This page separates reported trial results from statistical interpretation. Numerical results are taken from the ClinicalTrials.gov data posted on ClinicalTrials.gov for NCT01946204. The registry states that these results are up to the clinical cutoff date (CCO), 19 May 2017.
1. Trial at a Glance
SPARTAN was a randomized, parallel-group, triple-masked phase 3 trial evaluating apalutamide versus placebo in men with prostatic neoplasms. The registered primary endpoint was metastasis-free survival by blinded independent central review, a time-to-event endpoint analyzed using the log-rank test and hazard ratio.
| Feature | SPARTAN |
|---|---|
| Phase | Phase 3 |
| Brief title | A Study of Apalutamide (ARN-509) in Men With Non-Metastatic Castration-Resistant Prostate Cancer |
| Condition | Prostatic Neoplasms |
| Design | Randomized, parallel-group, triple-masked |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Interventions | Apalutamide; placebo |
| Enrollment | 1207 |
| Primary endpoint type | Time-to-event |
| Statistical method reported | Log-rank test |
| Effect measure | Hazard ratio |
| Hypothesis type | Superiority |
| ClinicalTrials.gov | NCT01946204 |
| Lead sponsor | Aragon Pharmaceuticals, Inc. |
2. Clinical Question
The central statistical question was whether treatment with apalutamide differed from placebo with respect to metastasis-free survival in men enrolled in this phase 3 study of non-metastatic castration-resistant prostate cancer.
Population
Men with non-metastatic castration-resistant prostate cancer, represented in the registry under the condition category Prostatic Neoplasms.
Intervention
Apalutamide (ARN-509).
Comparator
Placebo.
Primary question
Does apalutamide improve the time from randomization to first evidence of BICR-confirmed bone or soft tissue distant metastasis or death from any cause?
3. Trial Design
Apalutamide
- Drug intervention: apalutamide.
- Included in the randomized two-arm comparison.
- Evaluated against placebo for time-to-event outcomes.
Placebo
- Drug intervention: placebo.
- Included in the randomized two-arm comparison.
- Served as the comparator for the superiority hypothesis.
4. Trial Timeline
Trial start
The registry records the study start date as 14 October 2013.
Primary completion
The registry records primary completion on 19 May 2017.
Clinical cutoff
The registry states that the posted results are up to the clinical cutoff date, also 19 May 2017.
Active, not recruiting
The ClinicalTrials.gov record identifies the study as active, not recruiting.
5. Endpoints
| Endpoint | Registry time frame | Definition / statistical role |
|---|---|---|
| Metastasis-Free Survival (MFS) by Blinded Independent Central Review (BICR) | Up to approximately 43 Months | Primary time-to-event endpoint. MFS was defined as the time from randomization to the time of first evidence of BICR-confirmed bone or soft tissue distant metastasis or death due to any cause, whichever occurred first. |
| Time to Metastasis (TTM) | Up to approximately 43 Months | Secondary time-to-event endpoint analyzed using the log-rank test and hazard ratio. |
| Progression-free Survival (PFS) | Up to approximately 43 Months | Secondary time-to-event endpoint analyzed using the log-rank test and hazard ratio. |
| Time to Symptomatic Progression | Up to approximately 43 Months | Secondary time-to-event endpoint analyzed using the log-rank test and hazard ratio. |
6. Statistical Methodology
Intention-to-treat analysis
The primary and secondary efficacy analyses were conducted in the Intent-to-Treat (ITT) population. The registry defines this population as including all participants randomized into the study, with study-drug assignments designated according to initial randomization, regardless of whether participants received the assigned study treatment.
This is important because the treatment comparison remains anchored to the randomized allocation. A participant's later treatment exposure does not redefine which randomized group contributes to the primary efficacy analysis.
Log-rank test
The registry reports the log-rank test for MFS, TTM, PFS, and time to symptomatic progression. The log-rank test compares the event-time distributions between the randomized groups while accounting for the fact that participants may have different follow-up times and may be censored before experiencing the event.
The log-rank framework evaluates whether the observed pattern of events over follow-up is compatible with the null hypothesis of no difference in the underlying survival distributions.
Hazard ratio
The effect measure reported for every statistical analysis is the hazard ratio (HR). For the registry comparisons, the groups are listed as placebo versus apalutamide, so an HR below 1 indicates a lower estimated instantaneous event rate for apalutamide relative to placebo under the fitted time-to-event model.
For example, an HR of 0.280 corresponds to an estimated hazard that is 28.0% of the comparator hazard, or a 72.0% lower estimated hazard relative to placebo. This is a relative hazard interpretation, not a statement that 72.0% of participants avoided metastasis.
Confidence intervals
Each posted statistical analysis supplies a two-sided 95% confidence interval around the hazard ratio. The interval communicates statistical uncertainty around the estimated relative effect under the analysis framework.
A confidence interval is not a range containing 95% of individual patient outcomes. It instead describes uncertainty about the population-level effect estimate generated by the statistical procedure.
7. Statistical Methods Explained
Why was a log-rank test used?
MFS, TTM, PFS, and time to symptomatic progression are time-to-event endpoints. Participants can have different follow-up durations, and some may not experience the event during observation. The log-rank test is designed for comparing such event-time distributions while incorporating censored observations.
What does the MFS hazard ratio measure?
The MFS HR summarizes the relative instantaneous rate of experiencing the MFS event between the randomized groups under the fitted time-to-event framework. Because the MFS event includes either BICR-confirmed distant metastasis or death from any cause, the HR refers to that composite event definition rather than metastasis alone.
Why is the ITT population important?
Analyzing randomized participants according to their initial assignment preserves the treatment comparison created by randomization. It avoids redefining the comparison based on treatment actually received after randomization.
What does a 95% confidence interval tell us?
The confidence interval shows how precisely the hazard ratio has been estimated. A relatively narrow interval indicates greater statistical precision than a very wide interval. It does not describe individual treatment responses or guarantee that the true effect lies within the interval.
Why doesn't the p-value measure effect size?
A p-value addresses evidence against a null hypothesis under the specified statistical model. It does not tell us how large the treatment effect is. The hazard ratio describes relative effect size, while its confidence interval describes uncertainty around that estimate.
Why does censoring matter?
Time-to-event methods allow participants who have not yet experienced the event to contribute information up to their censoring time. The resulting estimates therefore depend not only on observed events but also on the timing and handling of censoring.
Why should the hazard ratio not be interpreted as an absolute risk reduction?
An HR is a relative time-to-event measure. It does not directly provide the difference in the probability of an event by a particular calendar time or the number of patients who would need treatment to prevent one event. Absolute measures require appropriate time-specific survival or cumulative-incidence information.
8. Primary Results: Metastasis-Free Survival
The registry contains two formal primary analyses of metastasis-free survival by blinded independent central review, both using the ITT population and the log-rank test. They correspond to the US regulatory and ex-US regulatory analyses.
US Regulatory Analysis
Hazard ratio for MFS
95% CI: 0.227–0.346 · P < 0.0001
Groups compared: Placebo vs Apalutamide
| Feature | Reported result |
|---|---|
| Endpoint | Metastasis-Free Survival (MFS) by Blinded Independent Central Review (BICR) |
| Time frame | Up to approximately 43 Months |
| Analysis population | Intent-to-Treat (ITT) |
| Method | Log Rank |
| Effect measure | Hazard Ratio (HR) |
| Estimate | 0.280 |
| 95% CI | 0.227–0.346 |
| P-value | <0.0001 |
| Hypothesis | Superiority |
| Analysis note | Statistical Analysis for MFS by BICR (US Regulatory) |
The HR of 0.280 means that, under the time-to-event analysis, the estimated instantaneous rate of the MFS event was 28.0% of the rate in the placebo group. Expressed as a relative hazard comparison, this corresponds to a 72.0% lower estimated hazard with apalutamide.
This does not mean that 72.0% of participants were protected from metastasis, that 72.0% of participants were cured, or that each participant experienced the same proportional reduction in risk. It is a population-level relative hazard estimate.
The 95% CI of 0.227–0.346 indicates that the estimated effect is relatively precise within this analysis framework. Because the entire interval is below 1, the interval is consistent with a lower estimated hazard in the apalutamide group relative to placebo.
The p-value of <0.0001 addresses the strength of evidence against the null hypothesis under the specified testing framework. It does not measure the magnitude of the treatment effect; the HR and confidence interval provide that information.
As with other hazard-ratio analyses, interpretation should recognize that the HR is a time-to-event summary and should not automatically be treated as a constant individual-level risk reduction throughout follow-up.
Ex-US Regulatory Analysis
Hazard ratio for MFS
95% CI: 0.244–0.362 · P < 0.0001
Groups compared: Placebo vs Apalutamide
| Feature | Reported result |
|---|---|
| Endpoint | Metastasis-Free Survival (MFS) by Blinded Independent Central Review (BICR) |
| Time frame | Up to approximately 43 Months |
| Analysis population | Intent-to-Treat (ITT) |
| Method | Log Rank |
| Effect measure | Hazard Ratio (HR) |
| Estimate | 0.297 |
| 95% CI | 0.244–0.362 |
| P-value | <0.0001 |
| Hypothesis | Superiority |
| Analysis note | Statistical Analysis for MFS by BICR (Ex-US Regulatory) |
The ex-US regulatory analysis produced an HR of 0.297. Under the same general hazard-ratio interpretation, the estimated instantaneous MFS event rate was 29.7% of the placebo-group rate, corresponding to a 70.3% lower estimated hazard relative to placebo.
The HR is not a probability and should not be translated into a percentage of patients who benefited. It is a relative measure of event hazard over the analyzed time-to-event framework.
The 95% CI of 0.244–0.362 indicates uncertainty around the point estimate. The interval remains entirely below 1, so the estimated direction of the treatment effect is consistently below the no-difference value within this analysis.
The p-value of <0.0001 indicates strong statistical evidence against the null hypothesis under the reported superiority test, but it does not quantify clinical magnitude. The HR and its confidence interval remain the more direct descriptors of the relative treatment effect.
9. Primary Endpoint Results Side by Side
| Primary analysis | HR | 95% CI | P-value | Method |
|---|---|---|---|---|
| MFS by BICR — US Regulatory | 0.280 | 0.227–0.346 | <0.0001 | Log Rank |
| MFS by BICR — Ex-US Regulatory | 0.297 | 0.244–0.362 | <0.0001 | Log Rank |
The two posted regulatory analyses produce similar conclusions at the level of the reported hazard ratios: both estimates are substantially below 1, both 95% confidence intervals are entirely below 1, and both p-values are reported as <0.0001. The appropriate interpretation is that the registry reports strong statistical evidence for a superiority comparison of apalutamide versus placebo on MFS under both regulatory analyses.
The two estimates should not be averaged or combined into a new hazard ratio. They are separate registry analyses, and the ClinicalTrials.gov record does not provide a statistical procedure for combining them.
10. Secondary Endpoint Results
Time to Metastasis
US Regulatory Analysis
95% CI: 0.219–0.335 · P < 0.0001
The US regulatory analysis of time to metastasis used the ITT population and the log-rank test. An HR of 0.271 corresponds to a 27.1% estimated hazard relative to placebo, or a 72.9% lower estimated hazard for the event under the model.
The confidence interval, 0.219–0.335, remains below 1 and provides the uncertainty range reported for the hazard-ratio estimate. The p-value of <0.0001 is evidence against the null hypothesis in the reported analysis, but it is not a measure of effect size. The HR itself supplies the relative-effect estimate.
Ex-US Regulatory Analysis
95% CI: 0.227–0.342 · P < 0.0001
The ex-US regulatory analysis produced an HR of 0.279, corresponding to a 27.9% estimated event hazard relative to placebo, or a 72.1% lower estimated hazard under the model.
The 95% CI of 0.227–0.342 is entirely below 1. The p-value of <0.0001 provides statistical evidence against the null hypothesis in this reported superiority analysis. Neither the p-value nor the HR should be interpreted as an absolute probability that an individual patient will or will not develop metastasis.
Progression-free Survival
US Regulatory Analysis
95% CI: 0.238–0.356 · P < 0.0001
The US regulatory PFS analysis used the ITT population and log-rank testing. The HR of 0.291 represents an estimated event hazard of 29.1% relative to placebo, corresponding to a 70.9% lower estimated hazard.
The 95% CI of 0.238–0.356 describes uncertainty around the estimated HR and remains below 1. The p-value of <0.0001 describes evidence against the null hypothesis under the reported test; it does not describe the magnitude or practical importance of the treatment effect.
Ex-US Regulatory Analysis
95% CI: 0.247–0.364 · P < 0.0001
The ex-US regulatory PFS analysis produced an HR of 0.300, corresponding to an estimated event hazard of 30.0% relative to placebo, or a 70.0% lower estimated hazard.
The confidence interval of 0.247–0.364 indicates the statistical uncertainty reported around the point estimate. The interval remains below 1, while the p-value of <0.0001 provides evidence against the no-difference null hypothesis within the reported analysis.
Time to Symptomatic Progression
Reported analysis
95% CI: 0.315–0.634 · P < 0.0001
The time-to-symptomatic-progression analysis used the ITT population and the log-rank test. The HR of 0.447 corresponds to an estimated event hazard of 44.7% relative to placebo, or a 55.3% lower estimated hazard.
The 95% CI of 0.315–0.634 remains below 1 and describes the uncertainty around the HR estimate. The p-value of <0.0001 indicates evidence against the null hypothesis under the reported analysis. As with the other endpoints, this is not an absolute risk reduction and should not be interpreted as the percentage of patients who avoided symptomatic progression.
11. Complete Registry Statistical Results
| Endpoint | Analysis | HR | 95% CI | P-value |
|---|---|---|---|---|
| MFS by BICR | US Regulatory | 0.280 | 0.227–0.346 | <0.0001 |
| MFS by BICR | Ex-US Regulatory | 0.297 | 0.244–0.362 | <0.0001 |
| Time to Metastasis | US Regulatory | 0.271 | 0.219–0.335 | <0.0001 |
| Time to Metastasis | Ex-US Regulatory | 0.279 | 0.227–0.342 | <0.0001 |
| Progression-free Survival | US Regulatory | 0.291 | 0.238–0.356 | <0.0001 |
| Progression-free Survival | Ex-US Regulatory | 0.300 | 0.247–0.364 | <0.0001 |
| Time to Symptomatic Progression | Registry analysis | 0.447 | 0.315–0.634 | <0.0001 |
The ClinicalTrials.gov record therefore contain seven formal statistical analyses: two for the primary MFS endpoint, two for TTM, two for PFS, and one for time to symptomatic progression. Every posted analysis uses the ITT population, compares placebo with apalutamide, uses the log-rank test, and reports a hazard ratio with a two-sided 95% confidence interval and a p-value of <0.0001.
12. Reading the Pattern Across Endpoints
The numerical pattern is internally consistent across the posted time-to-event endpoints. The two MFS analyses have HRs of 0.280 and 0.297. The TTM analyses have HRs of 0.271 and 0.279. The PFS analyses have HRs of 0.291 and 0.300. The time-to-symptomatic-progression analysis has an HR of 0.447.
These are not seven independent pieces of evidence in the sense of seven separately randomized experiments. They are multiple analyses of outcomes within the same randomized trial. Consequently, the reader should distinguish between consistency of the reported estimates and any claim that each p-value represents a separately powered confirmatory hypothesis.
The ClinicalTrials.gov record identifies the hypothesis type as superiority. They do not provide a multiplicity-adjustment strategy, alpha-allocation scheme, interim-analysis plan, or Bayesian analysis. Those design elements are therefore not inferred here.
13. Intention-to-Treat Analysis in SPARTAN
The ITT framework is particularly important for interpreting a randomized trial. The registry analysis text states that the ITT population included all participants randomized into the study, with treatment assignment designated according to the initial randomization.
What ITT preserves
ITT preserves the original randomized comparison and avoids changing treatment groups based on treatment received after randomization.
What ITT does not guarantee
ITT does not eliminate missing observations, censoring, treatment discontinuation, or other practical issues that can affect a time-to-event analysis.
Why it matters for interpretation
The reported hazard ratios describe the randomized treatment comparison rather than a comparison restricted only to participants who adhered perfectly to assigned treatment.
Population consistency
The statistical analyses posted on ClinicalTrials.gov consistently identify the ITT population as their analysis population.
14. BICR and the MFS Endpoint
The primary endpoint was Metastasis-Free Survival by Blinded Independent Central Review. The registry definition states that MFS begins at randomization and ends at the first occurrence of BICR-confirmed bone or soft tissue distant metastasis or death due to any cause.
This endpoint combines two clinically meaningful event pathways into a single time-to-event outcome. A participant therefore reaches the MFS event either through the specified evidence of distant metastasis or through death, whichever occurs first.
15. Safety Results
The ClinicalTrials.gov record provides serious adverse events by randomized arm as affected participants divided by participants at risk.
| Arm | Serious adverse events | At risk |
|---|---|---|
| Placebo | 92 affected | 398 |
| Apalutamide | 199 affected | 803 |
The raw registry counts should be interpreted alongside their denominators. There were 92 affected participants among 398 at risk in the placebo arm and 199 affected participants among 803 at risk in the apalutamide arm.
The ClinicalTrials.gov record does not provide a formal statistical analysis, confidence interval, or p-value for the serious-adverse-event comparison. This page therefore reports the registry counts without constructing an unreported inferential test.
16. What the Hazard Ratios Do — and Do Not — Mean
An HR of 0.280 for MFS means that the estimated instantaneous event hazard under the fitted time-to-event model is 28.0% of the comparator hazard. In relative terms, that is a 72.0% lower estimated hazard.
It does not mean that 72.0% of patients were spared metastasis or death, nor does it mean that every patient had exactly the same reduction in individual risk.
The 95% CI of 0.227–0.346 gives the statistical uncertainty around the US regulatory MFS HR. It does not describe the range of effects that individual patients experienced.
A p-value of <0.0001 indicates strong evidence against the null hypothesis in the reported analysis. It does not tell us that the treatment effect is "0.0001" in size, and it should not be used as a substitute for the hazard ratio and confidence interval.
The ClinicalTrials.gov record does not provide time-specific survival probabilities, median event times, or absolute event-risk differences. Those measures should therefore not be reconstructed from the hazard ratios alone.
17. Regulatory Analyses and Statistical Interpretation
The presence of both US and ex-US regulatory analyses is an important feature of the registry record. For MFS, TTM, and PFS, the ClinicalTrials.gov record contains separate US and ex-US analyses.
| Endpoint | US HR | Ex-US HR | US 95% CI | Ex-US 95% CI |
|---|---|---|---|---|
| MFS by BICR | 0.280 | 0.297 | 0.227–0.346 | 0.244–0.362 |
| Time to Metastasis | 0.271 | 0.279 | 0.219–0.335 | 0.227–0.342 |
| Progression-free Survival | 0.291 | 0.300 | 0.238–0.356 | 0.247–0.364 |
The estimates are close within each endpoint. That pattern can be described as numerical consistency between the posted regulatory analyses. It would not be appropriate to infer that the analyses are statistically identical, because the ClinicalTrials.gov record does not provide the underlying event data, covariance information, or a formal comparison between the regulatory analyses.
18. Multiplicity, Interim Analysis, and Other Design Features
The trial data identify a superiority hypothesis and provide seven statistical analyses. They do not provide an alpha-spending procedure, a multiplicity-adjustment method, an interim-analysis schedule, a non-inferiority margin, a crossover scheme, a factorial design, or Bayesian methodology.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Hypothesis type | Superiority |
| Non-inferiority margin | Not provided in the ClinicalTrials.gov record |
| Crossover | Not provided in the ClinicalTrials.gov record |
| Factorial design | Not reported; design model is parallel |
| Multiplicity strategy | Not provided in the ClinicalTrials.gov record |
| Interim analysis | Not provided in the ClinicalTrials.gov record |
| Missing-data / imputation method | Not provided in the ClinicalTrials.gov record |
| Bayesian methods | Not reported in the ClinicalTrials.gov record |
This distinction is important. The existence of multiple endpoint analyses does not by itself establish how familywise type I error was controlled. Similarly, the presence of posted results does not establish that an interim analysis occurred. Statistical design features should be described only when supported by the trial record.
19. Limitations
- Clinical cutoff: the registry states that the posted results are up to the clinical cutoff date of 19 May 2017. The estimates should therefore be understood in relation to that cutoff.
- Registry-level reporting: the ClinicalTrials.gov record provides hazard ratios, confidence intervals, p-values, endpoint definitions, and analysis populations, but do not provide the underlying patient-level event and censoring data.
- Absolute time-to-event measures: the ClinicalTrials.gov record does not include median MFS, median TTM, median PFS, or time-specific event-free probabilities, so these cannot be derived reliably from the HRs alone.
- Regulatory analyses: the presence of separate US and ex-US analyses means the estimates should be kept distinct rather than combined into a single derived estimate.
- Multiplicity: the ClinicalTrials.gov record does not specify how the seven posted statistical analyses were incorporated into a familywise error-control strategy.
- Proportional-hazards interpretation: a single hazard ratio is a compact summary of relative event hazard. Its interpretation can be less descriptive if the hazards do not behave approximately proportionally over time; the ClinicalTrials.gov record does not provide a diagnostic assessment of that assumption.
- Safety inference: the registry-reported serious-adverse-event data are counts with denominators, not a formal comparative statistical analysis.
20. Why This Trial Matters Statistically
SPARTAN is a useful teaching case because the registry record concentrates several fundamental survival-analysis concepts into one randomized phase 3 study. The primary endpoint is a time-to-event outcome, the analysis population is ITT, the comparison uses the log-rank test, and the treatment effect is summarized with hazard ratios and two-sided 95% confidence intervals.
| Concept | How it appears in SPARTAN |
|---|---|
| Randomization | Randomized parallel-group phase 3 design with 1207 participants enrolled. |
| Blinding | Triple masking was recorded in the registry. |
| ITT analysis | Posted efficacy analyses use the Intent-to-Treat population. |
| Time-to-event endpoints | MFS, TTM, PFS, and time to symptomatic progression. |
| BICR | The primary MFS endpoint was assessed by blinded independent central review. |
| Kaplan-Meier framework | Time-to-event outcomes are naturally represented using survival-function estimation, although the ClinicalTrials.gov record does not contain the underlying observations needed for a curve. |
| Log-rank test | Reported statistical method for all seven registry-reported efficacy analyses. |
| Hazard ratio | Reported effect measure for all registry-reported efficacy analyses. |
| Confidence intervals | Every registry-reported efficacy analysis reports a two-sided 95% CI. |
| P-values | Every registry-reported efficacy analysis reports P < 0.0001. |
| Superiority testing | The registry identifies superiority as the hypothesis type. |
| Regulatory analyses | US and ex-US analyses are separately reported for MFS, TTM, and PFS. |
21. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
22. Related Statistical Calculators
23. Sources
- ClinicalTrials.gov: NCT01946204 — SPARTAN.
- PubMed: PMID 40132221.
- PubMed: PMID 39889245.
- PubMed: PMID 39405060.
- PubMed: PMID 39163505.
- PubMed: PMID 37951974.
Continue through the Clinical Biostats trial-analysis pathway
Explore the statistical concepts behind randomized time-to-event trials, then apply them with statistical calculators and analysis tools.
24. Record Summary
SPARTAN provides a clear example of a randomized phase 3 time-to-event analysis. The ClinicalTrials.gov record describes a 1207-participant, two-arm, triple-masked study evaluating apalutamide versus placebo, with metastasis-free survival by blinded independent central review as the primary endpoint. The formal analyses use the ITT population, log-rank testing, hazard ratios, two-sided 95% confidence intervals, and a superiority hypothesis.
The primary MFS analyses report HRs of 0.280 and 0.297 for the US and ex-US regulatory analyses, respectively, with both p-values reported as <0.0001. Secondary analyses likewise report HRs below 1 for time to metastasis, progression-free survival, and time to symptomatic progression. The most statistically responsible reading combines these relative effect estimates with their confidence intervals while keeping the endpoint definitions, analysis population, regulatory-analysis distinction, censoring framework, and clinical cutoff in view.