← Clinical Trials
Breast Cancer Phase 3 Time-to-Event NCT00553410

SOLE: Complete Statistical Analysis of Letrozole in Early-Stage Breast Cancer

An independent statistical analysis of the randomized phase 3 SOLE trial comparing continuous versus intermittent letrozole in postmenopausal women who had received 4–6 years of hormone therapy for hormone receptor-positive, lymph node-positive, early-stage breast cancer.

Completed  ·  Enrollment: 4884  ·  Primary completion: 2018-04
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. Numerical results on this page are restricted to the trial data posted on ClinicalTrials.gov for NCT00553410.

1. Trial at a Glance

SOLE was a randomized phase 3 trial evaluating continuous versus intermittent letrozole in postmenopausal women who had received 4–6 years of hormone therapy for hormone receptor-positive, lymph node-positive, early-stage breast cancer. The registered primary endpoint was disease-free survival (DFS), analyzed as a time-to-event outcome.

4884
Enrollment
Randomized trial
2
Arms
Continuous vs intermittent
1.08
DFS HR
95% CI .93–1.26
.31
DFS P-value
Two-sided log-rank analysis
FeatureSOLE
PhasePhase 3
ConditionBreast Cancer
PopulationPostmenopausal women who have received 4–6 years of hormone therapy for hormone receptor-positive, lymph node-positive, early-stage breast cancer
AllocationRandomized
Design modelParallel
MaskingNone
Primary purposePrevention
Enrollment4884
Primary endpointDisease-free Survival (DFS)
Primary endpoint typeTime-to-event
Primary analysisIntention-to-treat; log-rank test; hazard ratio
Trial statusCompleted
Lead sponsorETOP IBCSG Partners Foundation

2. Clinical Question

The statistical question was whether the pattern of letrozole administration—continuous versus intermittent—was associated with a difference in disease-free survival in the trial population. The registry classified the primary hypothesis as superiority.

Population

Postmenopausal women who had received 4–6 years of hormone therapy for hormone receptor-positive, lymph node-positive, early-stage breast cancer.

Intervention

Continuous letrozole, identified in the registry as Arm A.

Comparator

Intermittent letrozole, identified in the registry as Arm B.

Primary question

Does continuous letrozole differ from intermittent letrozole with respect to disease-free survival?

3. Trial Design

SOLE used a randomized, parallel-group design with no masking. The trial enrolled 4884 participants and compared two letrozole treatment strategies. Because treatment assignment was randomized, the primary efficacy comparison was based on the intention-to-treat principle.

01
Randomize4884 participants
02
AssignContinuous or intermittent letrozole
03
FollowTime-to-event outcomes
04
CompareDFS and secondary endpoints
05
EstimateHazard ratios and 95% CIs
ARM A

Continuous Letrozole

  • Continuous letrozole strategy
  • Randomized treatment group
  • Included in the intention-to-treat efficacy analysis
  • 2411 participants were at risk for the reported serious-adverse-event analysis
ARM B

Intermittent Letrozole

  • Intermittent letrozole strategy
  • Randomized comparison group
  • Included in the intention-to-treat efficacy analysis
  • 2417 participants were at risk for the reported serious-adverse-event analysis
Open-label design: the registry reports masking as NONE. Thus the treatment strategies were not masked. The statistical analysis of the primary time-to-event endpoint was nevertheless based on randomized treatment assignment rather than on treatment received.

4. Endpoints

EndpointRegistered definition / time frameRole
Disease-free Survival (DFS) Duration of time from randomization to the first indication of the following events: invasive recurrence at local (including recurrence restricted to the breast after breast conserving treatment), regional or distant sites; a new invasive cancer in the contralateral breast; any second (non-breast) invasive malignancy; or a death without prior cancer event. Appearance of DCIS or LCIS either in the ... Primary
Overall Survival 5-year estimates, reported at a median follow-up of 60 months Secondary
Distant Recurrence-free Interval (DRFI) 5-year estimates, reported at a median follow-up of 60 months Secondary
Breast Cancer-free Interval 5-year estimates, reported at a median follow-up of 60 months Secondary

The registry definition for DFS is truncated after “Appearance of DCIS or LCIS either in the ...”. Consistent with the data rule for this page, no additional endpoint-definition text is reported beyond the registry wording provided in the trial data.

5. Statistical Methodology

Primary analysis population: intention-to-treat

The primary DFS analysis was conducted in the intention-to-treat population. In an ITT analysis, participants are analyzed according to their randomized assignment. This preserves the comparison created by randomization and avoids redefining the treatment groups according to adherence or treatment changes after randomization.

Time-to-event analysis

DFS is a time-to-event endpoint. Rather than asking only whether a participant eventually experienced an event, time-to-event analysis incorporates the timing of the event and allows participants without an observed event during follow-up to contribute information through their censoring time.

Log-rank test

The registry reports the log-rank test as the statistical method for the posted analyses. The log-rank test compares the survival experience of the randomized groups across follow-up, accounting for the timing of observed events and censoring.

Conceptual survival comparison
H0: survival experience is the same between randomized groups

For SOLE, the reported hypothesis type was superiority. The log-rank test supplies the hypothesis-testing component, while the hazard ratio supplies an estimate of the relative event hazard.

Hazard ratio

The reported effect measure was the hazard ratio (HR). For the primary DFS comparison, the hazard ratio was 1.08 for continuous letrozole versus intermittent letrozole.

Hazard-ratio interpretation
HR = 1.08  →  estimated hazard ratio: continuous letrozole / intermittent letrozole

An HR above 1 means the estimated instantaneous event rate is higher in the numerator group under the fitted time-to-event comparison. It does not mean that 8% more participants necessarily experienced an event, nor does it describe an absolute difference in disease-free survival.

Confidence interval

The primary DFS estimate was accompanied by a two-sided 95% confidence interval of .93 to 1.26. A confidence interval gives information about the precision of the estimated hazard ratio and should be considered alongside the point estimate and hypothesis test.

Secondary endpoint analyses

Overall survival, distant recurrence-free interval, and breast cancer-free interval were each analyzed in the intention-to-treat population using the log-rank test and hazard ratio, with two-sided 95% confidence intervals and superiority as the hypothesis type.

6. Primary Result: Disease-free Survival

The primary endpoint was disease-free survival, measured as 5-year estimates and reported at a median follow-up of 60 months. The comparison was continuous letrozole versus intermittent letrozole.

Hazard ratio for disease-free survival

1.08

95% CI: .93–1.26   ·   P = .31

Intention-to-treat analysis  ·  Two-sided log-rank test  ·  Superiority hypothesis

Primary endpointContinuous LetrozoleIntermittent LetrozoleEffect estimateP-value
Disease-free Survival (DFS) Randomized Arm A Randomized Arm B HR 1.08
95% CI .93–1.26
.31
Clinical Biostats interpretation

The hazard ratio of 1.08 means that the estimated instantaneous hazard of a DFS event for continuous letrozole relative to intermittent letrozole was 1.08 under the reported time-to-event analysis. Expressed descriptively, this is an estimated hazard 8% higher in the numerator group, not an 8-percentage-point difference in disease-free survival.

The estimate does not mean that continuous letrozole caused 8% more patients to experience a DFS event. A hazard ratio is a relative time-to-event measure, not an absolute event probability.

The two-sided 95% CI of .93–1.26 describes uncertainty around the estimated hazard ratio. It includes 1, meaning the interval is compatible with a range of relative hazards that includes no estimated hazard difference. The interval also shows that the point estimate should not be interpreted as a highly precise estimate of a small relative difference.

The P-value of .31 is evidence from the reported hypothesis test, not a measure of effect size. A P-value does not tell us that the treatment effect is “31%” or that the treatments are 31% similar. It addresses the statistical evidence against the specified null hypothesis under the analysis framework.

Because DFS is a censored time-to-event endpoint, interpretation also depends on the validity of the underlying time-to-event analysis and its censoring assumptions. The ClinicalTrials.gov record identifies the log-rank test and hazard ratio but do not report additional model diagnostics or proportional-hazards assessments.

7. Secondary Results

Overall Survival

Hazard ratio for overall survival

.85

95% CI: .68–1.06   ·   P = .16

5-year estimates, reported at a median follow-up of 60 months

Clinical Biostats interpretation

The reported overall-survival HR of .85 estimates a lower instantaneous hazard of death for continuous letrozole relative to intermittent letrozole under the reported analysis. Descriptively, an HR of .85 corresponds to an estimated hazard 15% lower in the numerator group; it does not mean that 15% more patients survived or that an individual patient's probability of survival increased by 15%.

The 95% CI of .68–1.06 spans 1. This indicates uncertainty that includes both a lower hazard and a value compatible with no hazard difference. The P-value of .16 is the reported result of the superiority hypothesis test and should not be interpreted as the magnitude or probability of a treatment effect.

Distant Recurrence-free Interval

Hazard ratio for DRFI

.88

95% CI: .71–1.09   ·   P = .25

5-year estimates, reported at a median follow-up of 60 months

Clinical Biostats interpretation

The DRFI HR of .88 is an estimated relative hazard for the continuous-letrozole group compared with the intermittent-letrozole group. An HR of .88 corresponds descriptively to an estimated instantaneous hazard 12% lower in the numerator group, but it does not represent a 12-percentage-point improvement in a recurrence-free probability.

The 95% CI of .71–1.09 includes 1, so the interval includes values compatible with no relative hazard difference. The P-value of .25 is a hypothesis-test result rather than a measure of the size or clinical importance of the observed HR.

Breast Cancer-free Interval

Hazard ratio for breast cancer-free interval

.98

95% CI: .81–1.18   ·   P = .84

5-year estimates, reported at a median follow-up of 60 months

Clinical Biostats interpretation

The breast cancer-free interval HR of .98 is very close to 1. Under the reported analysis, the estimated instantaneous event hazard was approximately similar between the randomized groups. The estimate should still be read with its uncertainty rather than treated as an exact equality.

The 95% CI of .81–1.18 contains 1 and spans a range of possible relative hazards around the point estimate. The P-value of .84 indicates little statistical evidence against the null hypothesis in the reported superiority test, but it does not establish that the two treatment strategies are equivalent or identical.

8. Results Summary

EndpointAnalysis populationMethodHR95% CIP-value
Disease-free Survival (DFS)
Primary
Intention-to-treatLog-rank1.08.93–1.26.31
Overall SurvivalIntention-to-treatLog-rank.85.68–1.06.16
Distant Recurrence-free Interval (DRFI)Intention-to-treatLog-rank.88.71–1.09.25
Breast Cancer-free IntervalIntention-to-treatLog-rank.98.81–1.18.84

All four posted analyses use the same basic statistical structure: intention-to-treat analysis, comparison of continuous versus intermittent letrozole, a log-rank test, a hazard ratio as the effect measure, a two-sided 95% confidence interval, and a superiority hypothesis. The primary endpoint is DFS; the other three are secondary endpoints.

9. Statistical Methods Explained

Why was a log-rank test used?

DFS, overall survival, DRFI, and breast cancer-free interval are time-to-event outcomes. A log-rank test is designed to compare the survival experience of randomized groups across follow-up while incorporating the timing of events and accounting for censored observations. The registry specifically identifies the log-rank test as the reported method for all four posted analyses.

What does a hazard ratio of 1.08 mean?

A hazard ratio compares the estimated instantaneous event hazards between two groups. Here, the reported HR of 1.08 uses continuous letrozole as the numerator and intermittent letrozole as the comparator. It therefore represents an estimated hazard 1.08 times that of the comparator, or descriptively an 8% higher estimated hazard. It is not an 8-percentage-point difference in the probability of remaining disease-free.

Why does the confidence interval matter?

The point estimate is only one estimate of the underlying treatment contrast. The 95% confidence interval communicates the statistical precision of that estimate under the analysis framework. For DFS, the interval is .93–1.26, so the data support a range of plausible hazard ratios around 1.08 that includes 1.

What does the P-value of .31 mean?

The P-value quantifies the statistical evidence against the null hypothesis under the specified testing procedure. It does not measure the size of the hazard ratio, the probability that the null hypothesis is true, or the probability that the treatment has no clinical value. Effect size and uncertainty are better described by the HR and its confidence interval.

Why use intention-to-treat analysis?

ITT analysis retains participants in the groups to which they were randomized. This is particularly important in a randomized trial because randomization creates the basis for a fair comparison of treatment strategies. Reassigning participants after randomization according to treatment actually received can introduce post-randomization selection into the efficacy comparison.

Why are the secondary endpoints interpreted separately from DFS?

DFS is the registered primary endpoint, whereas overall survival, DRFI, and breast cancer-free interval are secondary endpoints. Their HRs and P-values provide additional information about different time-to-event outcomes. A secondary endpoint should not be silently treated as if it were the primary endpoint, and a P-value for one endpoint does not establish the result of another.

Does an HR near 1 prove the treatments are equivalent?

No. For example, the breast cancer-free interval HR is .98, which is close to 1, but its 95% CI is .81–1.18. A conventional superiority analysis that does not reject a null hypothesis of no difference is not automatically a formal equivalence or non-inferiority demonstration. The ClinicalTrials.gov record identifies the hypothesis type as superiority.

10. Interpreting Time-to-Event Endpoints

A major statistical feature of SOLE is that all four posted analyses are time-to-event analyses. This means that both when an event occurs and whether a participant remains event-free through a censoring time contribute to the analysis.

Event timing

A participant who experiences a DFS event contributes information about the time at which that event occurred rather than merely being classified as having an event.

Censoring

A participant without an observed event during the relevant observation period can contribute follow-up information until the censoring time.

Relative effect

The hazard ratio summarizes the relative event hazard between the randomized treatment groups.

Hypothesis test

The log-rank test provides the reported formal comparison of survival experience between the two randomized groups.

A useful distinction
Hazard ratio ≠ risk ratio ≠ risk difference

An HR is a time-to-event measure. It should not be translated directly into an absolute probability difference without appropriate survival estimates and a specified time point.

11. Why the Primary Endpoint Requires Careful Reading

The primary DFS result has an HR of 1.08, with a two-sided 95% CI of .93–1.26 and a P-value of .31. These three quantities answer related but different questions.

QuantitySOLE DFS resultWhat it tells us
Hazard ratio1.08The estimated relative event hazard for continuous versus intermittent letrozole.
95% CI.93–1.26The statistical uncertainty around the estimated hazard ratio.
P-value.31The reported evidence from the two-sided superiority hypothesis test.

It is important not to collapse these three quantities into a single statement. The HR describes the estimated effect, the confidence interval describes uncertainty around that estimate, and the P-value describes evidence against the null hypothesis under the specified testing framework.

12. Primary Endpoint Definition and Censoring

The registered DFS endpoint defines disease-free survival as the duration from randomization to the first indication of specified events, including invasive recurrence at local, regional, or distant sites; a new invasive cancer in the contralateral breast; any second non-breast invasive malignancy; or death without a prior cancer event. The registry text also begins a statement concerning DCIS or LCIS but is truncated in the ClinicalTrials.gov record.

Why the definition matters: a composite time-to-event endpoint can count several clinically distinct event types. The statistical result therefore corresponds to the registered DFS definition rather than to recurrence alone or death alone.

13. Secondary Endpoint Structure

The three secondary outcomes broaden the statistical picture beyond DFS. Overall survival focuses on death, while DRFI and breast cancer-free interval represent additional disease-related time-to-event outcomes.

EndpointTime frameHR95% CIP-value
Overall Survival5-year estimates, reported at a median follow-up of 60 months.85.68–1.06.16
Distant Recurrence-free Interval (DRFI)5-year estimates, reported at a median follow-up of 60 months.88.71–1.09.25
Breast Cancer-free Interval5-year estimates, reported at a median follow-up of 60 months.98.81–1.18.84

The secondary HRs are all below 1 except the primary DFS HR, which is 1.08, but the point estimates alone should not be interpreted as evidence that one treatment strategy is uniformly better or worse. Each estimate has its own confidence interval, and all three secondary confidence intervals include 1.

14. Serious Adverse Events

The ClinicalTrials.gov record reports serious adverse events by randomized treatment arm as affected participants divided by those at risk.

Safety measureContinuous LetrozoleIntermittent Letrozole
Serious adverse events1004 / 24111052 / 2417
Serious adverse events: affected / at risk
Continuous Letrozole
1004 / 2411
Intermittent Letrozole
1052 / 2417

The affected counts and at-risk denominators are reported exactly as reported in the registry. They should be kept distinct from the efficacy analysis: safety data describe adverse-event occurrence, whereas DFS and the other posted endpoints are time-to-event efficacy outcomes analyzed under the intention-to-treat framework.

Denominator caution: the ClinicalTrials.gov record gives affected and at-risk counts but does not provide a registry-level statistical comparison, confidence interval, or P-value for serious adverse events. A formal safety comparison should therefore not be inferred from the counts alone.

15. What the Hazard Ratios Do — and Do Not — Mean

Primary DFS HR

An HR of 1.08 means the estimated instantaneous DFS-event hazard for continuous letrozole relative to intermittent letrozole was 1.08 under the reported time-to-event analysis. It does not mean that 8% of participants had an event, that 8% more participants experienced an event, or that an individual patient's disease-free probability changed by 8%.

Overall survival HR

An HR of .85 means the estimated instantaneous death hazard for continuous letrozole relative to intermittent letrozole was .85 under the reported analysis. It does not mean that survival probability increased by 15 percentage points or that every participant experienced a 15% reduction in mortality risk.

DRFI HR

An HR of .88 is an estimated relative hazard for the distant recurrence-free interval. It is not a direct estimate of the proportion of patients who avoided distant recurrence.

Breast cancer-free interval HR

An HR of .98 is close to 1, but the corresponding confidence interval of .81–1.18 demonstrates that the point estimate alone does not capture the uncertainty of the treatment comparison.

16. Confidence Intervals and Statistical Precision

The four reported confidence intervals provide a useful way to compare statistical precision across endpoints.

EndpointHR95% CIRelationship to 1
DFS1.08.93–1.26Includes 1
Overall Survival.85.68–1.06Includes 1
DRFI.88.71–1.09Includes 1
Breast Cancer-free Interval.98.81–1.18Includes 1

For a hazard ratio, 1 represents equal estimated hazard between the two groups. A confidence interval that includes 1 therefore includes the possibility of no relative hazard difference under the stated confidence framework. The width of each interval also matters: a wider interval indicates less precision than a narrower interval, all else being equal.

17. Multiplicity and the Endpoint Family

The ClinicalTrials.gov record identifies one primary endpoint and three secondary endpoints. They do not provide an alpha-allocation scheme, multiplicity-adjustment procedure, hierarchical testing sequence, or other formal familywise-error strategy for the four posted analyses.

Primary endpoint

DFS was the registered primary endpoint and therefore occupies the central confirmatory position in the ClinicalTrials.gov record.

Secondary endpoints

Overall survival, DRFI, and breast cancer-free interval were registered and posted as secondary outcomes.

What is reported

Each posted analysis has its own HR, two-sided 95% CI, and P-value.

What is not reported

The ClinicalTrials.gov record does not specify a multiplicity-adjustment procedure across the four endpoints.

This distinction is important because multiple endpoint tests create a broader statistical interpretation problem than looking at one P-value in isolation. Without an explicitly reported multiplicity strategy in the ClinicalTrials.gov record, the individual nominal P-values should be interpreted as the posted analyses rather than automatically treated as a complete familywise-error-controlled testing sequence.

18. Design Topics Not Reported in the Supplied Data

The registry-derived trial data identify the randomized parallel design, intention-to-treat population, log-rank method, hazard ratio, two-sided 95% confidence intervals, and superiority hypothesis. They do not provide sufficient information to support additional claims about several other design features.

Design topicWhat the ClinicalTrials.gov record establishes
Non-inferiority marginNot reported. The hypothesis type is superiority.
CrossoverNot reported.
Factorial designNot reported; the design model is parallel.
Interim analysisNot reported.
Missing-data / imputation strategyNot reported.
Stratification factorsNot reported in the ClinicalTrials.gov record.
Bayesian methodsNot reported.
Multiplicity-adjustment procedureNot reported.
Interpretation rule: absence of a reported method in the ClinicalTrials.gov record should not be treated as evidence that the method was absent from the full protocol or statistical analysis plan. This page therefore limits itself to methods actually supported by the ClinicalTrials.gov record.

19. Statistical Interpretation vs Clinical Interpretation

Statistical interpretation

The primary DFS analysis produced an HR of 1.08 with a two-sided 95% CI of .93–1.26 and P = .31. The three secondary time-to-event analyses produced HRs of .85, .88, and .98, with confidence intervals that included 1.

Clinical interpretation

The ClinicalTrials.gov record permits description of the randomized treatment strategies and their reported time-to-event estimates. They do not, by themselves, justify a broader treatment recommendation or an equivalence conclusion.

The distinction is deliberate. Statistical evidence describes the observed randomized comparison and its uncertainty. Clinical interpretation additionally requires consideration of the endpoint definition, treatment context, adverse events, patient priorities, and the complete evidence base. The ClinicalTrials.gov record is sufficient for the statistical description above but do not supply every component needed for a broader clinical assessment.

20. Limitations

21. Why This Trial Matters Statistically

SOLE is a useful teaching example because the central statistical question is not simply whether an event occurred. It is whether two randomized treatment strategies differ in the timing of clinically defined events, and how that difference should be quantified and tested.

ConceptHow it appears in SOLE
RandomizationParticipants were randomized between continuous and intermittent letrozole strategies.
Parallel designThe registry identifies the design model as parallel.
Intention-to-treatThe primary and secondary efficacy analyses were conducted in the ITT population.
Time-to-event analysisDFS, overall survival, DRFI, and breast cancer-free interval are time-to-event endpoints.
Log-rank testThe registry reports log-rank testing for all four posted analyses.
Hazard ratioEach posted analysis uses an HR as the effect measure.
Confidence intervalEach analysis reports a two-sided 95% CI.
P-valueEach analysis includes a reported P-value for the superiority hypothesis.
Endpoint hierarchyDFS is primary; overall survival, DRFI, and breast cancer-free interval are secondary.
Open-label designThe registry reports no masking.
Safety analysisSerious adverse events are reported by treatment arm with affected and at-risk counts.

22. A Practical Reading of the SOLE Results

A disciplined reading starts with the endpoint rather than the P-value. The primary endpoint is DFS, and the reported comparison gives an HR of 1.08. The next step is to examine the uncertainty: the 95% CI extends from .93 to 1.26. Only after considering the effect estimate and its precision should the P-value of .31 be incorporated as the formal hypothesis-test result.

The same sequence applies to the secondary outcomes. Overall survival has an HR of .85 with a 95% CI of .68–1.06 and P = .16. DRFI has an HR of .88 with a 95% CI of .71–1.09 and P = .25. Breast cancer-free interval has an HR of .98 with a 95% CI of .81–1.18 and P = .84.

Reading sequence
Endpoint → Analysis population → Effect estimate → Confidence interval → P-value → Design context

This sequence helps prevent a common error in trial interpretation: treating a P-value as if it were a complete description of treatment effect.

23. Trial Timeline

2007-08

Trial start

The SOLE trial began in August 2007 according to the ClinicalTrials.gov record.

2018-04

Primary completion

The ClinicalTrials.gov record identifies April 2018 as the primary completion date.

Completed

Registry status

The trial status is reported as completed, with results posted for four outcome measures and four statistical analyses.

24. Sources

Continue through the Clinical Biostats statistical pathway

Explore the methods that make randomized time-to-event trials interpretable, from Kaplan-Meier estimation and hazard ratios to log-rank testing, confidence intervals, and intention-to-treat analysis.

25. Related Tutorials

Learn more about the methods used in this trial:

26. Related Calculators

27. Record Summary

SOLE provides a clear example of randomized time-to-event analysis in which the primary endpoint is disease-free survival and the statistical comparison is based on an intention-to-treat population and log-rank testing. The reported primary HR of 1.08, 95% CI .93–1.26, and P-value .31 should be read together rather than interpreted independently. The three secondary analyses—overall survival, DRFI, and breast cancer-free interval—likewise report hazard ratios with two-sided 95% confidence intervals and P-values.

The most useful statistical lesson is that a clinical-trial result is more than a P-value. The endpoint definition establishes what is being measured; randomization establishes the treatment comparison; the ITT population preserves that randomized comparison; the log-rank test addresses the survival distributions; the hazard ratio quantifies relative event hazard; and the confidence interval communicates precision. Safety information then needs to be considered as a separate evidence stream rather than folded into the efficacy estimate.

Clinical Biostats methodology: This page separates the numerical results reported in the ClinicalTrials.gov record from educational statistical interpretation. Where the ClinicalTrials.gov record does not report a design feature, subgroup analysis, adjustment method, or additional result, the page does not infer one.