← Clinical Trials
Renal Cell Carcinoma Phase 3 Time-to-Event NCT02811861

CLEAR: Complete Statistical Analysis of Lenvatinib-Based Therapy in Renal Cell Carcinoma

An independent statistical analysis of the randomized phase 3 CLEAR trial evaluating lenvatinib plus everolimus or lenvatinib plus pembrolizumab versus sunitinib alone as treatment of advanced renal cell carcinoma, with emphasis on progression-free survival by independent imaging review.

Trial start: 2016-10-13  ·  Primary completion: 2020-08-28  ·  Status: Active, not recruiting
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

CLEAR was a randomized, open-label, parallel-group phase 3 trial in renal cell carcinoma. The registry records 1069 participants and three treatment arms. The posted primary analyses evaluate progression-free survival by independent imaging review using a stratified log-rank test and hazard ratio.

1069
Enrollment
Randomized trial
3
Treatment arms
Parallel design
0.65
PFS HR
Lenvatinib + everolimus vs sunitinib
0.39
PFS HR
Lenvatinib + pembrolizumab vs sunitinib
FeatureCLEAR
PhasePhase 3
ConditionRenal Cell Carcinoma
Brief titleLenvatinib/Everolimus or Lenvatinib/Pembrolizumab Versus Sunitinib Alone as Treatment of Advanced Renal Cell Carcinoma
DesignRandomized, parallel-group
MaskingNone
Primary purposeTreatment
Enrollment1069
Primary endpoint typeTime-to-event
Primary endpoint analyses posted2
Statistical methodStratified log-rank test
Effect measureHazard ratio
Hypothesis typeSuperiority
Lead sponsorEisai Inc.

2. Clinical Question

The registry describes a three-arm randomized comparison of two lenvatinib-based combinations against sunitinib alone in advanced renal cell carcinoma. The posted primary efficacy analyses address whether each lenvatinib combination produces a different progression-free survival experience from sunitinib.

Population

Participants with renal cell carcinoma, in a phase 3 randomized treatment trial with an enrollment of 1069.

Intervention arms

Lenvatinib 18 mg plus everolimus 5 mg, and lenvatinib 20 mg plus pembrolizumab 200 mg.

Comparator

Sunitinib 50 mg.

Primary question

For each lenvatinib combination, is progression-free survival by independent imaging review superior to sunitinib?

3. Trial Design

01
Randomize1069 participants
02
3 armsTwo lenvatinib combinations + sunitinib
03
FollowTime-to-event endpoint
04
AssessIndependent imaging review
05
CompareStratified log-rank + HR
ARM A

Lenvatinib + Everolimus

  • Lenvatinib 18 mg
  • Everolimus 5 mg
ARM B

Lenvatinib + Pembrolizumab

  • Lenvatinib 20 mg
  • Pembrolizumab 200 mg
ARM C

Sunitinib

  • Sunitinib 50 mg

The design is randomized, parallel, and unmasked. Randomization is the key structural feature for causal comparison: treatment assignment is determined before the outcome is observed, allowing the randomized groups to serve as the basis for the primary efficacy comparison.

Three-arm interpretation: the two posted primary analyses are separate treatment-versus-sunitinib comparisons. They should therefore be read as two distinct statistical comparisons rather than as a single three-group hazard ratio.

4. Endpoints

EndpointRegistered definition / time frameStatistical role
Progression-free Survival (PFS) by Independent Imaging Review (IIR)From the date of randomization to the date of the first documentation of PD or date of death, whichever occurred first, using RECIST 1.1. The registry's displayed time frame ends with “whichever occurred first or up to data cutoff date 28 Aug 2020 (up to approximately 46 months)”.Primary endpoint

The registry defines PFS by independent imaging review as a time-to-event endpoint. Progressive disease was defined using Response Evaluation Criteria in Solid Tumors (RECIST 1.1), with the registry definition describing progression in terms of an increase in the sum of diameters of target lesions.

5. Statistical Methodology

Analysis population

The posted primary analyses use the full analysis set (FAS). The registry defines this population as including all randomized participants regardless of the treatment actually received.

Why the analysis population matters
FAS = all randomized participants, regardless of treatment actually received

This definition keeps the primary treatment comparison anchored to randomization. It avoids changing the comparison simply because some participants receive treatment differently from their assigned arm.

Kaplan-Meier framework

PFS is a time-to-event outcome. Conceptually, Kaplan-Meier estimation describes the probability of remaining progression-free over time while accounting for participants whose follow-up ends before an event is observed.

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

Here, di represents events at time ti, while ni represents participants at risk immediately before that time. The registry extract does not provide the underlying event and censoring data needed to reconstruct the complete curve.

Stratified log-rank test

The registry reports a stratified log-rank test for both primary PFS comparisons. A log-rank test evaluates whether the observed time-to-event experience differs between randomized groups across follow-up, while stratification allows the comparison to account for prespecified strata.

Cox proportional-hazards model

The registry states that the hazard ratio is based on a Cox Proportional Hazards Model including treatment group as a factor. The resulting hazard ratio summarizes the relative event rate under that model.

Hazard-ratio interpretation
HR < 1  →  lower estimated instantaneous event rate in the treatment group

For a PFS endpoint, the event is progression or death according to the registered endpoint definition. The hazard ratio is not a percentage of patients who progress, a relative risk, or a median PFS difference.

6. Primary Results: Lenvatinib 18 mg Plus Everolimus 5 mg vs Sunitinib 50 mg

The first posted primary analysis compares lenvatinib 18 mg plus everolimus 5 mg with sunitinib 50 mg for progression-free survival by independent imaging review. The analysis uses the FAS and a stratified log-rank test.

Stratified hazard ratio for PFS

0.65

95% CI: 0.53–0.80   ·   P < 0.0001

Two-sided 95% confidence interval · Superiority hypothesis

FeatureLenvatinib 18 mg + Everolimus 5 mg vs Sunitinib 50 mg
EndpointProgression-free Survival (PFS) by Independent Imaging Review (IIR)
Endpoint typeTime-to-event
Analysis populationFull analysis set: all randomized participants regardless of treatment actually received
MethodStratified log-rank test
Effect measureStratified hazard ratio
Estimate0.65
95% CI0.53–0.80
P-value<0.0001
Model noteHazard ratio based on a Cox Proportional Hazards Model including treatment group as a factor
Clinical Biostats interpretation

A hazard ratio of 0.65 means that the estimated instantaneous rate of progression or death was about 65% as high in the lenvatinib-plus-everolimus group as in the sunitinib group under the fitted Cox model. Equivalently, 0.65 corresponds to a 35% lower estimated hazard relative to sunitinib.

It does not mean that 35% of participants avoided progression, that every participant experienced a 35% reduction in risk, or that PFS was 35% longer. A hazard ratio is a relative time-to-event measure, not an absolute probability.

The 95% CI of 0.53–0.80 describes uncertainty around the estimated hazard ratio under the model and sampling framework. It does not describe the range of treatment effects for individual participants.

The p-value of <0.0001 addresses the statistical evidence against the relevant null comparison under the trial's analysis framework. It does not measure the magnitude or clinical importance of the treatment effect.

Because the hazard ratio comes from a Cox proportional-hazards model, its interpretation depends on the model framework, including the proportional-hazards assumption. The registry extract does not provide information needed to assess that assumption directly.

7. Primary Results: Lenvatinib 20 mg Plus Pembrolizumab 200 mg vs Sunitinib 50 mg

The second posted primary analysis compares lenvatinib 20 mg plus pembrolizumab 200 mg with sunitinib 50 mg using the same registered PFS endpoint, FAS population, stratified log-rank test, and Cox-model-based hazard ratio framework.

Stratified hazard ratio for PFS

0.39

95% CI: 0.32–0.49   ·   P < 0.0001

Two-sided 95% confidence interval · Superiority hypothesis

FeatureLenvatinib 20 mg + Pembrolizumab 200 mg vs Sunitinib 50 mg
EndpointProgression-free Survival (PFS) by Independent Imaging Review (IIR)
Endpoint typeTime-to-event
Analysis populationFull analysis set: all randomized participants regardless of treatment actually received
MethodStratified log-rank test
Effect measureStratified hazard ratio
Estimate0.39
95% CI0.32–0.49
P-value<0.0001
Model noteHazard ratio based on a Cox Proportional Hazards Model including treatment group as a factor
Clinical Biostats interpretation

A hazard ratio of 0.39 means that the estimated instantaneous rate of progression or death was about 39% as high in the lenvatinib-plus-pembrolizumab group as in the sunitinib group under the fitted Cox model. Equivalently, 0.39 corresponds to a 61% lower estimated hazard relative to sunitinib.

It does not mean that 61% of participants were protected from progression, that 61% of participants benefited, or that PFS duration increased by exactly 61%. The hazard ratio describes a relative event-rate measure rather than an individual-level probability.

The 95% CI of 0.32–0.49 expresses uncertainty around the estimated hazard ratio. It is considerably different from saying that individual participants have hazards confined to this interval; the interval concerns the estimated treatment effect under the statistical model.

The p-value of <0.0001 indicates strong statistical evidence against the null comparison under the specified testing framework. It is not a measure of effect size. A very small p-value can accompany a modest effect in a sufficiently informative trial, while a larger p-value does not by itself establish that an effect is clinically unimportant.

As with the other primary comparison, the Cox-model interpretation depends on the proportional-hazards framework. The ClinicalTrials.gov record does not report diagnostics or alternative time-varying-effect analyses.

8. Comparing the Two Posted Primary Analyses

Primary comparisonHR95% CIP-valueHypothesis
Lenvatinib 18 mg + Everolimus 5 mg vs Sunitinib 50 mg0.650.53–0.80<0.0001Superiority
Lenvatinib 20 mg + Pembrolizumab 200 mg vs Sunitinib 50 mg0.390.32–0.49<0.0001Superiority

Both posted primary analyses have hazard ratios below 1, confidence intervals below 1, and p-values reported as <0.0001. Descriptively, the estimated relative hazard is smaller for the lenvatinib-plus-pembrolizumab comparison than for the lenvatinib-plus-everolimus comparison.

Do not turn descriptive differences into a formal head-to-head claim. The ClinicalTrials.gov record reports each combination against sunitinib; they do not report a formal statistical comparison between the two lenvatinib combinations. Therefore, the two hazard ratios should not be interpreted as establishing that one lenvatinib combination is statistically superior to the other.

The confidence intervals also provide important context. The interval around 0.65 is 0.53–0.80, whereas the interval around 0.39 is 0.32–0.49. These intervals quantify uncertainty for their respective treatment-versus-sunitinib estimates, not uncertainty about the difference between the two experimental regimens.

9. Statistical Methods Explained

Why was a stratified log-rank test used?

PFS is a time-to-event endpoint, so simply comparing proportions at one fixed time would discard information about when progression or death occurred. The log-rank test instead uses the ordering of event times across follow-up. Stratification allows the comparison to incorporate the trial's stratified analysis structure rather than treating all participants as belonging to one unstratified risk set.

What does a hazard ratio of 0.65 mean?

A hazard ratio of 0.65 indicates a modeled event rate of approximately 65% of the comparator rate at a given point in time, under the Cox proportional-hazards framework. It corresponds to a 35% lower estimated hazard, but it does not imply a 35% reduction in the probability of progression for every participant.

What does a hazard ratio of 0.39 mean?

A hazard ratio of 0.39 indicates an estimated event rate approximately 39% as high as the comparator under the fitted model. The complementary interpretation is a 61% lower estimated hazard. Again, this is not the same as saying that 61% of patients avoid progression or that individual PFS times increase by 61%.

Why does the confidence interval matter?

The point estimate is only one estimate of the treatment effect. The 95% confidence interval shows the statistical uncertainty surrounding that estimate. For the lenvatinib-plus-everolimus comparison, the interval is 0.53–0.80; for the lenvatinib-plus-pembrolizumab comparison, it is 0.32–0.49. Neither interval is a prediction interval for individual outcomes.

Why doesn't the p-value measure effect size?

The p-value reflects the compatibility of the observed data with a specified null hypothesis under the statistical testing framework. It depends on both the effect and the amount of information in the study. Consequently, <0.0001 does not mean that the effect is large, nor does it tell us how much longer an individual patient remains progression-free.

Why analyze all randomized participants in the FAS?

The registry defines the FAS as all randomized participants regardless of the treatment actually received. Anchoring efficacy analysis to randomized assignment preserves the comparison created by randomization and avoids selectively excluding participants because of post-randomization treatment behavior.

Why is the Cox proportional-hazards assumption important?

A single Cox hazard ratio provides a compact summary of relative event rates under a proportional-hazards framework. If the relative hazard changes materially over time, one number may not describe the treatment effect completely. The ClinicalTrials.gov record identifies the Cox model but do not provide the diagnostics needed to determine whether proportional hazards held.

10. Understanding PFS as a Time-to-Event Endpoint

Progression-free survival combines two possible events into one endpoint: the first documentation of progressive disease or death, whichever occurs first. This creates a clinically useful time-to-event measure while also creating statistical features that distinguish PFS from a simple binary response outcome.

Event timing matters

A participant progressing early and another progressing later are not treated as equivalent observations in a time-to-event analysis. Their event times contribute different information.

Censoring matters

Participants without an observed event by the end of their usable follow-up can contribute information up to the point at which follow-up ends. The registry extract does not provide the individual censoring records.

Death is an event

Under the registered definition, death can end PFS even if radiographic progression has not previously been documented.

Independent review

The endpoint is specifically described as PFS by Independent Imaging Review, providing the registry-defined assessment framework for the primary endpoint.

11. Safety Results

The ClinicalTrials.gov record reports serious adverse events by treatment arm. These figures should be interpreted as the number affected divided by the number at risk for the corresponding arm.

Treatment armSerious adverse eventsAt risk
Lenvatinib 18 mg Plus Everolimus 5 mg164355
Lenvatinib 20 mg Plus Pembrolizumab 200 mg178352
Sunitinib 50 mg113340
Serious adverse events: affected participants
Lenvatinib + Everolimus
164 / 355
Lenvatinib + Pembrolizumab
178 / 352
Sunitinib
113 / 340

The affected/at-risk figures are reported separately from the primary efficacy analysis. Serious adverse events and PFS answer different questions: one concerns an important category of safety events, while the other concerns time to progression or death. The ClinicalTrials.gov record does not provide a formal statistical comparison of serious adverse-event rates, so none is inferred here.

Safety denominator matters: the registry supplies both the number affected and the number at risk for each arm. Those denominators differ across arms, so comparing raw event counts alone would be misleading.

12. Multiplicity and Multiple Primary Comparisons

The registry records two primary-endpoint analyses, both evaluating the same registered PFS endpoint against sunitinib. This structure creates an important multiplicity question because there are two experimental-versus-control comparisons associated with the primary endpoint.

ComparisonPrimary endpointFormal methodReported p-value
Lenvatinib 18 mg + Everolimus 5 mg vs Sunitinib 50 mgPFS by IIRStratified log-rank test<0.0001
Lenvatinib 20 mg + Pembrolizumab 200 mg vs Sunitinib 50 mgPFS by IIRStratified log-rank test<0.0001

The ClinicalTrials.gov record does not specify an alpha-allocation procedure, hierarchical testing strategy, or other multiplicity-adjustment method for these two comparisons. Therefore, this page reports the posted p-values exactly as given and does not infer an additional multiplicity adjustment.

Statistical caution: two statistically significant treatment-versus-control comparisons do not automatically establish that the two experimental treatments differ from each other. A formal experimental-arm comparison would require its own specified analysis.

13. What the Results Do — and Do Not — Establish

What the registry results establish

The posted analyses report PFS hazard ratios below 1 for both lenvatinib-based comparisons against sunitinib, with two-sided 95% confidence intervals entirely below 1 and p-values reported as <0.0001. Both analyses are identified as superiority analyses.

What the hazard ratios do not establish

The hazard ratios do not provide median PFS, absolute PFS probabilities at a particular time, the percentage of patients who benefit, or individual patient outcomes. Those quantities require different data or summary measures.

What the two hazard ratios do not establish

The values 0.65 and 0.39 are treatment-versus-sunitinib estimates. Their numerical difference is not itself a formal test comparing lenvatinib plus everolimus with lenvatinib plus pembrolizumab.

What the p-values do not establish

The reported <0.0001 values quantify statistical evidence under the respective testing framework. They do not quantify effect size, probability that the treatment is effective, or probability that the null hypothesis is true.

14. Design Features That Shape Interpretation

Allocation
Randomized. Treatment assignment is randomized, creating the core framework for causal treatment comparisons.
Model
Parallel. The three treatment arms are followed as parallel randomized groups.
Masking
None. The registry identifies the study as unmasked.
Primary purpose
Treatment. The trial's primary purpose is treatment evaluation.

These design features have different statistical implications. Randomization primarily supports the validity of between-group efficacy comparisons. The absence of masking can matter particularly for outcomes that involve subjective assessment, although this primary endpoint uses independent imaging review. The parallel design means that participants are assigned to one of the three treatment strategies rather than sequentially receiving each intervention.

15. Trial Timeline

2016-10-13 · Trial start

Study begins

The registry records the CLEAR trial start date as October 13, 2016.

2020-08-28 · Primary completion

Primary completion recorded

The registry records August 28, 2020 as the primary completion date.

Current registry status

Active, not recruiting

The ClinicalTrials.gov record identifies the study status as ACTIVE_NOT_RECRUITING.

16. Limitations

17. Why This Trial Matters Statistically

CLEAR is a useful teaching example because a three-arm randomized trial can generate more than one primary treatment comparison while still using a common time-to-event endpoint and a common survival-analysis framework.

Statistical conceptHow it appears in CLEAR
RandomizationParticipants were randomized among three parallel treatment arms.
Multiple comparisonsTwo primary analyses compare separate lenvatinib combinations with the same sunitinib comparator.
Time-to-event endpointPFS is defined from randomization to progression or death, whichever occurs first.
Independent assessmentThe primary endpoint is PFS by Independent Imaging Review.
Stratified log-rank testThe registry reports this as the primary comparison method.
Hazard ratioThe effect measure for both primary PFS comparisons.
Cox modelThe registry states that the hazard ratios are based on a Cox Proportional Hazards Model including treatment group as a factor.
Full analysis setAll randomized participants are included regardless of treatment actually received.
Confidence intervalsTwo-sided 95% CIs quantify uncertainty around both hazard-ratio estimates.
Safety denominatorsSerious adverse events are reported as affected participants relative to participants at risk in each arm.

The most important statistical lesson is that a hazard ratio is only one part of the evidence. Its interpretation depends on the endpoint definition, analysis population, censoring framework, statistical model, comparison being made, and uncertainty interval. In a multi-arm trial, it is equally important to keep each reported contrast separate from comparisons that were not formally analyzed.

18. Statistical Interpretation of the Primary Evidence

The two posted PFS analyses point in the same descriptive direction: both experimental regimens have hazard-ratio estimates below 1 relative to sunitinib. The lenvatinib-plus-everolimus estimate is 0.65, while the lenvatinib-plus-pembrolizumab estimate is 0.39.

For the first comparison, the 95% confidence interval extends from 0.53 to 0.80. For the second, it extends from 0.32 to 0.49. Because these are intervals for separate treatment-versus-control effects, they should not be used as if they were the confidence interval for the difference between the two experimental treatments.

The two p-values are both reported as <0.0001. That provides strong statistical evidence for the respective superiority comparisons under the reported testing framework, but the p-values themselves do not distinguish the magnitude of the two treatment effects. The effect-size information is carried primarily by the hazard ratios and their confidence intervals.

Reading the evidence correctly: the most defensible summary from the ClinicalTrials.gov record is that both experimental-versus-sunitinib PFS analyses report hazard ratios below 1 with two-sided 95% confidence intervals below 1 and p-values <0.0001. Any conclusion about the relative efficacy of the two experimental regimens would require a separate direct comparison.

19. Related Tutorials

Learn more about the methods used in this trial:

20. Related Calculators

21. Sources

Continue through the Clinical Biostats statistical pathway

Use the related tutorials and calculators to explore the survival-analysis methods that underpin randomized time-to-event comparisons.

22. Record Summary

CLEAR is a phase 3 randomized, parallel-group trial with three treatment arms and 1069 enrolled participants. Its the ClinicalTrials.gov record provide two primary analyses of progression-free survival by independent imaging review: lenvatinib 18 mg plus everolimus 5 mg versus sunitinib 50 mg, with a hazard ratio of 0.65 (95% CI 0.53–0.80; P < 0.0001), and lenvatinib 20 mg plus pembrolizumab 200 mg versus sunitinib 50 mg, with a hazard ratio of 0.39 (95% CI 0.32–0.49; P < 0.0001).

Statistically, the trial illustrates how randomized treatment comparisons can be evaluated with a stratified log-rank test and Cox-model hazard ratios when the endpoint is time to progression or death. It also illustrates why confidence intervals, analysis populations, endpoint definitions, and multiplicity considerations matter alongside the p-value. The two reported hazard ratios describe separate comparisons against sunitinib; the ClinicalTrials.gov record does not establish a formal comparison between the two experimental regimens.

Clinical Biostats methodology: The purpose of this page is to distinguish the numerical evidence reported in the ClinicalTrials.gov record from the statistical interpretation of that evidence. Where the registry does not provide a result or methodological detail, this page does not fill the gap with an unsupported estimate.