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
EV-301 was a randomized, open-label, parallel phase 3 treatment trial with 608 enrolled participants. The registered primary endpoint was overall survival, a time-to-event endpoint analyzed using Kaplan-Meier estimates and a stratified log-rank comparison, with a stratified hazard ratio reported for enfortumab vedotin versus chemotherapy.
| Feature | EV-301 |
|---|---|
| Trial name | EV-301 |
| NCT ID | NCT03474107 |
| Phase | Phase 3 |
| Status | COMPLETED |
| Population | Subjects with previously treated locally advanced or metastatic urothelial cancer |
| Conditions | Ureteral Cancer; Urothelial Cancer; Bladder Cancer |
| Allocation | RANDOMIZED |
| Design | PARALLEL |
| Masking | NONE |
| Primary purpose | TREATMENT |
| Enrollment | 608 |
| Lead sponsor | Astellas Pharma Global Development, Inc. |
| Sponsor type | INDUSTRY |
2. Clinical Question
The primary statistical question was whether enfortumab vedotin produced a different overall-survival experience from chemotherapy in subjects with previously treated locally advanced or metastatic urothelial cancer. The registered hypothesis type was superiority.
Population
Subjects with previously treated locally advanced or metastatic urothelial cancer, including the registered conditions of ureteral cancer, urothelial cancer, and bladder cancer.
Intervention
Enfortumab vedotin. The reported primary comparison specifies enfortumab vedotin 1.25 mg/kg versus chemotherapy.
Comparator
Chemotherapy. The registered intervention list includes docetaxel, vinflunine, and paclitaxel as chemotherapy drugs.
Primary question
Does enfortumab vedotin improve overall survival relative to chemotherapy?
3. Trial Design
Enfortumab vedotin
- Enfortumab vedotin 1.25 mg/kg in the primary statistical comparison
- Compared with chemotherapy for the primary and reported secondary analyses
Chemotherapy
- Chemotherapy comparator in the primary statistical comparison
- Registered intervention list includes docetaxel, vinflunine, and paclitaxel
4. Trial Timeline
Trial start
The EV-301 study began on 27-Jun-2018.
Primary completion
The registry lists 15-Jul-2020 as the primary completion date.
Primary analysis cutoff
The posted overall-survival analysis uses data from randomization through the analysis cut-off date of 15-Jul-2020.
5. Endpoints
| Endpoint | Registry definition / time frame | Endpoint type |
|---|---|---|
| Overall Survival (OS) | OS was defined as the time from the date of randomization until the documented date of death from any cause. Participants who were still alive at the time of data cutoff date were to be censored at the last known alive date or at the data cutoff date, whichever was earlier. Time frame: from randomization until the analysis cut-off date of 15-Jul-2020 (median OS follow-up was 11.10 months). | Time-to-event |
| Progression Free Survival on Study Therapy (PFS1) as Per RECIST v1.1 | From randomization until the analysis cut-off date of 15-Jul-2020 (median OS follow-up was 11.10 months). | Time-to-event |
| Overall Response Rate (ORR) as Per RECIST V1.1 | From randomization until the analysis cut-off date of 15-Jul-2020 (median OS follow-up was 11.10 months). | Binary |
| Disease Control Rate (DCR) as Per RECIST V1.1 | From randomization until the analysis cut-off date of 15-Jul-2020 (median OS follow-up was 11.10 months). | Binary |
The registry lists 9 outcome measures in total and 4 posted statistical analyses. The ClinicalTrials.gov record identifies one primary endpoint analysis and three secondary endpoint analyses.
6. Statistical Methodology
Kaplan-Meier estimation
The registered definition of overall survival explicitly states that OS was analyzed using Kaplan-Meier estimates. This is appropriate for a time-to-event endpoint because not every participant necessarily experiences death before the analysis cutoff.
Here, di represents events at an event time and ni represents participants at risk immediately before that time. The resulting survival function estimates the probability of remaining event-free beyond a given time.
Stratified log-rank test
The primary OS comparison used a stratified log-rank test. The ClinicalTrials.gov record identifies this as the reported method but do not provide the stratification factors used for the test. The analysis therefore should not be described as using particular stratification variables unless those variables are explicitly reported in the underlying record.
Stratified Cox proportional-hazards model
The primary OS analysis reports a stratified hazard ratio. The analysis notes specify a Cox proportional hazards model with treatment, ECOG PS, geographic region and liver metastasis as the explanatory variables.
A hazard ratio is a relative time-to-event measure. It is not a median-survival ratio, an absolute risk difference, or the proportion of participants who benefit.
Cochran-Mantel-Haenszel test
ORR and DCR were analyzed using a stratified Cochran-Mantel-Haenszel test. This approach compares categorical outcomes across treatment groups while accounting for stratification.
Analysis populations
The primary OS and PFS1 analyses were reported for the FAS Population. The ORR analysis used a Response Evaluable Set, defined in the ClinicalTrials.gov record as all participants in the FAS who had measurable disease per RECIST v1.1 per investigator at baseline. DCR was analyzed in the RES Population.
| Endpoint | Analysis population | Statistical method | Effect measure |
|---|---|---|---|
| OS | FAS Population | Stratified log-rank; Cox proportional hazards model | Stratified hazard ratio |
| PFS1 | FAS Population | Stratified log-rank; Cox proportional hazards model | Stratified hazard ratio |
| ORR | Response Evaluable Set | Stratified Cochran-Mantel-Haenszel | P-value reported |
| DCR | RES Population | Stratified Cochran-Mantel-Haenszel | P-value reported |
7. Primary Result: Overall Survival
The primary endpoint was overall survival. The analysis compared enfortumab vedotin 1.25 mg/kg with chemotherapy in the FAS Population using a stratified log-rank test. The associated Cox proportional hazards model included treatment, ECOG PS, geographic region and liver metastasis as explanatory variables.
Stratified hazard ratio for overall survival
95% CI: 0.556–0.886 · P = 0.00142
Two-sided 95% confidence interval · Superiority hypothesis
| Primary endpoint | Enfortumab vedotin | Chemotherapy | Statistical result |
|---|---|---|---|
| Overall Survival (OS) | Enfortumab vedotin 1.25 mg/kg | Chemotherapy | Stratified HR 0.702; 95% CI 0.556–0.886; P = 0.00142 |
An OS hazard ratio of 0.702 means that, under the fitted time-to-event model, the estimated instantaneous hazard of death was approximately 29.8% lower in the enfortumab vedotin group relative to chemotherapy, because 1 − 0.702 = 0.298.
The hazard ratio does not mean that 29.8% of patients avoided death, that survival time increased by 29.8%, or that each individual participant experienced exactly a 29.8% reduction in risk. It is a model-based relative measure of the event rate over time.
The two-sided 95% confidence interval of 0.556–0.886 describes statistical uncertainty around the estimated hazard ratio under the model and sampling framework. It does not describe the range of effects experienced by individual patients.
The P = 0.00142 value addresses evidence against the null hypothesis in the specified superiority comparison. It does not measure the magnitude or clinical importance of the treatment effect. The effect magnitude is described by the hazard ratio and its confidence interval.
Because the estimate comes from a Cox proportional-hazards model, interpretation of a single hazard ratio also depends on the proportional-hazards framework. The ClinicalTrials.gov record does not provide a diagnostic assessment of that assumption.
8. Secondary Result: Progression-Free Survival
Progression Free Survival on Study Therapy (PFS1) as Per RECIST v1.1 was analyzed from randomization until the 15-Jul-2020 analysis cutoff. The analysis population was the FAS Population, and the reported comparison was enfortumab vedotin 1.25 mg/kg versus chemotherapy.
Stratified hazard ratio for PFS1
95% CI: 0.505–0.748 · P < 0.00001
Two-sided 95% confidence interval · Superiority hypothesis
| Secondary endpoint | Analysis population | Statistical result |
|---|---|---|
| Progression Free Survival on Study Therapy (PFS1) as Per RECIST v1.1 | FAS Population | Stratified HR 0.615; 95% CI 0.505–0.748; P < 0.00001 |
A PFS1 hazard ratio of 0.615 corresponds to an estimated instantaneous hazard approximately 38.5% lower in the enfortumab vedotin group relative to chemotherapy under the fitted model, because 1 − 0.615 = 0.385.
The estimate does not mean that 38.5% of participants were progression-free, nor does it imply that individual patients experienced an identical 38.5% reduction in their probability of progression or death.
The 95% confidence interval of 0.505–0.748 quantifies uncertainty around the estimated relative hazard. Because the entire interval is below 1, the reported interval is consistent with a lower estimated hazard in the enfortumab vedotin group under the model.
The reported P < 0.00001 is evidence against the null hypothesis specified for this superiority comparison. It should not be interpreted as a measure of how large the treatment effect is.
As with OS, a Cox-model hazard ratio is subject to the proportional-hazards framework. The ClinicalTrials.gov record does not provide a formal assessment of proportionality.
9. Secondary Results: Tumor Response
Overall Response Rate
Overall Response Rate (ORR) as Per RECIST V1.1 was analyzed in the Response Evaluable Set. The registry reports a stratified Cochran-Mantel-Haenszel analysis comparing enfortumab vedotin 1.25 mg/kg with chemotherapy.
Stratified Cochran-Mantel-Haenszel test
Superiority hypothesis
The ClinicalTrials.gov record reports the P-value but do not provide an ORR estimate or confidence interval.
The reported P < 0.001 indicates evidence against the null hypothesis for the specified superiority comparison of ORR using the stratified Cochran-Mantel-Haenszel test.
Because the ClinicalTrials.gov record does not contain the response percentages or a confidence interval, the statistical result cannot be translated into a numerical difference in response rates from the ClinicalTrials.gov record.
A P-value alone does not establish the magnitude of an effect. For a binary endpoint such as ORR, an effect estimate such as a risk difference, risk ratio, odds ratio, or the underlying response percentages would provide the magnitude and direction of the observed difference.
Disease Control Rate
Disease Control Rate (DCR) as Per RECIST V1.1 was analyzed in the RES Population using a stratified Cochran-Mantel-Haenszel test.
Stratified Cochran-Mantel-Haenszel test
Superiority hypothesis
The ClinicalTrials.gov record reports the P-value but do not provide a DCR estimate or confidence interval.
The reported P < 0.001 indicates evidence against the null hypothesis for the specified superiority comparison of DCR using the stratified Cochran-Mantel-Haenszel method.
The P-value does not quantify the size of the difference in disease-control rates. Because no DCR percentages or confidence interval are included in the ClinicalTrials.gov record, a numerical effect-size interpretation would require information beyond this registry dataset.
10. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized treatment grouping as affected participants over participants at risk. These figures should be interpreted as a safety summary rather than as an efficacy endpoint.
| Treatment group | Serious adverse events | Interpretation |
|---|---|---|
| Enfortumab Vedotin | 138/296 | 138 affected participants among 296 at risk |
| Chemotherapy | 128/291 | 128 affected participants among 291 at risk |
11. Analysis Populations and Censoring
The registry distinguishes analysis populations according to endpoint. This matters because the denominator and eligibility for analysis can change the interpretation of a treatment comparison.
| Endpoint | Population | Why it matters |
|---|---|---|
| Overall Survival | FAS Population | The primary time-to-event analysis is based on the specified full-analysis population. |
| PFS1 | FAS Population | Preserves the analysis population specified for the time-to-event efficacy comparison. |
| ORR | Response Evaluable Set | Restricted to participants in the FAS with measurable disease at baseline per investigator. |
| DCR | RES Population | Uses the response-evaluable population specified for the endpoint. |
For OS, participants who were alive at the data cutoff were to be censored at the last known alive date or the data cutoff date, whichever was earlier. This is a standard right-censoring mechanism for survival analysis: the participant contributes observed follow-up time without an observed death event during that period.
12. Statistical Methods Explained
Why was a stratified log-rank test used for overall survival?
Overall survival is a time-to-event endpoint, so the analysis needs to account for both whether an event occurred and when it occurred. The log-rank test compares survival experience across randomized groups while accommodating right-censored observations. The registry specifically reports a stratified log-rank test, indicating that the comparison was structured to account for stratification in the analysis.
What does a hazard ratio of 0.702 mean?
A hazard ratio of 0.702 indicates that the estimated instantaneous hazard of death in the enfortumab vedotin group was 70.2% of the estimated hazard in the chemotherapy group under the fitted model. Equivalently, 1 − 0.702 = 0.298, corresponding to an estimated 29.8% lower hazard. It is not the same as saying that 29.8% fewer participants died or that survival time increased by 29.8%.
Why is the confidence interval important?
The 95% confidence interval places statistical uncertainty around the estimated hazard ratio. For OS, the interval is 0.556–0.886. A point estimate alone can conceal how precisely an effect has been estimated; the interval provides additional information about that uncertainty.
Why does the P-value not measure effect size?
A P-value evaluates evidence against a null hypothesis under a specified statistical model and sampling framework. It is affected by both the magnitude of an observed effect and the amount of information available. The hazard ratio and confidence interval are therefore needed to describe the size and precision of the time-to-event effect.
Why was the Cochran-Mantel-Haenszel test used for ORR and DCR?
ORR and DCR are binary outcomes rather than time-to-event outcomes. The registry reports a stratified Cochran-Mantel-Haenszel test for these endpoints, which provides a way to compare categorical treatment outcomes while accounting for stratification.
Why are the OS and PFS results not simply percentages?
OS and PFS incorporate the timing of events and censoring. A participant who has not yet experienced an event at the analysis cutoff still contributes information to the survival analysis. Kaplan-Meier estimation and Cox modeling are designed to use that incomplete follow-up rather than treating all participants as if they had identical observation periods.
13. Understanding the Primary Statistical Result
The OS hazard ratio of 0.702 indicates a lower estimated hazard of death with enfortumab vedotin than chemotherapy in the specified analysis.
The 95% confidence interval of 0.556–0.886 describes uncertainty around that estimate. It does not describe individual-level variation in treatment benefit.
The reported P = 0.00142 provides evidence against the null hypothesis for the superiority comparison. It is not a measure of effect size or clinical importance.
The PFS1 analysis produced a hazard ratio of 0.615 with a 95% CI of 0.505–0.748 and P < 0.00001, providing a second time-to-event result in the ClinicalTrials.gov record.
14. Multiplicity, Interim Analysis, and Other Design Features
The ClinicalTrials.gov record identifies a superiority framework, one registered primary endpoint, four posted statistical analyses, and nine posted outcome measures. They do not provide a prespecified multiplicity procedure, alpha-spending scheme, interim-analysis schedule, or formal missing-data/imputation strategy.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Hypothesis type | Superiority |
| Primary endpoint count | 1 registered primary endpoint |
| Primary endpoint | Overall Survival (OS) |
| Statistical analyses posted | 4 |
| Multiplicity procedure | Not specified in the ClinicalTrials.gov record |
| Interim analysis | Not specified in the ClinicalTrials.gov record |
| Missing-data/imputation method | Not specified in the ClinicalTrials.gov record |
| Bayesian methods | Not reported in the ClinicalTrials.gov record |
| Crossover | Not reported in the ClinicalTrials.gov record |
| Non-inferiority margin | Not applicable to the reported superiority hypothesis |
| Factorial design | Not reported; design model is parallel |
This distinction is important. The absence of a method in the ClinicalTrials.gov record should not be interpreted as proof that the underlying protocol or statistical analysis plan contained no such procedure. It means only that the method is not supported by the ClinicalTrials.gov record.
15. Proportional-Hazards Interpretation
Both OS and PFS1 were summarized with stratified hazard ratios from Cox proportional-hazards models. A single hazard ratio provides a compact summary of relative event hazards, but its most straightforward interpretation assumes that the relative hazard is reasonably stable over time.
This is one reason a complete survival analysis normally benefits from viewing Kaplan-Meier curves or other time-specific summaries in addition to a single hazard ratio. The ClinicalTrials.gov record provides the Kaplan-Meier methodology for OS but do not include the underlying survival curve values needed to reconstruct one accurately.
16. What the Statistical Results Do — and Do Not — Establish
What the OS result establishes statistically
The posted analysis estimates a hazard ratio of 0.702 for death, with a two-sided 95% CI of 0.556–0.886 and P = 0.00142, under the reported stratified log-rank/Cox analysis.
What it does not establish
The hazard ratio does not identify an absolute survival difference, a median survival difference, or an individual patient's probability of benefit.
What the PFS result adds
The PFS1 hazard ratio of 0.615 provides a separate time-to-event comparison based on the FAS Population and RECIST v1.1 endpoint.
What the response analyses add
The ORR and DCR analyses use categorical-data methodology and report P-values, but the ClinicalTrials.gov record does not provide the corresponding response-rate estimates.
17. Important Limitations and Interpretation Issues
- Limited numerical result reporting: the ClinicalTrials.gov record provides the OS and PFS1 hazard ratios with confidence intervals and P-values, but do not provide median OS, median PFS, or numerical ORR/DCR estimates.
- Three-arm design: the trial is identified as having 3 arms, while the posted statistical analyses are reported as enfortumab vedotin versus chemotherapy. The ClinicalTrials.gov record does not provide the allocation counts for all three arms.
- Analysis-population differences: OS and PFS1 use the FAS Population, while ORR and DCR use response-evaluable populations. These populations should not be treated as interchangeable.
- Censoring: OS participants who remained alive at the cutoff were censored at the earlier of the last known alive date or the data cutoff date.
- Hazard-ratio assumptions: the Cox model is a model-based analysis and its conventional interpretation depends on the proportional-hazards framework. The ClinicalTrials.gov record does not report a formal diagnostic.
- Missing-data methods: no imputation or missing-data strategy is provided in the ClinicalTrials.gov record.
- Multiplicity: the ClinicalTrials.gov record does not identify a multiplicity-adjustment procedure. The four posted statistical analyses and nine outcome measures should therefore not automatically be interpreted as a single independently powered family of confirmatory tests.
- Safety denominators: serious adverse events are reported as affected/at-risk counts and should be interpreted using those denominators rather than the overall enrollment.
18. Why This Trial Matters Statistically
EV-301 is a useful teaching example because the registry combines randomized treatment comparison with several major clinical-trial statistical methods: time-to-event analysis for OS and PFS1, Kaplan-Meier estimation, stratified log-rank testing, Cox proportional-hazards modeling, and stratified Cochran-Mantel-Haenszel testing for binary response outcomes.
| Concept | How it appears in EV-301 |
|---|---|
| Randomization | The trial uses randomized allocation. |
| Parallel design | The registered design model is PARALLEL. |
| Time-to-event endpoint | Overall survival is the registered primary endpoint. |
| Kaplan-Meier estimation | OS was analyzed using Kaplan-Meier estimates. |
| Stratified log-rank test | Used for the primary OS comparison and the PFS1 comparison. |
| Hazard ratio | Reported for OS and PFS1 as the effect measure. |
| Cox model | Used for the reported stratified hazard-ratio analysis. |
| Confidence interval | OS and PFS1 hazard ratios have two-sided 95% confidence intervals. |
| Cochran-Mantel-Haenszel test | Used for ORR and DCR. |
| Analysis populations | FAS and response-evaluable populations are used for different endpoints. |
| Superiority testing | The reported hypothesis type is superiority. |
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Calculators
21. Sources
- ClinicalTrials.gov: NCT03474107.
- PubMed record: PMID 38418343.
- PubMed record: PMID 33577729.
Continue through the Clinical Biostats statistical pathway
Use the trial's endpoints and methods as a starting point for deeper study of survival analysis, categorical-data methods, confidence intervals, and clinical-trial design.
22. Record Summary
EV-301 provides a clear example of a randomized phase 3 oncology trial centered on a time-to-event primary endpoint. The ClinicalTrials.gov record reports an overall-survival hazard ratio of 0.702 with a two-sided 95% confidence interval of 0.556–0.886 and P = 0.00142 for enfortumab vedotin 1.25 mg/kg versus chemotherapy. The secondary PFS1 analysis reports a hazard ratio of 0.615, with a 95% confidence interval of 0.505–0.748 and P < 0.00001. ORR and DCR were analyzed with stratified Cochran-Mantel-Haenszel tests, with P < 0.001 reported for each, while the ClinicalTrials.gov record does not provide their numerical response-rate estimates.
The statistical story is therefore broader than a single P-value. It includes the choice of a time-to-event endpoint, Kaplan-Meier estimation, stratified log-rank testing, Cox modeling, confidence intervals, analysis-population definitions, categorical response analysis, and explicit handling of censoring. Reading these components together provides a more complete understanding of what the posted EV-301 statistical results do and do not show.