← Clinical Trials
Triple-Negative Breast Cancer Phase 3 Completed NCT02574455

ASCENT: Complete Statistical Analysis of Sacituzumab Govitecan in Triple-Negative Breast Cancer

An independent statistical analysis of the randomized phase 3 ASCENT trial evaluating sacituzumab govitecan versus Treatment of Physician's Choice (TPC) in participants with refractory or relapsed metastatic triple-negative breast cancer.

Trial start: November 7, 2017  ·  Primary completion: March 30, 2020  ·  Enrollment: 529
Scope of this record

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

ASCENT was a randomized, parallel-group, open-label phase 3 trial with 529 participants. The primary registered endpoint was progression-free survival (PFS) by Independent Review Committee (IRC) assessment in the brain metastasis-negative (BM-ve) population. The registry reports formal statistical analyses for the primary endpoint and multiple secondary endpoints.

529
Enrollment
Randomized trial
2
Arms
Sacituzumab govitecan vs TPC
0.387
Primary PFS HR
95% CI 0.305–0.492
<0.0001
Primary PFS P-value
Two-sided
FeatureASCENT
Trial nameASCENT
ClinicalTrials.gov identifierNCT02574455
PhasePhase 3
StatusCompleted
Therapeutic areaOncology
ConditionBreast Cancer
AllocationRandomized
Design modelParallel
MaskingNone
Primary purposeTreatment
Enrollment529
Lead sponsorGilead Sciences
Sponsor typeIndustry
Results postedYes
Outcome measures posted14
Statistical analyses posted12

2. Clinical Question

The registered trial evaluates sacituzumab govitecan against Treatment of Physician's Choice in participants with refractory or relapsed metastatic triple-negative breast cancer. Statistically, the principal question is whether the randomized treatment comparison changes the time from randomization to objective tumor progression or death in the prespecified BM-ve population.

Population

Participants with refractory or relapsed metastatic triple-negative breast cancer. The primary analysis population was the BM-ve Population: randomized participants in strata with no baseline brain metastasis.

Intervention

Sacituzumab govitecan.

Comparator

Treatment of Physician's Choice (TPC).

Primary question

Does sacituzumab govitecan alter progression-free survival relative to TPC in the BM-ve population?

3. Trial Design

01
Randomize529 participants
02
Parallel arms2 treatment groups
03
Open labelNo masking
04
AssessPFS, OS, response and other outcomes
05
AnalyzeLog-rank and CMH methods
ARM 1

Sacituzumab govitecan

  • Sacituzumab govitecan
  • Randomized treatment arm
ARM 2

Treatment of Physician's Choice

  • Treatment of Physician's Choice (TPC)
  • Registry intervention list includes eribulin, capecitabine, gemcitabine and vinorelbine
Design distinction: the registry describes ASCENT as randomized, parallel and unmasked. The primary efficacy comparison therefore rests on randomized assignment, while the absence of masking is an important design feature when interpreting outcomes that can involve investigator assessment.

Trial timeline

November 7, 2017

Trial start

The registry lists November 7, 2017 as the trial start date.

March 30, 2020

Primary completion

The registry lists March 30, 2020 as the primary completion date.

Completed

Results posted

ClinicalTrials.gov reports posted outcome measures and statistical analyses for the completed trial.

4. Endpoints

The registry identifies one primary endpoint and multiple secondary endpoints. The primary endpoint is a time-to-event measure; the posted secondary analyses include additional time-to-event and binary outcomes.

EndpointPopulationRegistry time frameType
Primary: Progression-Free Survival (PFS) by Independent Review Committee (IRC) Assessment in Brain Metastasis Negative (BM-ve) Population BM-ve Population From randomization until objective tumor progression or death (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Time-to-event
PFS by IRC Assessment in the ITT Population ITT Population From randomization until objective tumor progression or death (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Time-to-event
Overall Survival (OS) in BM-ve Population BM-ve Population From the randomization to death from any cause (maximum follow-up duration: 30.8 months) Time-to-event
Overall Survival (OS) in ITT Population ITT Population From the randomization to death from any cause (maximum follow-up duration: 30.8 months) Time-to-event
Objective Response Rate (ORR) by IRC and Investigator Assessment in BM-ve Population BM-ve Population with available data From randomization to the date of progression or death (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Binary
Duration of Response (DOR) by IRC and Investigator Assessment in BM-ve Population BM-ve Population with objective response From the first date of documented response of CR or PR to the date of progression or death (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Time-to-event
Time to Progression (TTP) by Investigator Assessment in BM-ve Population BM-ve Population From randomization until disease progression (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Time-to-event
Time to Progression (TTP) by IRC Assessment in BM-ve Population BM-ve Population From randomization until disease progression (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Time-to-event
Clinical Benefit Rate (CBR) by IRC and Investigator Assessment in BM-ve Population BM-ve Population From randomization to the date of progression or death (assessed every 6 weeks for 9 months and then every 9 weeks thereafter until the occurrence of progression of disease; maximum exposure: 29.6 months) Binary
Registry definition of the primary endpoint: PFS was defined as the time from randomization until objective tumor progression by Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 or death, whichever came first. The registry states that the date of progression was based on observation or radiological assessment of target lesions meeting the prespecified progression criteria.

5. Analysis Populations

PopulationRegistry definition / role
BM-ve PopulationIncluded all randomized participants who were randomized to the strata of no baseline brain metastasis at the time of randomization.
ITT PopulationIncluded all randomized participants.
BM-ve population with available dataUsed for the posted ORR analyses.
BM-ve population with objective responseUsed for the posted DOR analyses.

The distinction between the BM-ve population and the ITT population is statistically important. The primary endpoint was not analyzed in every randomized participant indiscriminately; it was defined for the randomized participants belonging to the no-baseline-brain-metastasis strata. By contrast, the registry separately reports PFS and OS analyses in the full ITT population.

6. Primary Endpoint Result: Progression-Free Survival

The primary registered endpoint was PFS by IRC assessment in the BM-ve population. The registry reports a two-sided log-rank analysis comparing sacituzumab govitecan with TPC.

Primary PFS hazard ratio

0.387

95% CI: 0.305–0.492   ·   P < 0.0001

Analysis: log-rank test   ·   Population: BM-ve

Primary endpointSacituzumab govitecan vs TPC
Effect measureHazard ratio
Estimate0.387
95% CI0.305–0.492
P-value<0.0001
Statistical methodLog-rank test
Analysis populationBM-ve Population
Clinical Biostats interpretation

The estimated hazard ratio of 0.387 is below 1, indicating a lower estimated instantaneous rate of progression or death in the sacituzumab govitecan group relative to TPC under the time-to-event analysis. Expressed as a simple relative-hazard transformation, 1 − 0.387 = 0.613, so the estimate corresponds to an approximately 61.3% lower estimated hazard.

The hazard ratio does not mean that 61.3% of participants avoided progression, that every participant experienced the same reduction, or that the probability of progression or death was reduced by exactly 61.3% at every time point. It is a relative time-to-event measure rather than an absolute risk difference.

The 95% confidence interval of 0.305–0.492 describes uncertainty around the estimated hazard ratio under the analysis framework. It is an interval for the treatment-effect estimate, not a range containing the individual treatment effects experienced by patients.

The P-value of <0.0001 addresses the evidence against the null hypothesis represented by the statistical test. It does not measure the size or clinical importance of the effect. Effect magnitude is better described by the hazard ratio and its confidence interval, together with absolute event-time measures when those are available.

Because this is a time-to-event analysis, interpretation also depends on censoring and on the assumptions underlying the hazard-ratio framework. The ClinicalTrials.gov record does not report a proportional-hazards diagnostic, so the HR should not be interpreted as proving that the relative hazard was constant throughout follow-up.

7. Secondary Progression-Free Survival Analysis

The registry also reports PFS by IRC assessment in the full ITT population. This provides a useful comparison with the primary BM-ve analysis because the ITT population includes all randomized participants.

ITT PFS hazard ratio

0.413

95% CI: 0.330–0.517   ·   P < 0.0001

Analysis: log-rank test   ·   Population: ITT

Clinical Biostats interpretation

The ITT PFS estimate of 0.413 indicates a lower estimated hazard of progression or death in the sacituzumab govitecan group relative to TPC. The estimate corresponds to an approximately 58.7% lower estimated hazard because 1 − 0.413 = 0.587.

The 95% CI of 0.330–0.517 indicates the statistical uncertainty surrounding that estimate. It does not provide a prediction interval for individual patients and should not be read as saying that the treatment effect for every patient lies within that range.

The P-value of <0.0001 provides evidence against the null hypothesis used by the log-rank test; it does not quantify the magnitude of benefit. The hazard ratio and confidence interval provide the principal information about relative effect size and precision.

8. Overall Survival Results

Overall Survival in the BM-ve Population

BM-ve OS hazard ratio

0.481

95% CI: 0.390–0.592   ·   P < 0.0001

Maximum follow-up duration: 30.8 months

Clinical Biostats interpretation

The BM-ve OS hazard ratio of 0.481 indicates a lower estimated instantaneous rate of death for sacituzumab govitecan relative to TPC in the analyzed BM-ve population. The simple relative-hazard interpretation corresponds to an approximately 51.9% lower estimated hazard.

The 95% CI of 0.390–0.592 describes uncertainty around the estimated relative hazard. It does not state the probability that the true effect lies in the interval, nor does it describe individual patient outcomes.

The P-value of <0.0001 is evidence from the specified statistical test and is not an effect-size measure. The absence of a median OS value in the ClinicalTrials.gov record means that a median-survival comparison should not be inferred from the hazard ratio alone.

Overall Survival in the ITT Population

ITT OS hazard ratio

0.514

95% CI: 0.422–0.625   ·   P < 0.0001

Maximum follow-up duration: 30.8 months

Clinical Biostats interpretation

The ITT OS hazard ratio of 0.514 indicates a lower estimated hazard of death in the sacituzumab govitecan group relative to TPC. The estimate corresponds to an approximately 48.6% lower estimated hazard.

The 95% CI of 0.422–0.625 quantifies uncertainty around the estimated hazard ratio. It does not establish that the treatment reduces the probability of death by the same percentage at every follow-up time.

The P-value of <0.0001 addresses statistical evidence under the reported test. It should be considered alongside the magnitude and precision of the estimated effect rather than treated as a measure of clinical importance.

EndpointPopulationHR95% CIP-value
Progression-Free SurvivalBM-ve0.3870.305–0.492<0.0001
Progression-Free SurvivalITT0.4130.330–0.517<0.0001
Overall SurvivalBM-ve0.4810.390–0.592<0.0001
Overall SurvivalITT0.5140.422–0.625<0.0001

9. Objective Response Rate

ORR was analyzed as a binary endpoint in the BM-ve population. The registry reports separate analyses according to IRC assessment and investigator assessment, using the Cochran-Mantel-Haenszel method and odds ratios as the effect measure.

ORR analysisMethodOdds ratio95% CIP-value
IRC assessmentCochran-Mantel-Haenszel10.8595.590–21.095<0.0001
Investigator assessmentCochran-Mantel-Haenszel7.3634.063–13.341<0.0001

IRC-assessed ORR

OR 10.859

95% CI: 5.590–21.095   ·   P < 0.0001

The IRC analysis estimates the odds of objective response to be 10.859 times the odds in the comparator group under the reported model. The corresponding 95% confidence interval is 5.590–21.095.

Clinical Biostats interpretation

An odds ratio is not a risk ratio. An OR of 10.859 means the estimated odds of response were substantially higher in the sacituzumab govitecan group relative to TPC; it does not mean that 10.859 times as many participants responded.

The 95% CI of 5.590–21.095 describes uncertainty around the odds-ratio estimate. The interval remains above 1, consistent with the reported P-value of <0.0001.

The ClinicalTrials.gov record reports the OR and P-value but do not provide the underlying response percentages. Therefore, an absolute response-rate difference cannot be calculated from the ClinicalTrials.gov record without introducing information outside the permitted trial dataset.

Investigator-assessed ORR

OR 7.363

95% CI: 4.063–13.341   ·   P < 0.0001

Clinical Biostats interpretation

The investigator-assessed ORR analysis produced an odds ratio of 7.363. This represents substantially higher estimated odds of response in the sacituzumab govitecan group relative to TPC under the reported analysis.

The 95% CI of 4.063–13.341 quantifies uncertainty around the odds ratio. As with the IRC analysis, it cannot be translated directly into a percentage-point response-rate difference without the underlying group-specific response proportions.

The two ORR analyses are based on different assessment sources. Their estimates should therefore be understood as separate analyses rather than treated as two independent confirmations that can simply be averaged.

10. Duration of Response

DOR was analyzed among BM-ve participants with objective response. The registry reports separate IRC and investigator assessments using log-rank tests and hazard ratios.

DOR analysisHR95% CIP-value
IRC assessment0.4070.150–1.1070.0683
Investigator assessment0.2120.103–0.435<0.0001

DOR by IRC Assessment

IRC-assessed DOR hazard ratio

0.407

95% CI: 0.150–1.107   ·   P = 0.0683

Clinical Biostats interpretation

The IRC-assessed DOR hazard ratio of 0.407 is below 1, corresponding to an approximately 59.3% lower estimated hazard of progression or death among responders under this analysis.

However, the 95% CI of 0.150–1.107 crosses 1, and the reported P-value is 0.0683. The estimate therefore has considerably greater statistical uncertainty than the primary PFS estimate. The point estimate alone should not be interpreted as establishing a definitive difference in DOR.

DOR is also a responder-only endpoint. Its analysis population is not the same as the ITT population, so it answers a different question from the primary PFS analysis.

DOR by Investigator Assessment

Investigator-assessed DOR hazard ratio

0.212

95% CI: 0.103–0.435   ·   P < 0.0001

Clinical Biostats interpretation

The investigator-assessed DOR hazard ratio of 0.212 corresponds to an approximately 78.8% lower estimated hazard of progression or death among responders in the treatment group.

The 95% CI of 0.103–0.435 indicates uncertainty around the estimated relative hazard while remaining below 1. The P-value of <0.0001 provides evidence against the null hypothesis used by the reported test.

The difference between the IRC and investigator estimates illustrates why assessment source matters. It is not appropriate to treat the larger numerical treatment effect from one assessment method as automatically establishing a stronger overall conclusion.

11. Time to Progression

The registry reports TTP separately by investigator and IRC assessment in the BM-ve population. Unlike PFS, TTP is defined around disease progression rather than progression or death in the registry-reported endpoint descriptions.

TTP analysisMethodHR95% CIP-value
Investigator assessmentLog-rank test0.3170.248–0.404<0.0001
IRC assessmentLog-rank test0.4060.315–0.525<0.0001

The investigator-assessed TTP hazard ratio of 0.317 corresponds to an approximately 68.3% lower estimated hazard of progression. The IRC-assessed TTP hazard ratio of 0.406 corresponds to an approximately 59.4% lower estimated hazard of progression. Both confidence intervals remain below 1, and both reported P-values are <0.0001.

Endpoint distinction: PFS and TTP should not be treated as interchangeable. In the registry definitions, PFS counts objective progression or death, whereas TTP is defined around disease progression. That difference can matter when deaths occur before documented progression.

12. Clinical Benefit Rate

CBR was analyzed as a binary endpoint in the BM-ve population using the Cochran-Mantel-Haenszel method. The registry reports separate IRC and investigator assessments.

CBR analysisMethodOdds ratio95% CIP-value
IRC assessmentCochran-Mantel-Haenszel8.5435.055–14.437<0.0001
Investigator assessmentCochran-Mantel-Haenszel7.4924.540–12.364<0.0001

The IRC-assessed CBR analysis produced an odds ratio of 8.543, with a 95% CI of 5.055–14.437. The investigator-assessed analysis produced an odds ratio of 7.492, with a 95% CI of 4.540–12.364. Both analyses reported P < 0.0001.

Clinical Biostats interpretation

Both CBR estimates are odds ratios, so they describe relative odds rather than absolute probabilities. An OR of 8.543 does not mean that 85.43% of participants benefited, nor does it mean that the absolute benefit was 8.543 times larger.

The confidence intervals remain above 1 for both assessment methods. The consistency between the IRC and investigator estimates is descriptive evidence that the direction of the binary endpoint comparison was similar under the two assessment sources, although the analyses remain distinct.

13. Statistical Methodology

Log-rank test for time-to-event endpoints

The registry identifies the log-rank test as the reported method for PFS, OS, DOR and TTP analyses. The log-rank test compares the survival experience of two groups across observed event times while incorporating right-censored observations.

Conceptual interpretation
H0: survival distributions are equivalent between treatment groups

The test evaluates evidence for a difference between the time-to-event distributions. It does not itself produce the hazard ratio; the HR is the accompanying effect measure reported for these analyses.

Hazard ratio

The hazard ratio is a relative measure comparing the instantaneous event rate between two groups. Values below 1 indicate a lower estimated event hazard in the numerator group relative to the comparator under the fitted time-to-event framework.

Conceptual form
HR = hazard in sacituzumab govitecan group ÷ hazard in TPC group

An HR below 1 therefore indicates a lower estimated hazard in the sacituzumab govitecan group. It is not equivalent to a risk ratio, odds ratio, absolute risk difference, or probability of being progression-free.

Kaplan-Meier estimation

Time-to-event endpoints such as PFS, OS, DOR and TTP are naturally suited to Kaplan-Meier estimation because participants can have different lengths of follow-up and may be censored before experiencing the event. The ClinicalTrials.gov record does not provide the Kaplan-Meier curves or the underlying event-by-time dataset, so this page does not reconstruct a curve.

Conceptual Kaplan-Meier estimator
S(t) = ∏ti ≤ t (1 − di/ni)

Here, di is the number of events at an event time and ni is the number at risk immediately before that time.

Cochran-Mantel-Haenszel test

The registry identifies the Cochran-Mantel-Haenszel method for the binary ORR and CBR analyses. This framework can compare treatment groups across categorical strata while accounting for the stratification structure represented in the analysis.

Odds ratios

For the binary endpoints, the registry reports odds ratios rather than risk ratios. The odds of an event are defined as the probability of the event divided by the probability of no event. Consequently, the odds ratio can differ substantially from a risk ratio when the outcome is not rare.

Intention-to-treat analysis

The registry explicitly identifies the ITT population as including all randomized participants and reports ITT analyses for PFS and OS. The ITT principle preserves the comparison created by randomization by analyzing participants according to randomized assignment rather than restricting efficacy analysis to participants who remain on treatment.

14. Statistical Methods Explained

Why was the primary endpoint analyzed with a log-rank test?

PFS is a time-to-event endpoint. Some participants progress or die while others remain event-free at the end of their available follow-up and are therefore censored. The log-rank test is designed to compare two such time-to-event distributions while incorporating the information contributed before censoring.

What does a hazard ratio of 0.387 mean?

It means that the estimated instantaneous rate of progression or death was 0.387 times the corresponding rate in the TPC group under the reported analysis. Equivalently, the simple transformation 1 − 0.387 gives an approximately 61.3% lower estimated hazard. It does not mean that 61.3% of participants avoided an event.

Why is the confidence interval important?

The point estimate alone does not communicate how precisely the treatment effect was estimated. The primary PFS 95% CI of 0.305–0.492 shows the uncertainty around the reported HR. A narrower interval generally indicates greater statistical precision than a wider interval, although precision and clinical importance are separate concepts.

Why doesn't a P-value measure treatment-effect size?

A P-value describes how compatible the observed data are with the null hypothesis used by the statistical test. It depends on both the size of the observed effect and the amount of information available. Therefore, a very small P-value does not mean that an effect is necessarily large, clinically important, or relevant to every patient.

Why does the BM-ve population matter?

The primary endpoint was specifically defined for participants randomized to strata with no baseline brain metastasis. This means the primary HR of 0.387 describes that prespecified analysis population, not automatically every participant enrolled in the trial. The separate ITT PFS analysis addresses the broader randomized population.

Why are IRC and investigator assessments reported separately?

Objective tumor response and progression can involve assessment procedures and judgment. Reporting IRC and investigator analyses separately allows readers to see whether the estimated treatment effect is similar across assessment sources. In ASCENT, the ClinicalTrials.gov record shows separate estimates for ORR, DOR, TTP and CBR according to assessment source.

Why is DOR analyzed only among responders?

DOR begins at the first documented response, so a participant must first have an objective response to enter the DOR analysis population. This makes DOR a different estimand from ITT PFS or OS, which begin at randomization. Consequently, DOR should not be interpreted as a substitute for a randomized comparison of all participants.

15. Interpreting Relative Effects Across Endpoints

EndpointEffect measureEstimateWhat the estimate describes
Primary PFS, BM-veHazard ratio0.387Relative hazard of progression or death
PFS, ITTHazard ratio0.413Relative hazard of progression or death
OS, BM-veHazard ratio0.481Relative hazard of death
OS, ITTHazard ratio0.514Relative hazard of death
ORR, IRCOdds ratio10.859Relative odds of objective response
ORR, investigatorOdds ratio7.363Relative odds of objective response
DOR, IRCHazard ratio0.407Relative hazard of progression or death among responders
DOR, investigatorHazard ratio0.212Relative hazard of progression or death among responders
TTP, investigatorHazard ratio0.317Relative hazard of progression
TTP, IRCHazard ratio0.406Relative hazard of progression
CBR, IRCOdds ratio8.543Relative odds of clinical benefit
CBR, investigatorOdds ratio7.492Relative odds of clinical benefit

These estimates should not be placed on a single numerical scale. A hazard ratio and an odds ratio have different mathematical meanings and refer to different endpoint structures. Even two hazard ratios can answer different questions when one concerns OS, another PFS, and another DOR among responders.

16. Safety Results

The ClinicalTrials.gov record reports serious adverse events by randomized treatment arm. The data do not provide a broader table of individual adverse-event categories, grades, or treatment-related events, so those outcomes are not added here.

Safety measureSacituzumab govitecanTPC
Serious adverse events69/25864/224
Serious adverse events — affected participants / at risk
Sacituzumab govitecan
69 / 258
TPC
64 / 224

The reported counts are presented as affected participants over participants at risk. The ClinicalTrials.gov record does not state a formal statistical comparison of serious adverse-event rates, so no inferential P-value or relative safety effect is added.

Safety interpretation: efficacy and safety use different estimands and should not be combined into a single numerical "net benefit" statistic. The available serious-adverse-event counts describe the number affected and the corresponding at-risk denominators, but the ClinicalTrials.gov record does not provide enough detail to reconstruct the complete safety profile.

17. What the Primary Result Does — and Does Not — Establish

Relative effect

The primary PFS HR of 0.387 indicates a lower estimated hazard of progression or death with sacituzumab govitecan relative to TPC in the BM-ve analysis population.

Absolute effect

The ClinicalTrials.gov record does not provide median PFS, time-specific PFS probabilities, or absolute numbers of PFS events by arm. Therefore, the primary HR should not be converted into an absolute months-of-benefit estimate.

Statistical evidence

The reported P-value of <0.0001 indicates strong statistical evidence against the null hypothesis represented by the reported log-rank test. It does not measure the magnitude of the treatment effect.

Precision

The 95% CI of 0.305–0.492 communicates uncertainty around the primary HR. The interval remains below 1, while still spanning a meaningful range of possible relative effects.

Population

The primary estimate applies to the BM-ve population as defined in the registry. The separate ITT PFS analysis provides a different population-level estimate of 0.413.

18. Multiplicity, Interim Analysis, and Other Design Topics

The ClinicalTrials.gov record identifies the primary endpoint, secondary endpoints, methods, effect measures and analysis populations. They do not provide a prespecified multiplicity-adjustment strategy, interim-analysis schedule, alpha-spending plan, missing-data imputation method, crossover procedure, non-inferiority margin, factorial structure, or Bayesian analysis.

Design topicWhat the ClinicalTrials.gov record supports
MultiplicityNot specified in the ClinicalTrials.gov record.
Interim analysisNot specified in the ClinicalTrials.gov record.
Alpha spendingNot specified in the ClinicalTrials.gov record.
Missing-data / imputationNot specified in the ClinicalTrials.gov record.
Non-inferiority marginNot applicable to the reported hypothesis description; no non-inferiority margin is reported.
CrossoverNot specified in the ClinicalTrials.gov record.
Factorial designNot present; the registered design model is parallel.
Bayesian methodsNot reported among the registry-reported statistical methods.
StratificationThe primary BM-ve population is explicitly defined by randomization strata with no baseline brain metastasis.

This distinction is important. Absence of a method in the ClinicalTrials.gov recordset does not establish that a method was never used elsewhere in the trial's protocol or statistical analysis plan. It means only that the method is not supported by the ClinicalTrials.gov record.

19. Limitations

20. Why This Trial Matters Statistically

ASCENT is a useful teaching case because it combines randomized treatment allocation with several distinct statistical estimands. The primary endpoint is a time-to-event outcome analyzed with a log-rank test and summarized using a hazard ratio, while secondary analyses extend the statistical framework to ITT survival outcomes, binary response endpoints, responder-only DOR, and progression endpoints assessed by different sources.

ConceptHow it appears in ASCENT
RandomizationParticipants were randomized in a parallel-group phase 3 design.
ITT analysisPFS and OS were separately analyzed in the ITT population.
BM-ve analysis populationThe primary PFS endpoint was analyzed among randomized participants in strata with no baseline brain metastasis.
Time-to-event endpointsPFS, OS, DOR and TTP were analyzed with survival-analysis methods.
Log-rank testUsed for the reported PFS, OS, DOR and TTP comparisons.
Hazard ratioUsed to quantify relative effects for time-to-event endpoints.
Cochran-Mantel-Haenszel testUsed for the reported ORR and CBR analyses.
Odds ratioUsed as the effect measure for binary ORR and CBR outcomes.
Confidence intervals95% two-sided intervals accompany the posted effect estimates.
Assessment sourceIRC and investigator assessments are reported separately for several secondary endpoints.
Responder populationDOR analyses were restricted to BM-ve participants with objective response.
Safety denominatorSerious adverse events are reported as affected participants over participants at risk by arm.

21. Related Tutorials

Learn more about the methods used in this trial:

22. Related Statistical Calculators

23. Sources

Continue through the Clinical Biostats statistical library

Explore tutorials and statistical calculators related to survival analysis, binary endpoints, confidence intervals, randomization and clinical-trial methodology.

24. Record Summary

ASCENT provides a detailed example of how a randomized phase 3 oncology trial can generate several related but distinct statistical questions. Its primary registered endpoint, PFS by IRC assessment in the BM-ve population, was analyzed with a log-rank test and produced an HR of 0.387 (95% CI 0.305–0.492; P < 0.0001). The registry also reports ITT PFS, BM-ve and ITT OS, ORR, DOR, TTP and CBR analyses using hazard ratios or odds ratios with two-sided 95% confidence intervals.

The most important statistical lesson is that these estimates should be interpreted according to their endpoint and analysis population. A hazard ratio describes a relative time-to-event effect; an odds ratio describes relative odds for a binary outcome; DOR applies only to responders; and the BM-ve primary analysis is distinct from the ITT analysis. Confidence intervals communicate precision, while P-values address statistical evidence rather than effect size.

Clinical Biostats methodology: A trial-results page should separate reported numerical evidence from statistical interpretation. The purpose is not merely to reproduce registry output, but to explain what each estimand means, which population it represents, what its uncertainty implies, and what conclusions the ClinicalTrials.gov record does and do not support.