← Clinical Trials
Metastatic Castration-Resistant Prostate Cancer Phase 3 Time-to-Event NCT04647526

SPLASH: Complete Statistical Analysis of [Lu-177]-PNT2002 in Metastatic Castration-Resistant Prostate Cancer

An independent statistical review of the randomized phase 3 SPLASH trial evaluating [Lu-177]-PNT2002 against abiraterone or enzalutamide in metastatic castration-resistant prostate cancer after second-line hormonal treatment.

Trial start: February 25, 2021  ·  Primary completion: November 1, 2023  ·  Enrollment: 455
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

SPLASH is a randomized, parallel, phase 3 oncology trial evaluating [Lu-177]-PNT2002 in metastatic castration-resistant prostate cancer. The registered primary endpoint is radiographic progression-free survival (rPFS), a time-to-event endpoint analyzed using a log-rank test and summarized with a hazard ratio.

455
Enrollment
Phase 3
0.71
rPFS HR
95% CI 0.55–0.92
0.0088
P-value
Superiority analysis
4
Arms
Randomized parallel design
FeatureSPLASH
TrialSPLASH
NCT IDNCT04647526
PhasePhase 3
ConditionMetastatic Castration-Resistant Prostate Cancer
AllocationRandomized
Design modelParallel
MaskingNone
Primary purposeTreatment
Enrollment455
Primary endpointRandomization Phase: Radiographic Progression-Free Survival (rPFS)
Primary endpoint typeTime-to-event
Primary analysisLog-rank test
Effect measureHazard ratio
Hypothesis typeSuperiority
Results postedYes
Statistical analyses posted1

2. Clinical Question

The central statistical question is whether randomized treatment with [Lu-177]-PNT2002 produces a different time to radiographic progression or death compared with the randomized comparator of abiraterone or enzalutamide in patients with metastatic castration-resistant prostate cancer.

Population

Participants with metastatic castration-resistant prostate cancer enrolled in the phase 3 SPLASH trial.

Intervention

[Lu-177]-PNT2002.

Comparator

Abiraterone or enzalutamide in the Randomization Phase.

Primary question

Does [Lu-177]-PNT2002 improve radiographic progression-free survival relative to abiraterone or enzalutamide?

3. Trial Design

The registry describes SPLASH as a randomized, parallel, unmasked phase 3 treatment study. Enrollment was 455 participants, with four arms recorded in the ClinicalTrials.gov record. The interventions listed in the registry are [Lu-177]-PNT2002, abiraterone, and enzalutamide.

01
Enroll455 participants
02
RandomizeParallel design
03
Treat[Lu-177]-PNT2002 vs comparator
04
AssessRadiographic progression
05
AnalyzeTime-to-event comparison
RANDOMIZATION PHASE · ARM A

[Lu-177]-PNT2002

  • Drug intervention: [Lu-177]-PNT2002
  • Primary rPFS comparison includes Arm A
  • 114 participants were censored in the reported ITT analysis
RANDOMIZATION PHASE · ARM B

Abiraterone or Enzalutamide

  • Comparator intervention: abiraterone or enzalutamide
  • Primary rPFS comparison includes Arm B
  • 40 participants were censored in the reported ITT analysis
Interpretive point: The primary statistical comparison reported in the registry is specifically between Randomization Phase [Lu-177]-PNT2002 (Arm A) and Randomization Phase Abiraterone or Enzalutamide (Arm B). the ClinicalTrials.gov record records four arms overall, but the registry-reported statistical analysis contains only this primary comparison.

4. Trial Timing and Status

February 25, 2021

Trial start

The registry lists February 25, 2021 as the study start date.

November 1, 2023

Primary completion

The registry lists November 1, 2023 as the primary completion date.

Current registry status

Active, not recruiting

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

5. Primary Endpoint

EndpointRegistry definition and time frameAnalysis
Randomization Phase: Radiographic Progression-Free Survival (rPFS) From the randomization date to the first documented progressive disease or death from any cause (up to 23 months). rPFS, as assessed by blinded independent central review (BICR), is the time from the randomization date to progression on soft tissue per Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST v1.1) or confirmed progression on bone lesions by Prostate Cancer Working Group 3 (PCWG3) criteria, or death from any cause. Log-rank test; hazard ratio; superiority hypothesis

This endpoint combines two clinically important event types into one time-to-event outcome: documented radiographic progression and death from any cause. A participant who experiences either event reaches the endpoint. Participants who have not experienced an event by the relevant observation point can contribute censored follow-up to the analysis.

6. Statistical Methodology

Intention-to-treat analysis

The reported primary analysis used all participants from the ITT analysis set. ITT analysis preserves the randomized comparison by analyzing participants according to their assigned treatment group rather than redefining the primary efficacy population according to subsequent treatment exposure.

Reported analysis population
ITT analysis set  →  Arm A vs Arm B

The registry-reported analysis reports 114 participants censored in [Lu-177]-PNT2002 (Arm A) and 40 participants censored in Abiraterone or Enzalutamide (Arm B).

Kaplan-Meier framework

Radiographic progression-free survival is a time-to-event endpoint, so the natural descriptive framework is Kaplan-Meier estimation. Kaplan-Meier methods allow participants with incomplete event follow-up to contribute information up to their censoring time rather than requiring every participant to have experienced progression or death.

Conceptual survival function
S(t) = ∏ti ≤ t (1 − di/ni)

Here, di represents events at time ti, while ni represents participants at risk immediately before that time.

Log-rank test

The registry identifies the log-rank test as the primary statistical method. The log-rank test compares the observed and expected numbers of events between randomized groups across follow-up, making it appropriate for a time-to-event comparison such as rPFS.

Hazard ratio

The reported effect measure is a hazard ratio. Unlike a simple proportion or risk ratio at one fixed time point, a hazard ratio summarizes the relative instantaneous event rate between treatment groups over the analyzed follow-up under the time-to-event modeling framework.

Conceptual interpretation
HR < 1  →  lower estimated instantaneous event rate in the treatment group

A hazard ratio below 1 does not mean that the same percentage of every individual patient's risk was reduced, nor does it directly give an absolute difference in progression-free survival.

Censoring

The reported ITT analysis includes censored participants: 114 in Arm A and 40 in Arm B. Censoring means that the analysis has follow-up information up to a specified point without observing the endpoint by that point. Correct interpretation of Kaplan-Meier and hazard-ratio analyses depends on how censoring relates to the underlying event process.

7. Primary rPFS Results

ClinicalTrials.gov reports one formal statistical analysis for the primary endpoint. The comparison is between Randomization Phase [Lu-177]-PNT2002 (Arm A) and Randomization Phase Abiraterone or Enzalutamide (Arm B).

Radiographic progression-free survival

HR 0.71

95% CI: 0.55–0.92   ·   P = 0.0088

Analysis: log-rank test   ·   Hypothesis: superiority

Primary endpoint[Lu-177]-PNT2002Abiraterone or EnzalutamideEffect estimateP-value
Randomization Phase: Radiographic Progression-Free Survival (rPFS) Arm A Arm B HR 0.71 (95% CI 0.55–0.92) 0.0088
Clinical Biostats interpretation

The reported HR of 0.71 means that the estimated instantaneous rate of the rPFS event was 0.71 times that of the comparator group under the reported time-to-event analysis. Expressed as a simple relative-hazard interpretation, this corresponds to an estimated 29% lower hazard in the [Lu-177]-PNT2002 group.

The HR does not mean that 29% of patients avoided progression, that each individual patient had exactly a 29% reduction in risk, or that the absolute probability of progression or death was reduced by 29 percentage points.

The 95% CI of 0.55–0.92 describes uncertainty around the estimated hazard ratio under the analysis framework. It is not a prediction interval for individual patients and does not describe the range of outcomes that individual patients will experience.

The P-value of 0.0088 addresses the statistical evidence against the null hypothesis under the specified testing framework. It does not measure the size or clinical importance of the treatment effect. Effect size and uncertainty are better represented by the HR and its confidence interval.

Because this is a time-to-event analysis, interpretation also depends on censoring and on the extent to which a single hazard ratio appropriately summarizes the treatment difference over follow-up. The ClinicalTrials.gov record does not provide additional diagnostics concerning the proportional-hazards assumption, so no conclusion about that assumption is made here.

What the result says statistically

The reported analysis is consistent with a statistically detectable difference in rPFS between the two randomized groups under the registry's stated superiority hypothesis. The confidence interval is entirely below 1.00, while the point estimate is 0.71.

What the result does not establish by itself

8. Secondary Endpoint Results

the ClinicalTrials.gov record reports 6 outcome measures, but only 1 statistical analysis is posted in the ClinicalTrials.gov record, corresponding to the primary rPFS endpoint.

Reporting boundary: no formal statistical analyses for secondary endpoints are included in the ClinicalTrials.gov record. This page therefore does not assign estimates, confidence intervals, or P-values to secondary outcomes that are not represented in the posted statistical analysis.

For a secondary time-to-event endpoint, a typical analysis would use a Kaplan-Meier framework with a between-group comparison such as a log-rank test and an effect measure such as a hazard ratio. However, that describes the standard statistical approach; it does not constitute a reported SPLASH secondary-endpoint result.

9. Safety Results

The ClinicalTrials.gov record reports serious adverse events by phase and arm as affected participants over participants at risk. These counts are reported separately from the primary rPFS analysis.

Phase / armSerious adverse events affected / at risk
Dosimetry Phase: [Lu-177]-PNT20026/27
Randomization Phase: [Lu-177]-PNT2002 (Arm A)47/269
Randomization Phase: Abiraterone or Enza30/130
Randomization Phase: Abiraterone or Enza14/77
PK Extension Phase: [Lu-177]-PNT20022/15
Safety interpretation: the ClinicalTrials.gov record contains phase-specific affected/at-risk counts, including two separately reported Randomization Phase entries for Abiraterone or Enza. They are reproduced exactly rather than combined or reinterpreted, because the ClinicalTrials.gov record does not provide a basis for determining whether those entries should be pooled.

10. Statistical Methods Explained

Why is rPFS a time-to-event endpoint?

rPFS records the amount of time from randomization until the first qualifying event: documented progressive disease or death from any cause. The endpoint therefore incorporates both whether an event occurred and when it occurred. Participants who have not experienced an event by the relevant observation point contribute censored follow-up.

What does an HR of 0.71 mean?

An HR of 0.71 means that the estimated instantaneous rate of the rPFS event in the [Lu-177]-PNT2002 group was 0.71 times the corresponding estimated rate in the comparator group under the reported analysis. The complementary calculation, 1 − 0.71, gives a 29% relative reduction in the estimated hazard. That is not the same as a 29% absolute reduction in the probability of progression or death.

Why use a log-rank test?

The log-rank test is designed for comparing survival-type distributions between groups when participants may be censored. Instead of reducing each patient to a single binary outcome at one fixed time, it uses information from the timing of observed events throughout follow-up.

Why is the confidence interval important?

The point estimate of 0.71 is only one estimate of the treatment effect. The 95% CI of 0.55–0.92 communicates uncertainty around that estimate. A narrower interval would indicate greater statistical precision, while a wider interval would indicate less precision. The interval does not describe individual-level variation in treatment response.

Why does the P-value not measure effect size?

The P-value of 0.0088 quantifies how compatible the observed statistical evidence is with the specified null hypothesis under the testing framework. It does not tell us that the treatment effect is "0.0088" in size. The HR describes relative effect, while the CI communicates uncertainty around the HR.

Why does censoring matter?

Censoring allows a participant to contribute observed follow-up even when the endpoint is not observed during that person's available follow-up. Time-to-event methods rely on assumptions about the relationship between censoring and the underlying event process. The ClinicalTrials.gov record identifies censored participants but do not provide further diagnostics about the censoring mechanism.

Does HR 0.71 prove proportional hazards?

No. A hazard ratio is a relative time-to-event measure, but the ClinicalTrials.gov record does not report a diagnostic assessment of the proportional-hazards assumption. The HR should therefore be interpreted as the reported model-based summary rather than as evidence that hazards were necessarily proportional at every point in time.

11. Randomization and the ITT Principle

Randomization is central to the causal interpretation of a randomized trial because treatment assignment is determined by the study design rather than by investigators or participants choosing the treatment based on prognosis. SPLASH is registered as randomized, and the primary analysis in the ClinicalTrials.gov record uses the ITT analysis set.

Why ITT is useful

Analyzing participants according to randomized assignment preserves the comparison created by randomization and avoids redefining the primary efficacy population solely according to subsequent treatment exposure.

What ITT does not solve

ITT analysis does not eliminate missing follow-up, censoring, treatment discontinuation, or every other source of uncertainty. Those features remain part of the interpretation of the time-to-event analysis.

12. Understanding the Hazard Ratio in Context

Relative effect

The reported HR of 0.71 indicates a lower estimated instantaneous rPFS-event rate for [Lu-177]-PNT2002 relative to abiraterone or enzalutamide in the reported randomized comparison.

Precision

The 95% CI of 0.55–0.92 gives the statistical uncertainty around the estimated HR. The interval remains below 1.00, consistent with the direction of the reported superiority analysis.

Statistical evidence

The reported P = 0.0088 indicates the strength of evidence against the null hypothesis under the specified analysis. It should not be used as a substitute for examining the HR, confidence interval, endpoint definition, and trial design.

Absolute effect

The ClinicalTrials.gov record does not report median rPFS or fixed-time rPFS probabilities. Consequently, the HR cannot be converted here into an absolute number of additional months without introducing information outside the ClinicalTrials.gov record.

13. What Is Known — and What Is Not Reported Here

QuestionInformation reported
Was the trial randomized?Yes.
Was it parallel?Yes.
Was it masked?No.
What was enrollment?455.
What was the primary endpoint?Randomization Phase: Radiographic Progression-Free Survival (rPFS).
What was the analysis method?Log-rank test.
What was the effect measure?Hazard ratio.
What was the primary estimate?0.71.
What was the 95% CI?0.55–0.92.
What was the P-value?0.0088.
What was the hypothesis type?Superiority.
Are median rPFS values reported?No.
Are fixed-time rPFS rates reported?No.
Are formal secondary statistical analyses reported?No; one statistical analysis is reported, for the primary endpoint.
Are subgroup statistical analyses reported?No.
Are interim-analysis details reported?No.
Is a non-inferiority margin reported?No; the registered hypothesis type is superiority.

14. Design Topics Not Supported by the Supplied Data

Several design questions are important in clinical-trial statistical interpretation, but they should not be inferred when the registry data posted on ClinicalTrials.gov for a particular analysis do not contain the relevant information.

Non-inferiority margin

The registered hypothesis type is superiority. No non-inferiority margin is reported, so no margin-based interpretation is appropriate.

Interim analysis

The ClinicalTrials.gov record does not describe an interim efficacy analysis, alpha-spending plan, or stopping boundary.

Multiplicity

The ClinicalTrials.gov record does not provide an alpha-allocation or multiplicity-adjustment strategy beyond identifying the primary hypothesis as superiority.

Missing-data imputation

No specific imputation method is reported. The primary endpoint is a time-to-event outcome with reported censoring rather than a simple continuous endpoint requiring a conventional imputation rule.

15. Why the Primary Analysis Is a Survival Analysis

A conventional two-arm comparison of means would discard much of the information contained in the timing of progression or death. rPFS instead records a time from randomization to an event, while allowing some participants to remain event-free at the time their available follow-up ends.

Statistical featureRole in SPLASH
RandomizationCreates the treatment-group comparison.
ITT analysisDefines the reported primary efficacy population.
Time-to-event endpointUses both event occurrence and event timing.
CensoringAllows participants without an observed endpoint to contribute available follow-up.
Log-rank testProvides the reported between-group survival comparison.
Hazard ratioProvides the reported relative effect measure.
Confidence intervalCommunicates uncertainty around the HR.
P-valueCommunicates statistical evidence under the stated hypothesis test.

16. Limitations

17. Why This Trial Matters Statistically

SPLASH is a useful teaching case because its primary endpoint illustrates the core structure of modern randomized time-to-event analysis: randomized treatment assignment, an ITT efficacy population, an event defined using radiographic progression or death, censoring, a log-rank comparison, a hazard ratio, a confidence interval, and a superiority hypothesis.

ConceptHow it appears in SPLASH
RandomizationThe trial is registered as randomized.
Parallel designThe registered design model is parallel.
Intention-to-treat analysisThe primary rPFS analysis uses all participants from the ITT analysis set.
Time-to-event endpointrPFS is measured from randomization to progression or death.
Kaplan-Meier frameworkAppropriate for describing a time-to-event endpoint with censoring.
Log-rank testThe registry-reported primary statistical method.
Hazard ratioThe registry-reported effect measure.
Confidence interval95% CI of 0.55–0.92 around the HR of 0.71.
P-value0.0088 for the reported superiority comparison.
Censoring114 participants in Arm A and 40 in Arm B were censored in the reported analysis.

18. Related Tutorials

Learn more about the methods used in this trial:

19. Related Calculators

20. Sources

Continue learning from the SPLASH statistical framework

Explore the broader Clinical Biostats collection of biostatistics tutorials, statistical calculators, and clinical-trial statistical analyses.

21. Record Summary

SPLASH provides a clear example of a randomized phase 3 time-to-event analysis. The registry identifies radiographic progression-free survival as the primary endpoint, defines it from randomization to documented progressive disease or death, and reports an ITT comparison between [Lu-177]-PNT2002 and abiraterone or enzalutamide using a log-rank test. The reported hazard ratio was 0.71, with a 95% CI of 0.55–0.92 and P = 0.0088 under a superiority hypothesis.

The most informative statistical reading of that result combines the relative effect estimate with its confidence interval and P-value while recognizing what those quantities do not provide. The HR does not give an absolute treatment benefit, the P-value does not measure effect size, and the confidence interval does not describe individual patient outcomes. The ClinicalTrials.gov record also establish important boundaries: only one formal statistical analysis is provided, secondary endpoint analyses are not included, and the reported analysis contains censored participants in both randomized groups.

Clinical Biostats methodology: A trial-results page should distinguish reported evidence from statistical interpretation. Where the ClinicalTrials.gov record does not contain an estimate, analysis detail, subgroup result, or design parameter, the appropriate approach is to leave that quantity unreported rather than infer it from external sources.