This page separates reported trial results from statistical interpretation. Trial-specific numerical results and endpoint definitions are taken only from the ClinicalTrials.gov record.
1. Trial at a Glance
IMvigor010 was a randomized, open-label, parallel-group phase 3 study evaluating atezolizumab versus observation as adjuvant therapy after surgical resection in participants with high-risk muscle-invasive urothelial carcinoma.
| Feature | IMvigor010 |
|---|---|
| Phase | Phase 3 |
| Condition | Carcinoma, Transitional Cell |
| Population | Participants with high-risk muscle-invasive urothelial carcinoma after surgical resection |
| Design | Randomized, parallel-group, unmasked |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 809 |
| Primary endpoint | Disease-Free Survival (DFS), as Assessed by Investigator |
| Results posted | Yes |
| Lead sponsor | Hoffmann-La Roche |
| Trial status | Terminated |
| ClinicalTrials.gov | NCT02450331 |
2. Clinical Question
The central statistical question was whether adjuvant atezolizumab changes disease-free survival compared with observation in participants with high-risk muscle-invasive urothelial carcinoma after surgical resection.
Population
Participants with high-risk muscle-invasive urothelial carcinoma after surgical resection.
Intervention
Atezolizumab as adjuvant therapy.
Comparator
Observation.
Primary question
Does adjuvant atezolizumab improve investigator-assessed disease-free survival relative to observation?
3. Trial Design
Observation
- Observation after surgical resection.
- Serious adverse events: 72 affected of 398 at risk.
Atezolizumab
- Atezolizumab as adjuvant therapy after surgical resection.
- Serious adverse events: 122 affected of 390 at risk.
The randomized parallel design establishes the treatment comparison before subsequent follow-up. Because the primary endpoint is time-to-event, the analysis does not simply compare the proportion of participants who eventually experience an event. Instead, it uses the timing of events and the information contributed by participants who have not yet experienced an event at the end of their observed follow-up.
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Disease-Free Survival (DFS), as Assessed by Investigator | DFS is defined as the time from randomization to the time of first occurrence of a DFS event. DFS events include: local (pelvic) recurrence of UC (including soft tissue and regional lymph nodes); urinary tract recurrence of UC (including all pathological stages and grades); distant metastasis of UC; or death from any cause. Tumor assessment will be performed using radiographic evaluations. | Randomization up to first occurrence of DFS event (up to approximately 50 months) |
| Overall Survival (OS) | Time from randomization until death due to any cause. | Randomization until death due to any cause (up to approximately 80 months) |
| Disease-Specific Survival (DSS), as Assessed by Investigator | Time from randomization until death due to UC. | Randomization until death due to UC (up to approximately 50 months) |
| Distant Metastasis-Free Survival (DMFS) | Time from randomization up to diagnosis of distant metastases or death from any cause. | Randomization up to diagnosis of distant metastases or death from any cause (up to approximately 50 months) |
| Non-Urinary Tract Recurrence-Free Survival (NURFS) | Time from randomization up to time of first occurrence of a NURFS event. | Randomization up to time of first occurrence of a NURFS event (up to approximately 50 months) |
All five posted statistical analyses are time-to-event comparisons. The primary endpoint is investigator-assessed DFS; OS, DSS, DMFS, and NURFS are secondary endpoints in the registry analysis data.
5. Analysis Population and Statistical Framework
The registry defines the ITT population as all randomized patients, whether or not the patient received the assigned treatment. This is the analysis population used for the primary and secondary efficacy analyses reported in the ClinicalTrials.gov record.
| Component | Registered / reported approach |
|---|---|
| Analysis population | Intention-to-treat |
| Groups compared | Observation vs Atezolizumab |
| Primary endpoint | Investigator-assessed DFS |
| Endpoint family | Time-to-event |
| Primary comparison | Log-rank test |
| Effect measure | Hazard ratio |
| Hypothesis type | Superiority |
| Confidence interval | Two-sided 95% |
The registry identifies the hypothesis as superiority and the reported method as a log-rank test. The hazard ratio supplies a relative effect estimate, while the confidence interval quantifies uncertainty around that estimate.
6. Statistical Methodology
Log-rank test
The reported statistical method for the primary DFS comparison is the log-rank test. This test is designed for comparing time-to-event distributions between randomized groups while accounting for the timing of events and censoring.
Rather than asking only whether the final number of DFS events differs between groups, the log-rank framework repeatedly compares the observed number of events with the number expected under the null hypothesis at event times where participants remain at risk.
Hazard ratio
The reported effect measure is the hazard ratio. A hazard ratio below 1 indicates a lower estimated instantaneous event rate in the atezolizumab group relative to observation under the fitted time-to-event comparison.
The hazard ratio is a relative time-to-event measure. It is not a probability, does not equal a percentage of patients who benefit, and should not be interpreted as an absolute reduction in the probability of an event.
Intention-to-treat analysis
The ITT principle keeps participants in the group to which they were randomized for the efficacy analysis. The registry explicitly defines the ITT population as all randomized patients, regardless of whether the assigned treatment was received.
This approach protects the comparison created by randomization. It also means that an ITT hazard ratio describes the effect associated with assignment to the randomized strategy, rather than simply the effect among participants who completed treatment exactly as planned.
Time-to-event analysis and censoring
DFS, OS, DSS, DMFS, and NURFS are all defined using time from randomization to an event. Participants who have not experienced the specified event during their observable follow-up contribute information until the point at which their event status is no longer observed. The ClinicalTrials.gov record does not provide individual-level event and censoring records, so this page does not attempt to reconstruct Kaplan-Meier curves.
Stratified analysis
The registry-reported analysis notes identify stratified analysis for OS, DSS, DMFS, and NURFS. For OS, the registry specifies stratification based on PDL1 status, tumor stage after resection, and nodal status. The ClinicalTrials.gov record does not identify the specific stratification variables for the other endpoints beyond the notation that the analysis was stratified.
7. Primary Result: Disease-Free Survival
The primary endpoint was investigator-assessed disease-free survival, defined from randomization to the first DFS event. The registry-reported statistical analysis compared observation with atezolizumab using a log-rank test in the ITT population.
Hazard ratio for disease-free survival
95% CI: 0.735–1.081 · P = 0.2446
Two-sided 95% confidence interval · Superiority hypothesis
| Primary endpoint | Analysis population | Method | Effect estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Disease-Free Survival (DFS), as Assessed by Investigator | ITT | Log-rank test | HR 0.892 | 0.735–1.081 | 0.2446 |
The estimated hazard ratio of 0.892 compares the estimated instantaneous rate of a DFS event in the atezolizumab group with that in the observation group. Numerically, an HR of 0.892 corresponds to an estimated hazard that is 89.2% of the comparator hazard under the analysis framework; equivalently, the point estimate is associated with an approximately 10.8% lower estimated hazard in the atezolizumab group.
The hazard ratio does not mean that 10.8% of participants avoided recurrence or death, nor does it mean that every participant experienced exactly a 10.8% reduction in risk. It is a relative time-to-event effect estimate.
The two-sided 95% confidence interval of 0.735–1.081 expresses uncertainty around the estimated hazard ratio. Because the interval includes 1, the registry-reported interval is compatible with both a lower and a higher hazard under the model. The width of the interval also shows that the point estimate should not be treated as an exact measure of the underlying treatment effect.
The P-value of 0.2446 addresses the statistical evidence against the specified null hypothesis in the reported test; it does not measure the size, clinical importance, or probability of the treatment effect. The primary analysis is explicitly a superiority analysis, not a non-inferiority analysis.
As with other hazard-ratio analyses, interpretation depends on the time-to-event framework and its assumptions. In particular, a single hazard ratio is most straightforward when the relative hazards are reasonably stable over time. The ClinicalTrials.gov record does not provide the underlying event-by-time information needed to assess that assumption directly.
8. Secondary Result: Overall Survival
Overall survival was analyzed from randomization until death due to any cause. The analysis used the ITT population, a log-rank test, and a hazard ratio. The registry-reported analysis also identifies a stratified analysis based on PDL1 status, tumor stage after resection, and nodal status.
Hazard ratio for overall survival
95% CI: 0.726–1.109 · P = 0.3172
Two-sided 95% confidence interval · Stratified analysis
The OS hazard ratio of 0.897 is the estimated relative hazard of death for atezolizumab versus observation under the reported analysis. The point estimate corresponds to an approximately 10.3% lower estimated hazard of death for the atezolizumab group relative to observation.
This does not mean that atezolizumab reduces an individual's probability of death by 10.3%, and it does not describe median survival or an absolute survival difference. Those are different estimands and are not reported in the ClinicalTrials.gov record.
The 95% CI of 0.726–1.109 crosses 1. Consequently, the point estimate alone should not be interpreted as establishing a treatment difference. The confidence interval provides the more informative description of the uncertainty surrounding the estimated hazard ratio.
The P-value of 0.3172 is evidence from the specified statistical test against its null hypothesis; it is not a measure of treatment effect magnitude. Because the analysis is a superiority analysis, it should not be reinterpreted using non-inferiority logic.
The analysis was stratified by PDL1 status, tumor stage after resection, and nodal status. Stratification can improve alignment between the analysis and the randomized design factors, but it does not eliminate uncertainty in the estimated treatment effect.
9. Secondary Result: Disease-Specific Survival
Disease-specific survival was defined as the time from randomization until death due to urothelial carcinoma. The registry-reported analysis used the ITT population and a log-rank test, with a reported stratified analysis.
Hazard ratio for disease-specific survival
95% CI: 0.626–1.116 · P = 0.2235
Two-sided 95% confidence interval · Stratified analysis
The DSS hazard ratio of 0.836 is the estimated relative hazard of death due to urothelial carcinoma in the atezolizumab group versus observation. The point estimate corresponds to an approximately 16.4% lower estimated hazard under the reported comparison.
The estimate is not an absolute mortality reduction and should not be interpreted as the proportion of participants whose cancer-specific survival was improved.
The 95% CI of 0.626–1.116 includes 1. Thus, the interval reflects uncertainty that encompasses both a lower and higher hazard relative to observation. The P-value of 0.2235 is a test statistic summary, not a measure of effect size or clinical relevance.
10. Secondary Result: Distant Metastasis-Free Survival
DMFS was defined as the time from randomization to diagnosis of distant metastases or death from any cause. The registry analysis used the ITT population and a log-rank test, with the registry-reported analysis marked as stratified.
Hazard ratio for distant metastasis-free survival
95% CI: 0.743–1.134 · P = 0.4291
Two-sided 95% confidence interval · Stratified analysis
The DMFS hazard ratio of 0.918 corresponds to an approximately 8.2% lower estimated hazard of the defined DMFS event for atezolizumab relative to observation at the point estimate.
The endpoint combines diagnosis of distant metastases with death from any cause, so its event definition is broader than distant metastasis alone. The hazard ratio therefore describes the composite time-to-event endpoint as registered rather than a pure measure of metastatic recurrence.
The 95% CI of 0.743–1.134 includes 1, and the P-value is 0.4291. Neither value should be translated into a probability that the treatment works or fails. The confidence interval is the appropriate place to examine the uncertainty around the estimated relative effect.
11. Secondary Result: Non-Urinary Tract Recurrence-Free Survival
NURFS was defined as the time from randomization to the first occurrence of a NURFS event. The registry-reported analysis used the ITT population and a log-rank test, with the analysis identified as stratified.
Hazard ratio for non-urinary tract recurrence-free survival
95% CI: 0.722–1.070 · P = 0.1994
Two-sided 95% confidence interval · Stratified analysis
The NURFS hazard ratio of 0.879 corresponds to an approximately 12.1% lower estimated hazard for the defined NURFS event with atezolizumab relative to observation at the point estimate.
Again, the hazard ratio is not an absolute risk difference and does not indicate that 12.1% of participants avoided recurrence. It summarizes a relative time-to-event comparison.
The 95% CI of 0.722–1.070 crosses 1. The P-value of 0.1994 should be understood as evidence from the reported superiority test, not as a measure of the magnitude or practical importance of the estimated effect.
12. Summary of Reported Efficacy Analyses
| Endpoint | Role | Method | HR | 95% CI | P-value | Stratified? |
|---|---|---|---|---|---|---|
| Disease-Free Survival | Primary | Log-rank | 0.892 | 0.735–1.081 | 0.2446 | Not specified in registry-reported primary analysis |
| Overall Survival | Secondary | Log-rank | 0.897 | 0.726–1.109 | 0.3172 | Yes |
| Disease-Specific Survival | Secondary | Log-rank | 0.836 | 0.626–1.116 | 0.2235 | Yes |
| Distant Metastasis-Free Survival | Secondary | Log-rank | 0.918 | 0.743–1.134 | 0.4291 | Yes |
| Non-Urinary Tract Recurrence-Free Survival | Secondary | Log-rank | 0.879 | 0.722–1.070 | 0.1994 | Yes |
Across the five registry-reported analyses, every point estimate is below 1. That descriptive observation should not be converted into a single overall conclusion, because the endpoints represent different clinical event definitions and the ClinicalTrials.gov record does not provide a complete multiplicity strategy for combining all five tests into one familywise inference.
13. Statistical Methods Explained
Why was a log-rank test used?
All five reported efficacy analyses are time-to-event endpoints. A log-rank test is appropriate for comparing the survival experience of two groups while incorporating the timing of observed events and accommodating censored observations. It therefore uses more information than a simple comparison of event proportions at one fixed time.
What does a hazard ratio of 0.892 mean for DFS?
An HR of 0.892 means that the estimated instantaneous rate of the defined DFS event in the atezolizumab group was 0.892 times the corresponding estimated rate in the observation group under the reported analysis. The point estimate can also be described as an approximately 10.8% lower estimated hazard. It does not mean a 10.8 percentage-point improvement in disease-free survival.
Why does the confidence interval matter?
A point estimate is only one estimate from the observed data. The two-sided 95% confidence interval shows the statistical uncertainty surrounding it. For DFS, the interval is 0.735–1.081; for OS, it is 0.726–1.109. These intervals should be considered alongside the point estimates rather than treating the point estimates as exact quantities.
Why does the P-value not measure effect size?
A P-value evaluates the compatibility of the observed data with a specified null hypothesis under the statistical test. It is affected by the amount of information and does not quantify how large or clinically meaningful the treatment effect is. The hazard ratio and its confidence interval provide the direct description of the estimated relative effect and its uncertainty.
Why use the ITT population?
The registry defines the ITT population as all randomized patients, whether or not they received the assigned treatment. An ITT analysis preserves the treatment assignment generated by randomization and therefore maintains the central basis for a randomized comparison.
What does stratification add to the analysis?
Stratified survival analysis allows the comparison to account for specified factors rather than assuming the same baseline hazard structure across all levels of those factors. For OS in this trial, the registry-reported analysis identifies PDL1 status, tumor stage after resection, and nodal status as stratification factors. Stratification does not remove sampling uncertainty or guarantee that treatment effects are identical across strata.
14. Interpreting Hazard Ratios in This Trial
| Endpoint | HR | Point-estimate interpretation |
|---|---|---|
| DFS | 0.892 | Approximately 10.8% lower estimated event hazard with atezolizumab |
| OS | 0.897 | Approximately 10.3% lower estimated hazard of death with atezolizumab |
| DSS | 0.836 | Approximately 16.4% lower estimated hazard of death due to UC with atezolizumab |
| DMFS | 0.918 | Approximately 8.2% lower estimated hazard of the defined DMFS event with atezolizumab |
| NURFS | 0.879 | Approximately 12.1% lower estimated hazard of the defined NURFS event with atezolizumab |
These are simple interpretations of the reported hazard-ratio point estimates. They are not additional trial results. The calculations follow directly from 1 − HR and are useful for translating a ratio into a relative-hazard description.
15. Multiplicity and Endpoint Hierarchy
The ClinicalTrials.gov record identifies one primary endpoint and four secondary endpoints with statistical analyses posted. They do not provide an alpha-allocation scheme, a formal multiplicity adjustment procedure, or a hierarchical testing sequence.
| Endpoint | Role in the ClinicalTrials.gov record | Statistical analysis |
|---|---|---|
| DFS | Primary | Log-rank test; HR 0.892; 95% CI 0.735–1.081; P = 0.2446 |
| OS | Secondary | Log-rank test; HR 0.897; 95% CI 0.726–1.109; P = 0.3172 |
| DSS | Secondary | Log-rank test; HR 0.836; 95% CI 0.626–1.116; P = 0.2235 |
| DMFS | Secondary | Log-rank test; HR 0.918; 95% CI 0.743–1.134; P = 0.4291 |
| NURFS | Secondary | Log-rank test; HR 0.879; 95% CI 0.722–1.070; P = 0.1994 |
Because the ClinicalTrials.gov record does not specify a multiplicity-control strategy, the five P-values should not be treated as though the page can reconstruct a familywise-error procedure. In particular, the presence of several secondary analyses means that each result should be interpreted according to its registered endpoint role and the statistical framework actually documented in the registry data.
16. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized arm using affected participants divided by those at risk.
| Arm | Serious adverse events affected | At risk | Reported measure |
|---|---|---|---|
| Observation | 72 | 398 | 72/398 |
| Atezolizumab | 122 | 390 | 122/390 |
The registry data do not provide a formal statistical comparison for serious adverse events in the registry-reported analysis set. Accordingly, this page reports the affected and at-risk counts without calculating an unreported P-value, risk ratio, odds ratio, or confidence interval.
17. What the Confidence Intervals Say
DFS
The HR is 0.892 with a two-sided 95% CI of 0.735–1.081. The interval includes 1, so the uncertainty interval spans both sides of the null hazard ratio.
OS
The HR is 0.897 with a two-sided 95% CI of 0.726–1.109. The interval includes 1 and therefore does not isolate a single direction of relative hazard.
DSS
The HR is 0.836 with a two-sided 95% CI of 0.626–1.116. The point estimate is below 1, while the interval extends above 1.
DMFS and NURFS
DMFS has a 95% CI of 0.743–1.134, while NURFS has a 95% CI of 0.722–1.070. Both intervals include 1.
A confidence interval that crosses 1 does not mean that the estimated hazard ratio is exactly 1. It means the data and statistical model produce an interval that includes the null value. Conversely, an HR below 1 should not be described as definitive evidence without considering the corresponding uncertainty and prespecified testing framework.
18. Time Frames and Follow-Up Structure
| Endpoint | Registered follow-up frame |
|---|---|
| DFS | Randomization up to first occurrence of DFS event (up to approximately 50 months) |
| OS | Randomization until death due to any cause (up to approximately 80 months) |
| DSS | Randomization until death due to UC (up to approximately 50 months) |
| DMFS | Randomization up to diagnosis of distant metastases or death from any cause (up to approximately 50 months) |
| NURFS | Randomization up to time of first occurrence of a NURFS event (up to approximately 50 months) |
The different time frames are statistically important because they define the observation window for each estimand. OS has a registered frame extending to approximately 80 months, whereas DFS, DSS, DMFS, and NURFS have registered frames extending to approximately 50 months. These are endpoint-specific registry definitions rather than interchangeable follow-up periods.
19. Limitations
- No individual-level survival data: the ClinicalTrials.gov record contains hazard ratios, confidence intervals, and P-values but not the event and censoring times required to reconstruct the underlying Kaplan-Meier estimates.
- No median survival values: median DFS, OS, DSS, DMFS, and NURFS are not included in the ClinicalTrials.gov record and therefore are not reported here.
- No baseline table: detailed baseline demographic and disease characteristics are not reported and are therefore not inferred.
- No subgroup estimates: aside from the OS stratification factors, the ClinicalTrials.gov record does not provide subgroup-specific treatment effects. No subgroup HRs or forest plots are presented.
- No multiplicity procedure: the ClinicalTrials.gov record does not specify an alpha-allocation strategy or hierarchical testing scheme across the primary and secondary endpoints.
- No interim-analysis information: the ClinicalTrials.gov record does not identify an interim analysis schedule, alpha-spending procedure, or stopping boundary.
- No missing-data or imputation specification: the ClinicalTrials.gov record does not describe a missing-data strategy. For time-to-event outcomes, the page therefore does not infer one.
- No non-inferiority margin: the hypothesis type is superiority, so a non-inferiority margin is not part of the registry-reported analysis framework.
- No crossover information: the ClinicalTrials.gov record does not report treatment crossover, so no crossover adjustment or interpretation is added.
- No Bayesian analysis: the statistical analyses posted on ClinicalTrials.gov identify frequentist log-rank testing and hazard ratios; no Bayesian method is reported.
- Open-label design: masking is listed as none. This is a design characteristic that should be considered when interpreting outcomes involving investigator assessment, although the ClinicalTrials.gov record does not quantify any resulting assessment bias.
20. Why This Trial Matters Statistically
IMvigor010 provides a compact teaching example of how a randomized clinical trial can be structured around a primary time-to-event endpoint and a sequence of related secondary time-to-event outcomes.
| Statistical concept | How it appears in IMvigor010 |
|---|---|
| Randomization | 809 participants were enrolled in a randomized two-arm parallel-group phase 3 study. |
| Intention-to-treat analysis | The ITT population includes all randomized patients regardless of whether assigned treatment was received. |
| Time-to-event endpoints | DFS, OS, DSS, DMFS, and NURFS are all defined by time from randomization to specified events. |
| Log-rank test | The reported statistical method for the primary and secondary efficacy analyses. |
| Hazard ratio | Used as the reported relative effect measure for each efficacy endpoint. |
| Confidence intervals | Two-sided 95% confidence intervals quantify uncertainty around the reported hazard ratios. |
| Stratified analysis | Specified for OS, DSS, DMFS, and NURFS; OS is stratified by PDL1 status, tumor stage after resection, and nodal status. |
| Superiority testing | The primary DFS analysis is explicitly identified as a superiority hypothesis. |
| Endpoint hierarchy | DFS is primary, while OS, DSS, DMFS, and NURFS are secondary in the ClinicalTrials.gov record. |
| Safety denominators | Serious adverse events are reported as affected participants over participants at risk in each arm. |
The educational value lies in the distinction between the estimand, the statistical test, and the effect measure. DFS is the endpoint; the log-rank test is the reported comparison method; and the hazard ratio is the relative effect measure. These are related components of the analysis, but they answer different statistical questions.
21. Primary Endpoint vs Secondary Endpoints
Primary endpoint
DFS is the registered primary endpoint. Its analysis is therefore the central efficacy analysis for the ClinicalTrials.gov record.
Secondary endpoints
OS, DSS, DMFS, and NURFS provide additional time-to-event perspectives on survival and recurrence.
Why endpoint definitions matter
Each endpoint uses a different event definition. DFS includes several recurrence categories and death, while OS is death from any cause and DSS is death due to UC.
Why results should stay separate
A treatment effect on one endpoint should not automatically be substituted for an effect on another endpoint because the estimands are different.
22. Reading the Primary Result Correctly
The primary estimand is investigator-assessed DFS: time from randomization to the first occurrence of a defined DFS event.
The registry defines the ITT population as all randomized patients, whether or not they received assigned treatment.
The reported comparison method is the log-rank test, a survival-analysis method for comparing time-to-event experience between groups.
The HR is 0.892. The point estimate is below 1, corresponding to an approximately 10.8% lower estimated event hazard for atezolizumab relative to observation.
The two-sided 95% CI is 0.735–1.081. It includes 1, so the uncertainty interval spans the null hazard ratio.
The P-value is 0.2446. It summarizes evidence under the reported superiority test and should not be treated as a measure of effect size or clinical importance.
23. Related Tutorials
Learn more about the methods used in this trial:
24. Related Calculators
25. Sources
- ClinicalTrials.gov: NCT02450331 — IMvigor010.
- PubMed: PMID 41720562.
- PubMed: PMID 41144872.
- PubMed: PMID 37601688.
- PubMed: PMID 37500339.
- PubMed: PMID 33721560.
Continue through the Clinical Biostats statistical pathway
Use the related tutorials and calculators to explore the survival-analysis concepts underlying randomized time-to-event trials.
26. Record Summary
IMvigor010 is a randomized phase 3 trial in which the primary endpoint, investigator-assessed DFS, was analyzed as a time-to-event outcome in the ITT population using a log-rank test and a hazard ratio. The reported DFS HR was 0.892, with a two-sided 95% CI of 0.735–1.081 and a P-value of 0.2446.
The registry also reports four secondary time-to-event analyses. OS had an HR of 0.897 (95% CI 0.726–1.109; P = 0.3172), DSS had an HR of 0.836 (95% CI 0.626–1.116; P = 0.2235), DMFS had an HR of 0.918 (95% CI 0.743–1.134; P = 0.4291), and NURFS had an HR of 0.879 (95% CI 0.722–1.070; P = 0.1994). Each result must be interpreted according to its endpoint definition, analysis population, and uncertainty rather than reducing the trial to a single number.
The statistical teaching value of the trial comes from the relationship among randomization, ITT analysis, time-to-event endpoint definitions, log-rank testing, hazard ratios, confidence intervals, and stratified analysis. The ClinicalTrials.gov record also illustrate why an independent analysis should distinguish the primary endpoint from secondary outcomes and should avoid adding unreported medians, subgroup estimates, multiplicity procedures, or survival curves.