHazard ratio
Hypothesis typeSuperiority
StatusActive, not recruiting
Start2015-11-27
Primary completion2020-09-07
Lead sponsorAragon Pharmaceuticals, Inc.
Sponsor typeIndustry

2. Clinical Question

The central statistical question is whether adding apalutamide to androgen deprivation therapy improves time-to-event outcomes compared with androgen deprivation therapy alone in participants with metastatic hormone-sensitive prostate cancer.

Population

Participants with metastatic hormone-sensitive prostate cancer (mHSPC).

Intervention

Apalutamide plus androgen deprivation therapy (ADT).

Comparator

Placebo plus androgen deprivation therapy (ADT).

Primary question

Does apalutamide plus ADT improve radiographic progression-free survival and overall survival relative to placebo plus ADT?

3. Trial Design

01
Randomize1052 participants
02
Assign2 parallel treatment groups
03
MaskTriple-masked design
04
AssessTime-to-event endpoints
05
CompareLog-rank test and hazard ratio
TREATMENT ARM

Apalutamide + ADT

  • Apalutamide
  • Androgen deprivation therapy (ADT)
CONTROL ARM

Placebo + ADT

  • Placebo
  • Androgen deprivation therapy (ADT)

The trial uses a parallel-group randomized design. This means each participant is assigned to one treatment strategy and the groups are followed for the prespecified outcomes. Randomization provides the foundation for comparing outcomes according to assigned treatment rather than according to treatment actually received.

The registry classifies the study as triple-masked. Masking can reduce the influence of knowledge of treatment assignment on participant behavior, investigator assessment, and other aspects of trial conduct. For an endpoint such as investigator-assessed radiographic progression, masking is particularly relevant because the timing of an event can otherwise be vulnerable to assessment expectations.

Design interpretation: The registry identifies TITAN as randomized, parallel, triple-masked, and treatment-purpose. It does not provide, in the ClinicalTrials.gov record, a randomization ratio, stratification factors, interim-analysis specifications, alpha-spending method, or imputation strategy. Those design features are therefore not reconstructed here.

4. Endpoints

EndpointTime frameRegistry definitionEndpoint type
Radiographic Progression-free Survival (rPFS) Up to 35 months rPFS as assessed by the investigator was defined as the duration from the date of randomization to the date of first documentation of radiographic progressive disease or death due to any cause, whichever occurred first. Radiographic progressive disease was defined as progression of soft tissue lesions measured by computed tomography (CT) or magnetic resonance imaging (MRI) as defined by modified Response evaluation criteria in solid tumors (RECIST) 1.1. Time-to-event
Overall Survival (OS) Up to 57 months OS was defined as the time from date of randomization to date of death from any cause. Time-to-event

Both primary endpoints begin at randomization. That common time origin is important statistically because the treatment comparison is anchored to the randomized allocation. The endpoints differ in the event being observed: rPFS reaches an event at radiographic progression or death, whereas OS reaches an event only at death from any cause.

rPFS

Combines radiographic progression and death into a time-to-first-event endpoint. A participant can therefore experience an rPFS event without having died.

OS

Uses death from any cause as the event. Participants who remain alive at the relevant follow-up are treated as censored observations for survival analysis.

5. Primary Results: Radiographic Progression-free Survival

The registry reports a formal superiority analysis of rPFS in the ITT population. Participants were classified according to their assigned treatment group regardless of the actual treatment received.

Hazard ratio for radiographic progression or death

0.484

95% CI: 0.391–0.600   ·   P <0.0001

Analysis time frame: Up to 35 months

EndpointComparisonAnalysis populationMethodEffect estimate95% CIP-value
Radiographic Progression-free Survival (rPFS) Placebo + ADT vs Apalutamide + ADT ITT Log-rank test HR 0.484 0.391–0.600 <0.0001
Clinical Biostats interpretation

An HR of 0.484 means that the estimated instantaneous rate of an rPFS event was about 48% of the corresponding rate in the comparison group under the reported time-to-event analysis. Equivalently, the estimate corresponds to an approximately 52% lower estimated hazard for the apalutamide-plus-ADT group relative to placebo plus ADT.

The hazard ratio does not mean that 52% of participants avoided progression, nor does it mean that every participant experienced exactly a 52% reduction in risk. It is a relative time-to-event measure summarizing the treatment comparison over follow-up.

The 95% CI of 0.391–0.600 describes statistical uncertainty around the estimated hazard ratio. It does not describe the range of outcomes for individual participants. The entire interval is below 1, which is consistent with the direction of the reported superiority comparison.

The P-value of <0.0001 addresses the strength of evidence against the relevant null hypothesis under the specified statistical framework. It is not a measure of the magnitude or clinical importance of the treatment effect. Effect size is conveyed by the hazard ratio and its confidence interval.

Because this is a time-to-event analysis, interpretation also depends on censoring and the assumptions underlying the survival model used to express the hazard ratio. The ClinicalTrials.gov record identifies the log-rank test and hazard ratio but does not provide a proportional-hazards diagnostic or additional model specification.

Why the rPFS result is a time-to-event result

rPFS is not simply the percentage of participants who eventually progressed. The analysis incorporates when progression or death occurs and accommodates participants whose event status is not observed by the end of their available follow-up. This is why a hazard ratio is a natural effect measure for the endpoint.

Conceptual interpretation
HR = 0.484  →  estimated event hazard is 48.4% of the comparison hazard

The complementary interpretation is approximately a 51.6% relative reduction in the estimated instantaneous hazard, calculated as 1 − 0.484.

6. Primary Results: Overall Survival

The second primary endpoint was overall survival, defined from randomization to death from any cause. The registry reports a formal superiority analysis in the ITT population.

Hazard ratio for death

0.651

95% CI: 0.534–0.793   ·   P <0.0001

Analysis time frame: Up to 57 months

EndpointComparisonAnalysis populationMethodEffect estimate95% CIP-value
Overall Survival (OS) Placebo + ADT vs Apalutamide + ADT ITT Log-rank test HR 0.651 0.534–0.793 <0.0001
Clinical Biostats interpretation

An OS HR of 0.651 means that the estimated instantaneous rate of death in the apalutamide-plus-ADT group was approximately 65.1% of that in the placebo-plus-ADT group under the reported analysis. The complementary interpretation is an approximately 34.9% lower estimated hazard of death.

This does not mean that 34.9% of participants survived or that an individual participant's probability of death was reduced by exactly 34.9%. A hazard ratio is a relative time-to-event measure, not an absolute survival probability.

The 95% CI of 0.534–0.793 quantifies uncertainty around the estimated hazard ratio. It remains below 1 throughout the reported interval, so the point estimate and its uncertainty interval are both in the direction of lower estimated mortality hazard for apalutamide plus ADT.

The P-value of <0.0001 indicates strong statistical evidence against the corresponding null hypothesis under the reported superiority analysis. It should not be interpreted as the probability that the null hypothesis is true, nor as a measure of how large the treatment effect is.

OS is particularly important statistically because the endpoint is death from any cause rather than an investigator-assessed disease event. Nevertheless, the hazard-ratio interpretation remains model- and censoring-dependent, and the ClinicalTrials.gov record does not report additional information about proportional-hazards diagnostics or alternative survival estimands.

Comparing the two primary hazard ratios

Primary endpointHR95% CIP-valueTime frame
rPFS0.4840.391–0.600<0.0001Up to 35 months
OS0.6510.534–0.793<0.0001Up to 57 months

The two estimates address different events and use different reported time frames, so they should not be treated as interchangeable measures. The rPFS endpoint captures radiographic progression or death, while OS captures death from any cause. A treatment effect on rPFS can therefore be observed without requiring the same numerical hazard ratio for OS.

7. Secondary Endpoint Results

The registry also reports four secondary time-to-event analyses. All use the ITT population, compare placebo plus ADT with apalutamide plus ADT, and use the log-rank test with hazard ratio as the effect measure. Each analysis is reported as a superiority hypothesis.

Secondary endpointTime frameHR95% CIP-value
Time to Initiation of Cytotoxic Chemotherapy Up to 57 months 0.469 0.350–0.630 <.0001
Time to Pain Progression Up to 57 months 0.868 0.700–1.076 0.1966
Time to Chronic Opioid Use Up to 57 months 0.794 0.576–1.094 0.1563
Time to Skeletal-related Event (SRE) Up to 57 months 0.857 0.615–1.194 0.3608

Time to Initiation of Cytotoxic Chemotherapy

Hazard ratio

0.469

95% CI: 0.350–0.630   ·   P <.0001

The HR of 0.469 corresponds to an estimated instantaneous rate of initiation of cytotoxic chemotherapy of about 46.9% of the comparison-group rate under the reported analysis. The complementary interpretation is approximately a 53.1% lower estimated hazard of the event. The confidence interval remains below 1, while the P-value is <.0001.

Time to Pain Progression

Hazard ratio

0.868

95% CI: 0.700–1.076   ·   P = 0.1966

The point estimate corresponds to an estimated event hazard about 86.8% of the comparison-group hazard, but the 95% CI extends from 0.700 to 1.076. The reported P-value is 0.1966. The appropriate statistical description is therefore that the registry reports an estimated hazard ratio below 1 with an uncertainty interval that includes 1; the ClinicalTrials.gov record does not establish a statistically significant superiority result for this secondary endpoint.

Time to Chronic Opioid Use

Hazard ratio

0.794

95% CI: 0.576–1.094   ·   P = 0.1563

The HR of 0.794 corresponds to an estimated event hazard about 79.4% of the comparison-group hazard. However, the 95% CI of 0.576–1.094 crosses 1, and the reported P-value is 0.1563. The point estimate alone should therefore not be interpreted as demonstrating superiority.

Time to Skeletal-related Event (SRE)

Hazard ratio

0.857

95% CI: 0.615–1.194   ·   P = 0.3608

The HR of 0.857 represents an estimated event hazard about 85.7% of the comparison-group hazard. The 95% CI of 0.615–1.194 includes 1, and the P-value is 0.3608. As with the other secondary endpoints, the estimate should be interpreted together with its uncertainty rather than by the point estimate alone.

Multiplicity matters: Six formal statistical analyses are posted in the ClinicalTrials.gov record: two primary and four secondary endpoints. The ClinicalTrials.gov record does not specify an endpoint hierarchy, alpha allocation, multiplicity-adjustment procedure, or other familywise-error strategy. The reported P-values should therefore be presented as reported registry results rather than assigned an unreported multiplicity interpretation.

8. Statistical Methodology

Intention-to-treat analysis

For every posted analysis, the registry identifies the intent-to-treat (ITT) population. It defines this population as including all randomized participants, classified according to their assigned treatment group regardless of the actual treatment received.

Core ITT principle
Randomized assignment → analysis according to assigned group

The ITT framework preserves the comparison created by randomization. Treatment discontinuation, treatment switching, or other deviations do not ordinarily cause a participant to be reclassified into another randomized group for the primary efficacy comparison.

Log-rank test

The registry identifies the log-rank test as the method used for every posted analysis. The log-rank test is designed to compare time-to-event experience between groups while incorporating the timing of events and allowing for right-censored observations.

Conceptually, the test compares the observed number of events in each group with the number that would be expected if the groups had the same underlying survival experience. It therefore uses the full ordering of event times rather than reducing the outcome to a single binary endpoint.

Hazard ratio

The effect measure reported for every posted analysis is the hazard ratio. A hazard is an instantaneous event rate conditional on remaining event-free up to a particular time. A hazard ratio below 1 indicates a lower estimated event hazard in the apalutamide-plus-ADT group relative to the placebo-plus-ADT group under the reported comparison.

Generic hazard-ratio interpretation
HR < 1  →  lower estimated instantaneous event hazard in the treatment group

The quantity 1 − HR can be used for a simple relative-hazard interpretation when the HR is below 1, but that calculation should not be confused with an absolute risk reduction or a probability of benefit for an individual participant.

Confidence intervals

Each of the six posted statistical analyses includes a two-sided 95% confidence interval. For the primary endpoints, the intervals are 0.391–0.600 for rPFS and 0.534–0.793 for OS. A confidence interval communicates the statistical precision of the effect estimate; it does not give the probability that the true effect lies inside that particular realized interval.

Time-to-event censoring

Time-to-event analysis is necessary because not every participant necessarily experiences the event during the period in which the outcome is observed. Participants without a recorded event by their available follow-up can contribute information up to the time they are censored. This is one reason why Kaplan-Meier methods and log-rank tests are commonly used for endpoints such as rPFS and OS.

Superiority testing

The registry classifies all six reported analyses under a superiority hypothesis type. This is distinct from a non-inferiority design: the statistical objective is to evaluate evidence for a difference favoring the treatment comparison rather than to demonstrate that treatment is no worse than control within a prespecified margin.

9. Statistical Methods Explained

Why is a log-rank test appropriate for these endpoints?

rPFS, OS, time to initiation of cytotoxic chemotherapy, time to pain progression, time to chronic opioid use, and time to skeletal-related event are all time-to-event outcomes in the posted analyses. The log-rank test compares the event-time distributions between randomized groups while incorporating censored observations rather than treating them as if they had experienced no event.

What does an HR of 0.484 mean?

An HR of 0.484 means that the estimated instantaneous rate of the rPFS event in the apalutamide-plus-ADT group was 48.4% of the corresponding rate in the placebo-plus-ADT group under the reported analysis. The complementary 51.6% figure is a relative hazard reduction, not a percentage of participants who benefited.

Why does the confidence interval matter?

A point estimate alone does not communicate how precisely the treatment effect has been estimated. For rPFS, the 95% CI is 0.391–0.600; for OS, it is 0.534–0.793. The width of these intervals gives information about statistical precision, while their location describes the range of hazard-ratio values compatible with the stated confidence framework.

Why doesn't the P-value measure effect size?

A P-value describes how compatible the observed data are with a specified null hypothesis under the statistical model and testing procedure. It does not measure the magnitude of the treatment effect. Two studies can produce different P-values despite similar effect estimates because sample size, event information, variability, and other design characteristics differ.

Why is ITT analysis important?

ITT analysis retains participants in their randomized groups regardless of the treatment actually received. This protects the original randomized comparison and avoids redefining treatment groups after randomization based on post-randomization events. The TITAN registry analyses explicitly identify the ITT population for all six posted endpoints.

Why are rPFS and OS not interchangeable?

rPFS counts radiographic progression or death as the first event, whereas OS counts death from any cause. Consequently, the endpoints answer different clinical questions and can have different hazard ratios, follow-up periods, censoring patterns, and statistical behavior. The TITAN results illustrate this directly: the reported HR is 0.484 for rPFS and 0.651 for OS.

10. Reading the Primary Results Together

The two primary analyses tell a related but distinct statistical story. The rPFS analysis reports an HR of 0.484, while the OS analysis reports an HR of 0.651. Both confidence intervals are entirely below 1, and both reported P-values are <0.0001.

Relative effect

The hazard ratios summarize the relative event rates between randomized groups. They are useful for comparing treatment effects on the time-to-event scale.

Precision

The confidence intervals quantify uncertainty around the two hazard-ratio estimates and show that the estimates are not exact measurements.

Statistical evidence

Both primary P-values are reported as <0.0001 under superiority analyses. The P-values address statistical evidence, not effect magnitude.

Different estimands

rPFS and OS involve different events. Their numerical hazard ratios should therefore not be expected to be identical.

A useful discipline when interpreting clinical-trial statistics is to keep these four ideas separate: what happened, how large the estimated difference was, how precisely it was estimated, and how the hypothesis test evaluated the null hypothesis. A hazard ratio addresses the second question, its confidence interval addresses the third, and the P-value addresses the fourth.

11. Secondary Results: What the Pattern Shows Statistically

The four secondary analyses are informative because they demonstrate why a collection of hazard ratios should not be summarized simply by counting estimates below 1.

EndpointHR95% CI includes 1?P-valueStatistical reading
Time to Initiation of Cytotoxic Chemotherapy0.469No<.0001Reported estimate and interval are below 1.
Time to Pain Progression0.868Yes0.1966Point estimate below 1, but interval includes 1.
Time to Chronic Opioid Use0.794Yes0.1563Point estimate below 1, but interval includes 1.
Time to Skeletal-related Event (SRE)0.857Yes0.3608Point estimate below 1, but interval includes 1.

The distinction is important. An HR below 1 is a point estimate; it does not by itself establish statistical evidence of superiority. The confidence interval and hypothesis test provide complementary information about uncertainty and compatibility with the null value of 1.

The secondary endpoint results also demonstrate why statistical interpretation should not be reduced to a single threshold. For example, the HR of 0.794 for time to chronic opioid use is numerically lower than 1, but the 95% CI extends to 1.094 and the P-value is 0.1563. The same principle applies to time to pain progression and time to skeletal-related event.

12. Safety Results

The ClinicalTrials.gov record reports serious adverse events by arm as affected participants over participants at risk. The safety data include the two principal treatment groups and a separately identified group described as placebo + ADT to apalutamide + ADT.

Safety groupSerious adverse events affected / at riskObserved proportion
Placebo + ADT115 / 527115/527
Placebo + ADT to Apalutamide + ADT29 / 20829/208
Apalutamide + ADT153 / 524153/524

The registry's safety field reports affected / at-risk counts, rather than a formal comparative effect measure. These counts should therefore be kept separate from the efficacy hazard ratios. A serious adverse-event count does not answer the same statistical question as a time-to-event efficacy analysis.

Safety interpretation: The ClinicalTrials.gov record identifies the serious-adverse-event groups exactly as reported, including the separately named placebo + ADT to apalutamide + ADT group. The ClinicalTrials.gov record does not provide a formal statistical comparison, confidence interval, P-value, or detailed adverse-event definitions for these safety figures.

13. Randomization and the ITT Principle

Randomization is the mechanism that creates the basis for a causal treatment comparison. Before outcomes are observed, participants are assigned to treatment groups according to the trial's randomization process. The ITT analysis then keeps participants associated with their original randomized assignments.

Why the distinction matters
Randomized group ≠ treatment actually received

The TITAN registry explicitly states that the ITT population includes all randomized participants and classifies them according to assigned treatment regardless of actual treatment received. This makes the efficacy analysis answer a randomized-treatment question rather than a question based only on participants who remained on their assigned treatment.

This distinction becomes particularly important when treatment exposure changes after randomization. An analysis that excludes or reassigns participants after randomization can answer a different question and may lose some of the protection against post-randomization selection that randomization provides.

14. Crossover and Treatment Switching

The ClinicalTrials.gov record contains a separately named group, “Placebo + ADT to Apalutamide + ADT”, with serious adverse events reported as 29/208. This terminology indicates that the registry distinguishes participants associated with a transition from placebo plus ADT to apalutamide plus ADT in the posted safety data.

For statistical interpretation, however, the presence of a treatment-transition category should not be used to infer an unreported crossover rate, timing, or crossover-adjusted survival estimate. The ClinicalTrials.gov record does not provide those quantities.

ITT consequence: Even when treatment changes after randomization, the posted efficacy analyses remain ITT analyses. Participants are classified according to their randomized treatment assignment, not reassigned according to subsequent treatment exposure.

15. Multiplicity and Multiple Endpoints

TITAN has two registered primary endpoints and four posted secondary analyses. That structure creates a fundamental statistical issue: multiple hypotheses are being examined.

Endpoint roleNumber reportedStatistical methodHypothesis type
Primary2Log-rank testSuperiority
Secondary4Log-rank testSuperiority

When multiple hypotheses are tested, the interpretation of individual P-values can depend on the prespecified multiplicity strategy. A trial may control familywise type I error through hierarchical testing, alpha allocation, adjusted thresholds, or another procedure. The registry-reported TITAN data do not identify such a procedure.

Accordingly, the most precise description here is to report each registry P-value exactly as posted and avoid assigning an unreported multiplicity-adjusted interpretation. The two primary endpoints should also be distinguished from the four secondary endpoints rather than treating all six analyses as one undifferentiated set.

16. Interim Analysis and Alpha Spending

Interim-analysis details are not provided in the registry-reported TITAN trial data. The registry information identifies the formal statistical analyses, their endpoints, populations, methods, effect measures, confidence intervals, and P-values, but it does not specify an interim-analysis schedule, alpha-spending method, stopping boundary, or information fraction.

What can be established

The registry reports six formal analyses, all using log-rank testing and hazard ratios, with superiority as the hypothesis type.

What cannot be established here

The ClinicalTrials.gov record does not specify whether the reported analyses were subject to an interim-monitoring or alpha-spending scheme.

This distinction matters because repeatedly examining accumulating trial data can affect type I error unless the monitoring strategy is incorporated into the statistical design. Without the relevant protocol or statistical-analysis-plan details, no specific alpha-spending conclusion should be attached to the posted TITAN results.

17. Missing Data, Censoring, and Model Assumptions

Time-to-event analyses naturally involve censoring. A participant can remain alive or event-free at the end of available follow-up and therefore contribute information without contributing an observed event. The ClinicalTrials.gov record does not provide the detailed censoring rules or missing-data/imputation strategy.

Censoring

Survival methods can incorporate participants whose event time is not observed during their available follow-up.

Missing data

No specific imputation method is reported in the ClinicalTrials.gov record, so none is assumed for this analysis.

Proportional hazards

A hazard ratio is most naturally interpreted as a relative hazard under the model's assumptions. The ClinicalTrials.gov record does not report diagnostics for proportional hazards.

Endpoint assessment

For rPFS, the registry explicitly identifies investigator assessment in its endpoint definition.

These limitations do not invalidate the reported analyses. They define what can and cannot be concluded from the registry-level ClinicalTrials.gov record. A rigorous statistical review should distinguish reported methods from methods that would need to be confirmed from a protocol or statistical-analysis plan.

18. Understanding the Hazard Ratios Across TITAN

Reported hazard ratios
rPFS
0.484
OS
0.651
Cytotoxic chemotherapy
0.469
Pain progression
0.868
Chronic opioid use
0.794
Skeletal-related event
0.857

The visualization highlights an important statistical lesson: the six estimates are not all equally informative merely because they are below 1. The primary rPFS and OS intervals are entirely below 1, as is the interval for time to initiation of cytotoxic chemotherapy. The other three secondary confidence intervals extend across 1.

The graph also should not be read as a ranking of clinical importance. Each endpoint represents a different event, and the registry provides different time frames and uncertainty intervals for each. A hazard ratio is most useful when interpreted in the context of its endpoint definition, confidence interval, P-value, analysis population, and prespecified role in the trial.

19. What the Hazard Ratio Does — and Does Not — Mean

Relative treatment effect

An HR below 1 indicates a lower estimated instantaneous event rate in the apalutamide-plus-ADT group than in the placebo-plus-ADT group under the reported analysis. For example, the rPFS HR of 0.484 corresponds to an estimated hazard approximately 51.6% lower than the comparison hazard.

Not an individual probability

A hazard ratio does not tell an individual participant their probability of avoiding an event. It also does not mean that a fixed percentage of participants experienced the same relative reduction. The measure describes a relative comparison of event hazards across time.

Confidence interval

The 95% confidence interval describes uncertainty around the estimated hazard ratio. For OS, the interval is 0.534–0.793. It is not a prediction interval for individual survival times and should not be interpreted as saying that individual participants' effects fall within those limits.

P-value

The P-value evaluates the evidence against a specified null hypothesis under the statistical testing framework. A smaller P-value does not automatically imply a larger treatment effect. Effect magnitude and statistical evidence are separate concepts.

20. Clinical Interpretation vs Statistical Interpretation

Statistical interpretation

The registry reports hazard ratios below 1 for both primary endpoints, with 95% confidence intervals entirely below 1 and P-values <0.0001. The primary analyses were performed in the ITT population using log-rank tests and superiority hypotheses.

Clinical interpretation

The trial evaluates whether adding apalutamide to ADT changes clinically relevant time-to-event outcomes in participants with mHSPC. The registry reports distinct effects across progression, survival, chemotherapy initiation, pain progression, chronic opioid use, and skeletal-related events.

Keeping these interpretations separate prevents a common statistical error: treating statistical significance as synonymous with clinical magnitude. The TITAN registry results provide quantitative estimates, confidence intervals, and hypothesis-test results; the clinical meaning of each endpoint depends on what event it measures and how that event relates to patient outcomes.

21. Limitations

22. Why This Trial Matters Statistically

TITAN is a useful teaching case because it places several fundamental clinical-trial concepts into one randomized time-to-event framework. The same trial illustrates how an endpoint's definition, analysis population, effect measure, confidence interval, and P-value must be interpreted together.

ConceptHow it appears in TITAN
RandomizationThe trial is randomized with 2 parallel treatment arms.
BlindingThe registry classifies the study as triple-masked.
ITT analysisAll six posted analyses use the ITT population.
Time-to-event endpointsBoth primary endpoints and all four secondary analyses are time-to-event outcomes.
Log-rank testThe reported method for all six statistical analyses.
Hazard ratioThe reported effect measure for every posted analysis.
Confidence intervalEach statistical analysis includes a two-sided 95% CI.
P-valuesEach of the six posted analyses includes a reported P-value.
SuperiorityAll posted analyses are classified with a superiority hypothesis type.
Multiple endpointsTwo primary and four secondary analyses are posted.
Crossover / treatment transitionThe safety data separately identify a placebo + ADT to apalutamide + ADT group.
Safety analysisSerious adverse events are reported as affected participants over those at risk.

The strongest statistical lesson is that an effect estimate is inseparable from its endpoint and analysis framework. An HR of 0.484 for rPFS and an HR of 0.651 for OS cannot be interpreted as competing versions of the same outcome. They quantify different event processes over different reported time frames.

23. Related Tutorials

Learn more about the methods used in this trial:

24. Related Calculators

25. Sources

Continue through the Clinical Biostats statistical library

Explore tutorials on survival analysis and clinical-trial methodology, then apply the concepts with statistical calculators.

26. Record Summary

TITAN provides a clear example of randomized clinical-trial analysis centered on time-to-event outcomes. The ClinicalTrials.gov record reports two primary endpoints—radiographic progression-free survival and overall survival—and four secondary time-to-event endpoints. All six analyses use the ITT population, compare placebo plus ADT with apalutamide plus ADT, use the log-rank test, report hazard ratios with two-sided 95% confidence intervals, and are classified as superiority analyses.

The primary rPFS analysis reports an HR of 0.484 (95% CI 0.391–0.600; P <0.0001), while the primary OS analysis reports an HR of 0.651 (95% CI 0.534–0.793; P <0.0001). Among the secondary endpoints, the reported HR is 0.469 for time to initiation of cytotoxic chemotherapy, while the confidence intervals for time to pain progression, time to chronic opioid use, and time to skeletal-related event include 1.

Clinical Biostats methodology: The goal of a trial-results page is not simply to reproduce statistical outputs. It is to explain what each estimate measures, how the analysis supports the randomized comparison, how uncertainty should be read, and where the available registry data do not support additional conclusions.