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
CASSIOPEIA was a randomized, parallel-group, phase 3 trial in transplant-eligible participants with previously untreated multiple myeloma. The registry describes four arms and two primary endpoints: post-consolidation stringent complete response (sCR) rate and progression-free survival (PFS) post completion of maintenance therapy.
| Feature | CASSIOPEIA |
|---|---|
| Phase | Phase 3 |
| Condition | Multiple Myeloma |
| Population | Transplant-eligible participants with previously untreated multiple myeloma |
| Design | Randomized, parallel |
| Masking | None |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Primary endpoints | Post-Consolidation Stringent Complete Response (sCR) Rate; Progression Free Survival (PFS) Post Completion of Maintenance Therapy |
| Primary endpoint types | Binary; Time-to-event |
| Results posted | Yes |
| Lead sponsor | Intergroupe Francophone du Myelome |
| Sponsor type | Network |
| ClinicalTrials.gov | NCT02541383 |
2. Clinical Question
The registry describes CASSIOPEIA as a study evaluating daratumumab in transplant-eligible participants with previously untreated multiple myeloma. Statistically, the trial addresses two different questions: whether treatment assignment is associated with a higher post-consolidation sCR rate, and whether treatment assignment is associated with longer PFS after the completion of maintenance therapy.
Population
Transplant-eligible participants with previously untreated multiple myeloma.
Interventions
The registry lists bortezomib, thalidomide, and dexamethasone (VTD); VTD plus daratumumab; and daratumumab among the study interventions.
Comparator structure
The posted analyses compare Arm A with Arm B separately in Part 1 and Part 2. The ClinicalTrials.gov record does not provide additional arm-level treatment descriptions sufficient to reconstruct every treatment sequence.
Primary questions
Is post-consolidation sCR rate different between Arm A Part 1 and Arm B Part 1, and is post-maintenance PFS different between Arm A Part 2 and Arm B Part 2?
3. Trial Design
The ClinicalTrials.gov record identifies a first randomization and a second randomization. The primary sCR analysis uses all participants randomized in the first randomization, whereas the primary post-maintenance PFS analysis uses all participants randomized in the second randomization, even if they did not receive any study treatment dose.
4. Study Arms and Randomization Structure
Arm A Part 1
- Included in the first primary comparison.
- Compared with Arm B Part 1 for post-consolidation sCR rate.
- Analysis population: all participants randomized in the first randomization.
Arm B Part 1
- Included in the first primary comparison.
- Compared with Arm A Part 1 for post-consolidation sCR rate.
- Analysis population: all participants randomized in the first randomization.
Arm A Part 2
- Included in the second primary comparison.
- Compared with Arm B Part 2 for PFS post completion of maintenance therapy.
- Analysis population: all participants randomized in the second randomization, even if they did not receive any study treatment dose.
Arm B Part 2
- Included in the second primary comparison.
- Compared with Arm A Part 2 for PFS post completion of maintenance therapy.
- Analysis population: all participants randomized in the second randomization, even if they did not receive any study treatment dose.
| Registry intervention information | Listed intervention |
|---|---|
| Drug intervention | Bortezomib (VELCADE), Thalidomide, and Dexamethasone (VTD) |
| Drug intervention | Bortezomib, Thalidomide, Dexamethasone (VTD) + daratumumab |
| Drug intervention | Daratumumab |
The ClinicalTrials.gov record does not provide complete arm-by-arm dosing schedules or treatment durations. Accordingly, this page does not reconstruct treatment sequences beyond the intervention names and the Part 1/Part 2 comparison structure explicitly supported by the registry data.
5. Primary Endpoints
| Endpoint | Registry definition / time frame | Endpoint type | Primary method |
|---|---|---|---|
| Post-Consolidation Stringent Complete Response (sCR) Rate | At day 100 post Autologous Stem Cell Transplant (ASCT), up to 114 days post ASCT. Defined as the percentage of ITT subjects who achieved or maintained sCR status within 30 days of Day 100 post ASCT. sCR status is assessed using the computerized algorithm according to IMWG response criteria and must be achieved on or prior to start of subsequent therapies. Subjects must not die or progress by Day 100 post ASCT. | Binary | Cochran-Mantel-Haenszel test; odds ratio |
| Progression Free Survival (PFS) Post Completion of Maintenance Therapy | From the date of second randomization to either progressive disease or death whichever occurred first, with a median follow-up time of 35.4 months (cut-off for analysis was 26 months after the last rando 2 date).. The definition specifies progressive disease according to the IMWG criteria specified in the protocol, or death, whichever occurs first, with all these considered as events, at the completion of Maintenance therapy. | Time-to-event | Log-rank test; hazard ratio |
6. Statistical Methodology
Cochran-Mantel-Haenszel test
The primary sCR comparison used a Cochran-Mantel-Haenszel test, a method for comparing categorical outcomes across treatment groups while accounting for stratification or other defined strata. The associated effect measure reported by the registry is an odds ratio.
An odds ratio above 1 indicates higher estimated odds of the binary outcome in the numerator group, while an odds ratio below 1 indicates lower estimated odds. Odds are not the same as probabilities or risk.
Log-rank test
The primary PFS comparison used a log-rank test. This is a standard method for comparing time-to-event distributions between randomized groups while accounting for the timing of events and censoring.
Hazard ratio
The registry reports a hazard ratio as the effect measure for PFS. A hazard ratio compares the estimated instantaneous event rates between groups over the analyzed time period. For a progression-free survival endpoint, the event is either progressive disease or death according to the registered definition.
A hazard ratio is a relative time-to-event measure. It is not a median difference, an absolute probability difference, or the proportion of participants who benefit.
Kaplan-Meier estimation
For a time-to-event endpoint such as PFS, Kaplan-Meier estimation is the natural descriptive framework for displaying the estimated event-free probability over time. The registry's posted formal comparison is identified as a log-rank test with a hazard ratio; the ClinicalTrials.gov record does not provide Kaplan-Meier estimates at specific time points or median PFS values, so none are added here.
Confidence intervals
The registry reports two-sided 95% confidence intervals for both primary effect estimates. These intervals quantify uncertainty around the estimated odds ratio or hazard ratio under the corresponding statistical framework. They do not describe the range of outcomes that individual participants will experience.
Superiority testing
Both posted primary analyses are identified as superiority hypotheses. The statistical question is therefore whether the treatment comparison provides evidence of a difference in the prespecified direction of the effect measure, rather than whether one treatment is merely no worse than another within a non-inferiority margin.
7. Primary Result: Post-Consolidation sCR Rate
The first primary analysis compares Arm A Part 1 with Arm B Part 1 using all participants randomized in the first randomization. The endpoint is binary: whether a participant achieved or maintained stringent complete response within the prespecified post-ASCT assessment window.
Odds ratio for post-consolidation sCR rate
95% CI: 1.21–2.12 · P = 0.0010
Cochran-Mantel-Haenszel test · two-sided 95% CI · superiority hypothesis
| Primary endpoint | Comparison | Method | Effect estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Post-Consolidation Stringent Complete Response (sCR) Rate | Arm A Part 1 vs Arm B Part 1 | Cochran-Mantel-Haenszel test | OR 1.6 | 1.21–2.12 | 0.0010 |
An odds ratio of 1.6 means that the estimated odds of the binary sCR outcome were 1.6 times as high in the numerator group, Arm A Part 1, relative to Arm B Part 1, under the reported Cochran-Mantel-Haenszel analysis.
This does not mean that the sCR probability was 1.6 times as high, nor does it mean that 60% more participants achieved sCR. Odds ratios and probability ratios are different quantities, especially when the outcome is not rare.
The two-sided 95% confidence interval of 1.21–2.12 describes statistical uncertainty around the estimated odds ratio. Because the interval lies above 1, the registry's result is consistent with higher odds of sCR in Arm A Part 1 under the stated comparison.
The P-value of 0.0010 measures the evidence against the null hypothesis within the specified testing framework. It does not measure the size or clinical importance of the treatment effect, and it does not represent the probability that the null hypothesis is true.
The analysis is based on the first-randomization population. The definition also imposes important event and timing conditions: participants must not have died or progressed by Day 100 post ASCT, and sCR must be achieved on or before the start of subsequent therapies.
Why the binary endpoint requires careful interpretation
The sCR analysis reduces the outcome at the prespecified assessment window to a binary classification. This makes the result straightforward to compare statistically, but the odds ratio does not communicate the underlying event rates unless those rates are also reported. The ClinicalTrials.gov record provides the odds ratio, confidence interval, and P-value, but not the arm-specific sCR percentages. They are therefore not reconstructed here.
8. Primary Result: PFS Post Completion of Maintenance Therapy
The second primary analysis evaluates progression-free survival from the date of second randomization to progressive disease or death, whichever occurred first. The registry states that all such outcomes are considered events at the completion of maintenance therapy.
Hazard ratio for PFS
95% CI: 0.42–0.68 · P < 0.0001
Log-rank test · two-sided 95% CI · superiority hypothesis
| Primary endpoint | Comparison | Method | Effect estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Progression Free Survival (PFS) Post Completion of Maintenance Therapy | Arm A Part 2 vs Arm B Part 2 | Log-rank test | HR 0.53 | 0.42–0.68 | <0.0001 |
A hazard ratio of 0.53 means that the estimated instantaneous rate of progression or death in Arm A Part 2 was approximately 53% of that in Arm B Part 2 under the reported time-to-event comparison. Equivalently, the estimated hazard was approximately 47% lower in the numerator group.
This does not mean that 47% of participants avoided progression, that 47% of participants were cured, or that every participant experienced a 47% reduction in risk. A hazard ratio is a relative measure of event rates over time.
The two-sided 95% confidence interval of 0.42–0.68 gives the statistical uncertainty around the estimated hazard ratio. The entire interval is below 1, providing evidence of a lower estimated event hazard in Arm A Part 2 under the reported analysis.
The P-value of <0.0001 quantifies evidence against the null hypothesis in the stated log-rank framework. It does not measure effect size, clinical importance, or the probability that the observed treatment effect will be exactly reproduced in another population.
The analysis population is important: it includes all participants randomized in the second randomization even if they did not receive any study treatment dose. That preserves the randomized comparison for the efficacy analysis.
As with other Cox-type interpretations of hazard ratios, a single HR is easiest to interpret when the relative hazards are reasonably represented by a common proportional relationship over time. The ClinicalTrials.gov record does not report a formal assessment of the proportional-hazards assumption, so that assumption should not be treated as verified from the registry result alone.
9. Secondary Endpoint Result: PFS From First Randomization to End of Study
The registry also reports a secondary PFS analysis beginning at the date of first randomization and continuing to progressive disease or death, whichever occurred first. The comparison is Arm A Part 1 up to End of Study versus Arm B Part 1 up to End of Study.
Hazard ratio for secondary PFS
95% CI: 0.52–0.72 · P < 0.0001
Log-rank test · two-sided 95% CI · superiority hypothesis
| Endpoint role | Comparison | Analysis population | Method | Effect | 95% CI | P-value |
|---|---|---|---|---|---|---|
| Secondary | Arm A Part 1 up to End of Study vs Arm B Part 1 up to End of Study | All participants randomized in the first randomization, even if they did not receive any study treatment dose | Log-rank test | HR 0.61 | 0.52–0.72 | <0.0001 |
The reported hazard ratio of 0.61 indicates a lower estimated instantaneous rate of progression or death in Arm A Part 1 relative to Arm B Part 1 over the analyzed follow-up. Expressed as a simple derived interpretation, the estimated hazard was approximately 39% lower in Arm A Part 1.
The 95% confidence interval of 0.52–0.72 indicates the statistical precision of the estimate. It does not mean that individual treatment effects range from 0.52 to 0.72.
The P-value of <0.0001 indicates strong statistical evidence against the null comparison under the reported log-rank framework. It does not tell us how large the treatment effect is; that information comes from the hazard ratio and its confidence interval.
This is a secondary endpoint, not one of the two registered primary endpoints. Its statistical interpretation should therefore remain distinct from the primary efficacy conclusions, particularly when considering the broader family of analyses conducted in a multi-endpoint trial.
10. Primary Results in Context
| Endpoint | Role | Outcome type | Comparison | Effect | 95% CI | P-value |
|---|---|---|---|---|---|---|
| Post-Consolidation sCR Rate | Primary | Binary | Arm A Part 1 vs Arm B Part 1 | OR 1.6 | 1.21–2.12 | 0.0010 |
| PFS Post Completion of Maintenance Therapy | Primary | Time-to-event | Arm A Part 2 vs Arm B Part 2 | HR 0.53 | 0.42–0.68 | <0.0001 |
| PFS From First Randomization up to End of Study | Secondary | Time-to-event | Arm A Part 1 up to End of Study vs Arm B Part 1 up to End of Study | HR 0.61 | 0.52–0.72 | <0.0001 |
These three analyses illustrate why the endpoint definition and analysis clock must accompany every effect estimate. The first primary endpoint is a binary response measure after ASCT, the second primary endpoint begins at second randomization after the maintenance sequence, and the secondary PFS analysis begins at first randomization and extends to the end of the study.
11. Statistical Methods Explained
Why was the Cochran-Mantel-Haenszel test used?
The sCR endpoint is binary, so the analysis asks whether the probability of achieving or maintaining the defined sCR status differs between randomized groups. The Cochran-Mantel-Haenszel framework is designed for categorical comparisons and can account for defined strata when appropriate. The registry reports this method directly, so the analysis is not simply an unadjusted comparison of two percentages.
What does an odds ratio of 1.6 mean?
An odds ratio of 1.6 means the odds of sCR in Arm A Part 1 were estimated to be 1.6 times the odds in Arm B Part 1. The distinction between odds and probability matters. For example, an odds ratio cannot be interpreted as saying that the response probability increased by exactly 60% without knowing the underlying event probabilities.
Why is the PFS endpoint analysed differently from sCR?
PFS records when progression or death occurs, rather than simply whether a participant experienced an outcome by one fixed assessment. Participants can also be censored. The log-rank test is therefore appropriate for comparing the time-to-event distributions, while the hazard ratio provides a relative measure of the event rate over time.
What does an HR of 0.53 mean?
The HR of 0.53 for the primary post-maintenance PFS comparison indicates an estimated instantaneous event rate approximately 53% as large in Arm A Part 2 as in Arm B Part 2. A useful derived expression is that this corresponds to an approximately 47% lower estimated hazard. It does not mean a 47% absolute improvement in PFS probability.
Why does the analysis population matter?
The first primary endpoint uses all participants randomized in the first randomization. The second primary endpoint uses all participants randomized in the second randomization, even if they did not receive any study treatment dose. Analysing participants according to randomized assignment helps preserve the treatment comparison created by randomization.
Why are the confidence intervals important?
The sCR odds-ratio estimate is 1.6 with a 95% CI of 1.21–2.12, while the primary PFS hazard ratio is 0.53 with a 95% CI of 0.42–0.68. The intervals show the uncertainty around those estimates. A P-value alone cannot provide this information, and neither a P-value nor a confidence interval tells us how large an effect individual participants will experience.
12. Reading the Binary Endpoint Correctly
What is measured?
The post-consolidation sCR endpoint classifies participants according to whether they achieved or maintained stringent complete response within the specified post-ASCT window.
What is compared?
Arm A Part 1 is compared with Arm B Part 1 using all participants randomized in the first randomization.
What is estimated?
The registry reports an odds ratio of 1.6 rather than a risk ratio or risk difference.
What is not provided?
The ClinicalTrials.gov record does not include the arm-specific sCR percentages or counts, so they cannot be derived without additional information.
The endpoint definition contains several safeguards that affect interpretation. The sCR status is assessed using a computerized algorithm according to IMWG response criteria, must be achieved on or before the start of subsequent therapies, and requires that the participant has not died or progressed by Day 100 post ASCT. These rules make the endpoint more than a simple laboratory or clinical status measurement: it is a prespecified classification tied to a particular assessment window and event history.
13. Reading the Time-to-Event Endpoints Correctly
The PFS endpoints use progressive disease or death as events. This is important because death is treated as an event even if a participant has not previously been recorded as having progressive disease. The time origin also changes between the primary and secondary PFS analyses.
| Time-to-event analysis | Time origin | Event definition | Method |
|---|---|---|---|
| Primary PFS | Date of second randomization | Progressive disease or death, whichever occurred first | Log-rank test; HR |
| Secondary PFS | Date of first randomization | Progressive disease or death, whichever occurred first | Log-rank test; HR |
This distinction prevents a common statistical error: treating a hazard ratio as though it were independent of the time origin. The same biological event can produce a different statistical estimand depending on when follow-up begins.
14. Confidence Intervals and P-values
The odds ratio of 1.6 has a two-sided 95% CI of 1.21–2.12. The interval lies above 1, the null value for an odds ratio.
The hazard ratio of 0.53 has a two-sided 95% CI of 0.42–0.68. The interval lies below 1, the null value for a hazard ratio.
The hazard ratio of 0.61 has a two-sided 95% CI of 0.52–0.72. The interval lies below 1.
All three posted analyses have P-values below conventional significance thresholds, with the primary sCR analysis reporting 0.0010 and the two PFS analyses reporting <0.0001. The P-values provide evidence against their respective null comparisons, but they should not be used as substitutes for effect estimates or confidence intervals.
15. Safety Results
The registry data posted on ClinicalTrials.gov for CASSIOPEIA reports serious adverse events by arm. Because the available safety information is expressed as affected participants over participants at risk, the figures below are reported exactly as reported in the registry rather than converting them into calculated percentages.
| Arm / part | Serious adverse events affected / at risk |
|---|---|
| VTd - Part 1 | 261/538 |
| DVTd - Part 1 | 264/536 |
| VTd-OBS - Part 2 | 35/215 |
| VTd-DARA - Part 2 | 58/211 |
| DVTd-OBS - Part 2 | 50/229 |
| DVTd-DARA - Part 2 | 43/229 |
These safety figures should be kept conceptually separate from the efficacy analyses. The primary sCR and PFS analyses are efficacy comparisons anchored to randomization, whereas serious adverse events describe safety outcomes among the registry-reported arm-specific at-risk populations.
16. What the Hazard Ratios Do — and Do Not — Mean
The reported HR of 0.53 means that the estimated instantaneous rate of progression or death was approximately 53% as large in Arm A Part 2 as in Arm B Part 2 under the reported analysis. It does not mean that 47% of participants avoided progression or death.
The reported HR of 0.61 means that the estimated instantaneous rate of progression or death was approximately 61% as large in Arm A Part 1 as in Arm B Part 1. The corresponding simple derived interpretation is an approximately 39% lower estimated hazard.
The 95% confidence interval communicates uncertainty around each hazard-ratio estimate. The primary PFS interval of 0.42–0.68 and the secondary PFS interval of 0.52–0.72 indicate that the point estimate should not be treated as an exact population parameter.
A hazard ratio summarizes relative event rates. It does not tell the reader the probability of remaining progression-free at a particular month, the median PFS, or the absolute number of participants experiencing events. Those quantities require additional information that is not included in the ClinicalTrials.gov record.
17. Limitations
- Incomplete arm-level treatment detail: The ClinicalTrials.gov record lists several interventions and Part 1/Part 2 arm labels, but it does not provide complete arm-by-arm dosing schedules or treatment sequences.
- No arm-specific sCR rates: The reported sCR analysis provides an odds ratio, confidence interval, and P-value, but the ClinicalTrials.gov record does not provide the underlying percentages or event counts.
- No median PFS: The ClinicalTrials.gov record does not report median PFS for the primary or secondary PFS analyses.
- No time-specific PFS estimates: Without Kaplan-Meier event and censoring information, time-specific progression-free probabilities cannot be reconstructed.
- Endpoint timing: The primary sCR endpoint is anchored to Day 100 after ASCT, whereas the primary PFS endpoint begins at second randomization. These are distinct estimands.
- Analysis populations: The first primary endpoint uses the first-randomization population, while the second primary endpoint uses the second-randomization population. This difference must be retained when comparing results.
- Proportional hazards: A hazard ratio is most straightforward to summarize when a proportional-hazards interpretation is reasonable. The ClinicalTrials.gov record does not report a formal proportional-hazards assessment.
- Multiplicity information: Two primary endpoints are reported, but the ClinicalTrials.gov record does not provide the alpha-allocation or multiplicity-adjustment scheme.
- Missing-data and censoring rules: The ClinicalTrials.gov record does not provide a complete statistical analysis plan describing all missing-data, censoring, or sensitivity-analysis rules.
- Safety detail: Serious adverse events are available by arm as affected/at-risk counts, but the ClinicalTrials.gov record does not include a fuller safety table.
18. Why This Trial Matters Statistically
CASSIOPEIA is a useful statistical teaching case because its registry results combine a binary efficacy endpoint with time-to-event endpoints and use different randomization stages as the starting point for different analyses.
| Concept | How it appears in CASSIOPEIA |
|---|---|
| Randomization | The study is randomized and has four arms. |
| Parallel design | The registry identifies a parallel design model. |
| Binary endpoint | Post-consolidation sCR rate is a binary primary endpoint. |
| Cochran-Mantel-Haenszel test | Used for the primary sCR comparison. |
| Odds ratio | Reported effect measure for the primary sCR analysis. |
| Confidence interval | Two-sided 95% CIs are reported for the primary effect estimates. |
| Time-to-event endpoint | PFS is defined from randomization to progressive disease or death. |
| Log-rank test | Used for both posted PFS comparisons. |
| Hazard ratio | Reported effect measure for the primary and secondary PFS analyses. |
| Multiple randomization stages | The primary sCR analysis uses first randomization, while primary post-maintenance PFS uses second randomization. |
| Analysis populations | The registry explicitly defines the populations used for each posted analysis. |
| Safety by arm | Serious adverse events are reported for Part 1 and Part 2 arm groups. |
19. A Statistical Reading of the Complete Evidence
The CASSIOPEIA results should be read as a collection of related but distinct estimands rather than as one overall statistic. The first primary result concerns a binary response state after ASCT. Its odds ratio of 1.6 summarizes the relative odds of achieving or maintaining the registered sCR status between Arm A Part 1 and Arm B Part 1.
The second primary result concerns time to progression or death after the second randomization. Its hazard ratio of 0.53 summarizes the relative event rate between Arm A Part 2 and Arm B Part 2. Because the time origin changes, it cannot simply be treated as another version of the first PFS analysis.
The secondary PFS analysis provides a complementary time-to-event comparison beginning at first randomization. Its HR of 0.61 addresses progression or death over the broader study period defined from that earlier time origin.
Together, these analyses illustrate a central principle of clinical-trial statistics: an effect estimate has meaning only in the context of its endpoint definition, time origin, analysis population, comparison groups, statistical method, and uncertainty interval.
20. Important Statistical Interpretation Issues
Binary vs time-to-event
sCR is reduced to a binary outcome at a defined assessment window; PFS preserves information about when progression or death occurs.
Odds ratio vs hazard ratio
The OR and HR are different effect measures and should not be compared numerically as though they represented the same estimand.
First vs second randomization
The primary endpoints use different randomization stages, so their populations and time origins are not interchangeable.
P-value vs effect size
A small P-value indicates evidence against a null hypothesis under the specified model; it does not quantify the magnitude of the treatment effect.
21. Longitudinal Trial Record
Trial start
The registry identifies September 2015 as the study start.
Randomized treatment evaluation
The study is a randomized phase 3 trial with a parallel design and no masking.
Post-consolidation sCR
The first primary endpoint is assessed at Day 100 post ASCT, up to 114 days post ASCT.
Post-maintenance PFS analysis
The second primary endpoint measures PFS from the date of second randomization to progressive disease or death.
Primary completion
The registry identifies August 27, 2020 as the primary completion date.
22. Related Tutorials
Learn more about the methods used in this trial:
23. Related Calculators
24. Sources
- ClinicalTrials.gov: CASSIOPEIA, NCT02541383.
- PubMed: PMID 31171419.
- PubMed: PMID 34529931.
- PubMed: PMID 38889735.
- PubMed: PMID 42607072.
- PubMed: PMID 41894685.
Continue with the underlying statistical methods
Explore the clinical-trial, survival-analysis, and categorical-data methods that provide the statistical framework for interpreting CASSIOPEIA.
25. Record Summary
CASSIOPEIA provides a useful example of how different endpoint types require different statistical estimands. The first primary analysis is a Cochran-Mantel-Haenszel comparison of a binary post-consolidation sCR endpoint, reported as an odds ratio of 1.6 with a two-sided 95% CI of 1.21–2.12 and P = 0.0010. The second primary analysis is a log-rank comparison of post-maintenance PFS, reported as a hazard ratio of 0.53 with a two-sided 95% CI of 0.42–0.68 and P < 0.0001.
A secondary PFS analysis beginning at first randomization reports an HR of 0.61 with a two-sided 95% CI of 0.52–0.72 and P < 0.0001. The most important statistical distinction is that these analyses use different endpoint definitions and different randomization stages as their analysis starting points.
The ClinicalTrials.gov record additionally reports serious adverse events by Part 1 and Part 2 arm. These descriptive safety data should be interpreted separately from the randomized efficacy estimands. Overall, the registry results demonstrate how binary response analysis, odds ratios, time-to-event analysis, hazard ratios, confidence intervals, P-values, and multi-stage randomization can coexist within a single phase 3 clinical-trial framework.