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
SURTIME is a phase 3 randomized trial concerning the timing of surgery relative to sunitinib malate treatment in patients with metastatic kidney cancer. The registry records 99 participants, 2 arms, and overall progression-free survival as the primary endpoint.
| Feature | SURTIME |
|---|---|
| Trial acronym | SURTIME |
| ClinicalTrials.gov identifier | NCT01099423 |
| Phase | Phase 3 |
| Status | UNKNOWN |
| Condition | Kidney Cancer |
| Enrollment | 99 |
| Allocation | Randomized |
| Design description | The record describes the design model as single-group while also recording randomized allocation and 2 arms. |
| Masking | None |
| Primary purpose | Treatment |
| Lead sponsor | European Organisation for Research and Treatment of Cancer - EORTC |
| Sponsor type | Network |
| Start date | 2010-04 |
| Primary completion date | 2016-04 |
2. Clinical Question
The clinical question concerns whether the timing of surgery relative to sunitinib malate should be evaluated as an experimental treatment strategy in patients with metastatic kidney cancer.
Population
Patients with kidney cancer in the metastatic setting.
Intervention concept
Timing of surgery in relation to treatment with sunitinib malate.
Comparator concept
Immediate surgery versus surgery after sunitinib malate.
Primary question
How does the timing of surgery relate to overall progression-free survival?
The registry also lists gene expression analysis, biologic sample preservation, laboratory biomarker analysis, and therapeutic conventional surgery among the interventions. These entries indicate that the study incorporated biological and laboratory components in addition to the treatment-related procedures, but the registry information presented here does not provide statistical results for those components.
3. Trial Design
What randomization contributes statistically
Randomization is important because it provides the framework for comparing outcomes between treatment strategies without assigning treatment according to investigators' or patients' prognostic characteristics. In a randomized comparison, the treatment assignment itself is the factor that defines the primary comparison, while baseline characteristics and subsequent events contribute to the uncertainty surrounding the estimated treatment effect.
The absence of masking is also relevant. When participants and investigators know treatment assignment, knowledge of the assigned strategy can potentially influence aspects of treatment delivery, assessment, or subsequent care. For an objective time-to-event endpoint, outcome definitions and censoring rules remain important components of the statistical design.
4. Endpoints
| Endpoint | Registry wording | Time frame |
|---|---|---|
| Primary endpoint | Overall progression-free survival | The ClinicalTrials.gov record does not specify a time frame for this endpoint. |
The primary endpoint is therefore a time-to-event outcome. The term overall progression-free survival indicates that progression-free survival is being considered as an overall study outcome rather than as a subgroup-specific endpoint. The ClinicalTrials.gov record does not give a more detailed definition of progression, the censoring rule, an assessment schedule, or a prespecified landmark time.
5. Planned Analysis
No statistical analyses are posted to ClinicalTrials.gov for SURTIME. The registry does, however, identify overall progression-free survival as the primary endpoint.
How an endpoint of this type is typically analyzed
Progression-free survival is ordinarily analyzed as a time-to-event endpoint. Each participant contributes a time from a defined starting point until progression or another prespecified event, or until the participant is censored according to the study's rules.
Here, S(t) represents the probability of remaining event-free beyond time t. The exact definition of T, including the starting point, event definition, and censoring convention, must come from the trial's prespecified statistical documentation.
A standard analysis of a randomized progression-free survival endpoint commonly uses Kaplan-Meier estimation to describe the event-time distribution. A log-rank comparison can be used to compare the randomized groups, and a Cox proportional-hazards model can provide a hazard ratio with a confidence interval. These are general statistical approaches for time-to-event endpoints; the ClinicalTrials.gov record does not post a formal statistical analysis specifying which of them was used for SURTIME.
The central statistical question is whether the distribution of progression-free survival differs between the randomized strategies. A complete analysis would require the protocol-defined event and censoring rules as well as the observed participant-level or appropriately summarized time-to-event information.
6. Statistical Methodology
Time-to-event analysis
Progression-free survival differs from a simple binary endpoint because both whether an event occurs and when it occurs contribute information. A participant who progresses early and a participant who remains progression-free for a long period are not equivalent observations.
This structure makes survival-analysis methods useful. They allow participants who have not yet experienced the endpoint at the time of analysis to contribute information through their last known event-free time rather than being treated as if their eventual outcome were known.
Kaplan-Meier estimation
The Kaplan-Meier estimator is a nonparametric way to estimate the probability of remaining event-free over time. At each observed event time, the estimate is updated according to the number of events and the number of participants still at risk.
where di is the number of events at time ti and ni is the number at risk immediately before that event time.
For SURTIME, a Kaplan-Meier curve would provide a visual description of how overall progression-free survival evolves in the randomized groups. It would also make the timing of events and censoring more visible than a single summary statistic.
Hazard ratios
If a Cox proportional-hazards model were used, the hazard ratio would summarize the relative instantaneous event rate between the treatment strategies under the fitted model. A hazard ratio below 1 would correspond to a lower estimated instantaneous event rate in the numerator group, while a hazard ratio above 1 would correspond to a higher estimated instantaneous event rate.
The hazard ratio is a relative time-to-event measure. It is not a probability, an absolute risk difference, or a statement that a fixed percentage of individual patients will experience a particular outcome.
Log-rank comparison
A log-rank test is commonly used to compare survival distributions between randomized groups. It evaluates the observed pattern of events over follow-up rather than comparing only one prespecified time point.
For a randomized time-to-event trial, this is useful because participants can have different lengths of follow-up. The analysis can incorporate the available event-time information while accounting for censoring under the prespecified rules.
7. Statistical Methods Explained
Why is progression-free survival a time-to-event endpoint?
Because progression-free survival incorporates the amount of time a participant remains free of progression or the relevant event. A binary endpoint assessed at one date would discard information about whether an event occurred much earlier or much later.
Why are censored observations important?
Some participants may not have experienced the event by the time follow-up ends or may leave observation before an event is recorded. Survival methods allow these participants to contribute the information available up to their censoring time. Censoring is therefore not the same thing as an observed progression-free outcome.
What would a Kaplan-Meier curve show?
It would estimate the proportion of participants remaining progression-free as follow-up progresses. The curve can show when events accumulate, how quickly the estimated event-free probability changes, and how much of the follow-up is based on observed events versus censoring.
What would a hazard ratio mean?
A hazard ratio would compare the estimated instantaneous event rates between the two randomized strategies under a fitted time-to-event model. For example, an HR of 0.70 would describe a 30% lower estimated hazard because 1 − 0.70 = 0.30. It would not mean that 30% of participants avoid progression or that each participant experiences a 30% reduction in personal risk.
Why is the confidence interval important?
A confidence interval communicates statistical uncertainty around an estimated effect. A narrow interval indicates greater precision than a wide interval, all else equal. The interval is not a range containing the individual treatment effect experienced by every participant.
Why should the p-value not be interpreted as the size of an effect?
A p-value addresses compatibility of the observed data with a specified null hypothesis under the statistical model. It does not quantify the magnitude or clinical importance of an effect. Effect estimates, confidence intervals, absolute event-time measures, and the underlying study design provide the information needed to understand magnitude and precision.
8. Interpreting a Potential Survival Analysis
If an eventual analysis reports a hazard ratio, the estimate should be interpreted as a relative time-to-event measure. It describes the relationship between the event rates in the randomized groups under the fitted model, rather than directly describing the percentage of patients who experience progression.
Time-specific progression-free survival probabilities and median progression-free survival, if reported, would provide complementary absolute measures. These can make the practical magnitude of a time-to-event difference easier to understand than a hazard ratio alone.
The confidence interval around a hazard ratio would indicate how precisely the treatment comparison has been estimated. With 99 participants enrolled, the amount of information available for the primary endpoint depends not only on enrollment but also on the number and timing of progression-free survival events and the amount of censoring.
If a Cox proportional-hazards model were used, interpretation of a single hazard ratio would depend on the proportional-hazards assumption. If the relative hazards change substantially over time, one summary hazard ratio may not describe the treatment difference equally well throughout follow-up.
9. What Enrollment Tells Us About Statistical Precision
The registry records an enrollment of 99 participants. Enrollment is an important design quantity, but for a time-to-event endpoint it is not the only determinant of statistical information.
Participants
99 participants are recorded as enrolled. A larger randomized sample generally provides more information, but the actual precision of a survival comparison also depends on observed events.
Events
For progression-free survival, the number and timing of events are central to the information available for estimating a treatment effect.
Censoring
Censored observations contribute follow-up information but do not provide the same information as observed endpoint events.
Follow-up
The registry records a primary completion date of 2016-04, but the ClinicalTrials.gov record does not provide a statistical follow-up summary for progression-free survival.
This distinction is especially important when interpreting a study with a relatively modest enrollment. A sample of 99 participants does not automatically imply a particular confidence-interval width or statistical power for a time-to-event comparison. Those quantities depend on the prespecified design assumptions and the information actually accumulated through follow-up.
10. Design Features That Affect Interpretation
Randomization
Randomized allocation is the central design feature supporting a causal comparison between treatment strategies. If randomization is implemented as intended, prognostic factors are balanced in expectation across groups rather than being assigned according to treatment preference.
No masking
The registry records no masking. Knowledge of treatment assignment can matter more for outcomes that require judgment or influence subsequent management than for outcomes defined by objective events. Nevertheless, assessment procedures remain part of the statistical validity of a progression-free survival endpoint.
Two-arm comparison
The registry records 2 arms. A two-arm randomized comparison naturally lends itself to estimating the relative difference between the two assigned treatment strategies, provided that the endpoint and analysis population are prespecified appropriately.
Single-group design description
The registry also describes the design model as single-group. Because that description does not align naturally with the separately recorded randomized allocation and 2 arms, it should be treated as a registry-design descriptor rather than used to infer additional treatment assignments or comparisons.
11. Biological and Laboratory Components
In addition to the treatment-related procedures, the registry lists several biological and laboratory interventions:
| Intervention category | Registry description |
|---|---|
| Treatment-related | Timing of surgery; therapeutic conventional surgery |
| Genetic | Gene expression analysis |
| Biological sample | Biologic sample preservation procedure |
| Laboratory | Laboratory biomarker analysis |
These components broaden the scientific scope of the study beyond the primary treatment comparison. Statistically, however, a biological or biomarker measurement should not automatically be treated as a confirmatory clinical endpoint. Its interpretation depends on the prespecified objective, measurement procedure, analysis population, and multiplicity structure.
The registry information available here does not post statistical analyses for these components. They therefore do not provide an additional numerical treatment effect to interpret alongside overall progression-free survival.
12. Missing Statistical Results
The ClinicalTrials.gov record identifies the primary endpoint but contains no posted statistical analyses. There are therefore no registry-posted estimates, confidence intervals, or p-values for overall progression-free survival to interpret on this page.
This distinction is fundamental in clinical-trial statistics. A trial registration establishes the planned or registered scientific framework, while an inferential result requires observed outcome data and a defined analysis. The two should not be conflated.
13. Limitations
- No posted statistical analyses: The registry contains no formal statistical analyses for the primary endpoint.
- Incomplete endpoint detail: The ClinicalTrials.gov record names overall progression-free survival but does not specify the detailed event definition, starting point, assessment schedule, or censoring rule.
- No numerical efficacy estimates: No hazard ratio, confidence interval, median progression-free survival, or time-specific progression-free survival estimate is posted in the ClinicalTrials.gov record.
- Design-description inconsistency: The record describes randomized allocation and 2 arms while also describing the design model as single-group.
- No treatment-arm details: The ClinicalTrials.gov record identifies the intervention categories but does not assign the listed interventions to specific randomized arms in enough detail to construct a complete arm-by-arm treatment schedule.
- No baseline characteristics: The ClinicalTrials.gov record does not provide a baseline demographic or disease-characteristic table.
- No safety results: Serious adverse-event counts by arm are not posted in the ClinicalTrials.gov record, so a comparative safety analysis cannot be constructed from it.
- No subgroup results: No subgroup estimates or interaction analyses are posted for the primary endpoint.
- No multiplicity framework: The ClinicalTrials.gov record does not report an alpha allocation, multiplicity adjustment, or hierarchy among multiple endpoints.
- No interim-analysis information: The ClinicalTrials.gov record does not describe interim looks, stopping boundaries, or alpha-spending procedures.
- No missing-data or imputation strategy: The ClinicalTrials.gov record does not describe how missing endpoint information would be handled.
- No Bayesian analysis: The ClinicalTrials.gov record does not report a Bayesian statistical framework.
14. Why This Trial Matters Statistically
SURTIME is a useful teaching case because its central scientific question is about when an intervention should occur, rather than simply whether an intervention should occur. That creates a natural randomized comparison of treatment strategies in which the timing of surgery is part of the intervention itself.
| Statistical concept | How it applies to SURTIME |
|---|---|
| Randomization | The registry records randomized allocation to 2 arms. |
| Time-to-event analysis | Overall progression-free survival is the registered primary endpoint. |
| Censoring | A progression-free survival analysis would ordinarily need rules for participants whose endpoint status is not observed through the end of follow-up. |
| Kaplan-Meier estimation | A standard method for describing progression-free survival over time. |
| Hazard ratio | A common relative measure for comparing time-to-event rates between randomized groups. |
| Confidence interval | Provides information about uncertainty around an estimated treatment effect. |
| Statistical power | Depends on the prespecified effect, event information, significance level, allocation, and follow-up rather than enrollment alone. |
| Endpoint definition | The exact event and censoring rules are essential for interpreting progression-free survival. |
| Design consistency | The record contains randomized and single-group design descriptions that should be distinguished rather than silently reconciled. |
15. Clinical Interpretation vs Statistical Interpretation
Clinical question
The trial addresses whether immediate surgery and surgery after sunitinib malate represent different treatment-timing strategies for patients with metastatic kidney cancer.
Statistical question
The registered primary endpoint asks whether the randomized treatment strategies differ in overall progression-free survival.
What the endpoint can measure
Progression-free survival can capture both the occurrence of progression and the time until that event, making it sensitive to differences in the temporal pattern of disease control.
What is not established by the registry
The registry does not provide an estimated treatment effect, its precision, or a hypothesis-test result for the primary endpoint.
A useful statistical interpretation therefore separates the question posed by the trial from the answer produced by an analysis. The registry clearly identifies the former. It does not post the latter in the record summarized here.
16. Understanding Progression-Free Survival in This Setting
Progression-free survival is especially relevant when the timing of treatment may alter how quickly disease progression occurs. Unlike overall survival, progression-free survival focuses on the interval before progression or another prespecified event, rather than death alone.
Two treatment groups could have the same proportion of participants who eventually progress but very different timing of those events. Conversely, similar median progression-free survival values could coexist with different shapes of the survival curves. This is why the full time-to-event distribution can be more informative than a single percentage.
If a median progression-free survival were available, it would identify the time at which the estimated event-free survival crosses 50%. It would not describe the entire curve, the relative hazard throughout follow-up, or the absolute difference at every clinically relevant time point.
A hazard ratio compares event rates under a model, whereas a Kaplan-Meier estimate describes an event-free probability at a particular time. These quantities are related but not interchangeable.
17. A Practical Statistical Reading of the Trial
For readers evaluating SURTIME, the most useful sequence is to begin with the randomized treatment strategies, identify the exact primary endpoint, establish how progression-free survival was defined and censored, and then examine the treatment-effect estimate with its confidence interval.
Identify the randomized comparison
The record describes randomized allocation and 2 arms, with the clinical question centered on immediate surgery versus surgery after sunitinib malate.
Define the endpoint
The registered primary endpoint is overall progression-free survival. The detailed event and censoring definition is not reported in the ClinicalTrials.gov record summarized here.
Choose the time-to-event framework
Kaplan-Meier estimation, log-rank comparison, and Cox modeling are standard approaches for this class of endpoint, although the registry does not post a formal analysis plan identifying which were used.
Quantify effect and precision
An appropriate analysis would normally report an effect estimate together with a confidence interval, allowing both magnitude and statistical uncertainty to be assessed.
Interpret in context
The treatment effect should be considered alongside the endpoint definition, censoring, follow-up, analysis population, and the limitations of the trial design.
18. Sources
- ClinicalTrials.gov: NCT01099423, the official registry record for SURTIME.
- PubMed: PubMed record 30543350.
- PubMed: PubMed record 24338498.
- PubMed: PubMed record 23228299.
Continue through the Clinical Biostats trial-analysis library
Clinical trial results are most useful when the clinical question, endpoint definition, study design, and statistical interpretation are considered together.
19. Record Summary
SURTIME is a phase 3 randomized trial with 99 participants and 2 arms, addressing the timing of surgery in relation to sunitinib malate for patients with metastatic kidney cancer. Its registered primary endpoint is overall progression-free survival, making it fundamentally a time-to-event statistical question. The registry records randomized allocation, no masking, a treatment-focused primary purpose, and a set of treatment, surgical, genetic, biological-sample, and laboratory interventions.
The principal statistical lesson is that a registered endpoint and an analyzed endpoint are not the same thing. A progression-free survival endpoint requires a precisely defined event, starting time, censoring strategy, follow-up structure, and analysis population before an effect estimate can be interpreted. Standard survival methods such as Kaplan-Meier estimation, log-rank comparison, and Cox regression provide a framework for that analysis, but the registry record does not post a formal statistical analysis or numerical primary-endpoint result.