Sunitinib to Treat Recurrent Brain Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Sunitinib to Treat Recurrent Brain Cancer (NCT00923117), a randomized phase 2 study. The primary outcome is 6-month Progression-free Survival..

2. Clinical Question

3. Trial Design

This was a phase II, randomized, open-label, parallel trial sponsored by the National Cancer Institute (NCI). The registry lists total enrollment as 87 and status as terminated. The record gives a start date of 2008-06 and a primary completion date of 2012-06, but does not report why the trial was terminated.

The intervention was sunitinib, described in the brief summary as a pill taken once a day in 4-week cycles. The summary describes recurring MRI scans to monitor tumor growth, along with blood tests, neurological and physical examinations, and health-related quality-of-life assessments. PET scans were planned at baseline and at the end of the first cycle. These are descriptions of the trial procedures, not evidence that every assessment was completed for every participant.

The primary outcome population was narrower than total enrollment: 63 patients with GBM were evaluated, split into bevacizumab-naive and bevacizumab-resistant groups. The group labels describe prior treatment history, not different study drugs. Both groups are listed under the sunitinib intervention. Although the registry classifies allocation as randomized, it does not explain what was randomized, how assignment was performed, or whether randomization related to a treatment contrast. No separate control treatment is reported.

The registry lists one secondary outcome, participants with adverse events. It does not report the protocol's statistical analysis plan, a sample-size justification, details of interim analyses, or further information about the reason for termination. These omissions limit how fully the design and intended analysis can be reconstructed from the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Analysis population for the primary outcome, as defined in the registry: 63 patients enrolled with glioblastoma multiforme (GBM) (32 Bev naive, 31 Bev resistant) and were evaluated for this outcome measure.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as 6-month Progression-free Survival., measured over 6 months.

The primary endpoint is progression-free survival, a time-to-event outcome. The description defines the interval from treatment start to progression, with progression determined by RANO criteria. Those criteria include an increase in the measured enhancing lesions, a new lesion, or clinical deterioration. The endpoint therefore combines radiographic change and clinical status rather than measuring tumor size alone.

The registry's title for the outcome refers to 6-month progression-free survival, but the posted parameter is the median time to progression in months, with a 95% confidence interval, for each GBM group. The reported median is the point by which half of the observed time-to-event distribution has reached the event, subject to the study's handling of follow-up and censoring. That handling is not described in the posted results.

General statistical point: a time-to-event endpoint can use information about when an event occurs, rather than treating all participants as if they had identical follow-up. It is commonly used when the timing of progression is important and some participants may not have progression recorded during observation. The registry record does not state how many observations were censored or how censoring was handled, so those details cannot be assessed from the posted summary alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a median time-to-event result mean?

General statistical point: for a time-to-event outcome, each participant contributes information about the time until an event or the end of observable follow-up. The median is a summary of the event-time distribution, not an average and not a count of events. In this trial, the event was progression as defined in the registry description. The two posted medians describe the bevacizumab-history groups separately.

The outcome title refers to 6-month progression-free survival, but the posted statistic is median time to progression, expressed in months. These are different ways to describe a time-to-event outcome. A survival probability at a particular time describes the estimated proportion that has not yet experienced the event by then; a median describes the time at which the estimated event-time distribution reaches its midpoint. The registry record does not post a distinct 6-month probability estimate.

How should the confidence intervals be read?

General statistical point: a confidence interval communicates uncertainty in an estimate under the assumptions of the analysis. A narrower interval indicates greater precision in the reported estimate than a wider interval, but it does not establish that the analysis is unbiased or that a difference between groups is meaningful. Here, the posted interval for the bevacizumab-naive group extends from 0.92 to 2.47 months, while the interval for the resistant group extends from 0.89 to 0.92 months. The registry does not explain the methods used to estimate these intervals.

The interval for an individual group estimate is not a formal test of the difference between the two groups. A between-group test or model would need to be reported to support a statistical comparison. No such result is posted. The registry also reports no hazard ratio, which is a comparative measure often used for time-to-event outcomes. No comparative effect should be inferred by calculating from the posted medians or intervals.

What does randomization establish in this record?

General statistical point: random assignment can reduce systematic differences between groups when participants are assigned to distinct study conditions according to a defined process. It does not by itself clarify what the comparison is, and it does not guarantee that all analyses are unbiased. The registry classifies allocation as randomized and the study as parallel, but the posted arm descriptions distinguish groups by prior bevacizumab exposure and list sunitinib for both. The assignment procedure and the intended randomized contrast are not reported in the registry record. The randomized label alone therefore cannot be used to treat the two outcome groups as randomized sunitinib-versus-control arms.

What is known about the analysis?

The primary outcome is presented as a median with a 95% confidence interval for each group. The registry does not post a hazard ratio, p-value, adjusted model, or formal test of group differences. It does not report the statistical analysis plan, the method for handling censoring, or the number of progression events. Those omissions matter because time-to-event results depend on follow-up and on how participants without a recorded event are incorporated into the analysis.

General statistical point: a hazard ratio compares event rates over follow-up under a model and is not interchangeable with a ratio of medians. Interpretation of a hazard ratio also depends on the analysis model and its assumptions. Since no hazard ratio is reported, there is no sponsor-reported comparative estimate to interpret. No recalculated comparison is appropriate from the posted medians alone.

The registry reports an outcome population of 63 patients with GBM, including 32 bevacizumab-naive and 31 bevacizumab-resistant patients. It separately reports overall enrollment of 87. No explanation is posted for why the primary outcome population is smaller than total enrollment or for whether the difference reflects eligibility, available outcome data, or another reason. These distinctions cannot be resolved from the record.

Why does over-dispersion matter in general?

General statistical point: over-dispersion occurs when observed variability exceeds what a statistical model assumes. If a model ignores over-dispersion, its uncertainty estimates may be too small and tests may appear more decisive than they should. This is a general issue in some count-data analyses; the posted primary result is a time-to-event median, and the registry does not report an analysis involving a count model or an over-dispersion adjustment. It is therefore not possible to assess whether over-dispersion was relevant to the sponsor's analysis.

The record does not describe planned interim analyses, multiplicity adjustments, or a sample-size justification. There is also no basis in the posted results for reconstructing those choices. The reported estimates should be interpreted as the results available in the registry, without attributing unreported methods or assumptions to the trial team.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists the primary progression-free survival outcome and one secondary outcome: the number of participants with adverse events. It does not report whether the study tested additional outcomes, how multiple comparisons were handled, or whether any multiplicity adjustment was planned. The statistical analysis plan and any interim-analysis plan are not reported in the registry record. Without those details, it is not possible to determine how the posted outcomes relate to the trial's full prespecified testing strategy.

The primary outcome is reported for 63 patients with GBM, while total enrollment is 87. The record does not explain the difference or specify whether the outcome population was defined in advance. It does not state how missing outcome information, incomplete follow-up, or censoring was handled. These details are important for interpreting a time-to-event result because participants may have different amounts of observable follow-up.

The adverse-event figures use different denominators from the primary outcome groups: 44 patients at risk in the resistant group and 43 in the naive group. The registry does not explain how those safety populations relate to the outcome population. It also does not describe the period during which adverse events were collected, the event-attribution rules, or whether participants could contribute more than one event of a type. The figures can be reported as posted, but those missing definitions limit interpretation.

Because the analysis population definitions, missing-data handling, and multiplicity approach are not reported, the posted estimates do not provide a complete account of the trial's statistical plan. No conclusion about selective outcome reporting or the direction of any resulting bias can be made from the registry record alone.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: thrombosis/thrombus/embolism (9.3%), death not associated with ctcae term: disease progression nos (9.1%), seizure (6.8%), vomiting (6.8%), infection with normal anc or grade 1 or 2 neutrophils::lung (pneumonia) (4.7%).

The adverse-event results list 44 patients at risk in the bevacizumab-resistant group, of whom 16 were affected by serious adverse events. In the bevacizumab-naive group, 43 patients were at risk and 13 were affected. The record also reports 44 of 44 and 43 of 43, respectively, for the other adverse-event measure. These are participant counts, not rates per unit of follow-up, and the registry does not specify the observation period or detailed event-collection rules.

The registry lists serious-event terms with percentages, including thrombosis/thrombus/embolism at 9.3%, death not associated with a CTCAE term: Disease progression NOS at 9.1%, seizure at 6.8%, vomiting at 6.8%, infection with normal ANC or Grade 1 or 2 neutrophils::Lung (pneumonia) at 4.7%, and pain: Head/Headache at 4.7%. The denominator and counting convention for these term-specific percentages are not reported in the registry record.

General statistical point: a serious adverse event is a regulatory classification and is not identical to an event's severity grade. Counts also do not, by themselves, establish that an event was caused by the study drug. The posted figures describe the safety record; attribution details and a complete account of event timing are not reported.

12. What the Result Does, and Does Not, Mean

The registry results describe median time to progression for two groups of patients with GBM receiving sunitinib, distinguished by whether they were bevacizumab-naive or bevacizumab-resistant. The posted median was 0.92 months for the resistant group and 1.08 months for the naive group, with the respective 95% confidence intervals reported in the primary results. The results also give participant counts for serious and other adverse events.

These data do not establish that sunitinib caused a change in progression-free survival. The registry does not post a comparison with a non-sunitinib control treatment, a hazard ratio, or a formal test contrasting the two history groups. Prior bevacizumab status is not itself a randomized treatment assignment as described in the arm definitions. A direct treatment-effect conclusion would therefore go beyond the information reported.

The figures also do not establish how the results would apply to every enrolled participant or to all malignant gliomas. The primary outcome population was 63 patients with GBM; total enrollment was 87. The registry does not explain the difference, and the results are not reported separately for every condition listed in the trial record.

General statistical point: an estimate is best interpreted together with its uncertainty, population definition, and analysis method. The confidence intervals give a measure of precision for the medians, but the record does not provide important supporting details such as censoring, follow-up, or the analysis plan. The adverse-event counts describe reported participants and do not, on their own, establish causation or comparative safety.

The appropriate reading is therefore descriptive and limited to the posted registry results. No conclusion about treatment benefit, lack of benefit, or a causal difference between the groups follows from these summaries alone.

13. Why This Trial Matters Statistically

This trial illustrates why the label “randomized” is not enough to interpret a clinical comparison. The registry classifies the study as randomized, yet the posted outcome groups are defined by prior bevacizumab exposure and both receive sunitinib. Without a description of what was randomized, the labels do not establish a randomized contrast between treatments. Identifying the actual comparison is essential before interpreting group differences causally.

The record also highlights the distinction between an outcome's name and the statistic reported for it. The primary outcome title refers to 6-month progression-free survival, but the posted result is a median time-to-event estimate. A median and a probability at a specified time answer different questions. Clear interpretation requires attending to the measure actually reported, not relying on the title alone.

Uncertainty and population definitions matter as well. The registry gives confidence intervals for each median, but no comparative test, hazard ratio, or analysis details. Total enrollment and the GBM outcome population are not the same, and the reason is not reported. Safety counts use their own denominators. Each of these distinctions affects how far a reader can reasonably generalize or compare the results.

Statistically, the record is a useful example of the limits of registry summaries: descriptive estimates can be reported precisely as posted while causal claims remain unsupported. Missing analysis details should be acknowledged rather than filled in with assumptions about the protocol or study methods.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary result is available only for the 63 patients with GBM evaluated for that outcome. Although total enrollment was 87 and the trial included several malignant glioma conditions, the record does not explain why the primary population is narrower or provide the same outcome results for all enrolled participants. Generalization beyond the reported GBM population is therefore uncertain.

The treatment comparison is also unclear. Both outcome groups received sunitinib, and their definitions concern prior bevacizumab exposure. While allocation is listed as randomized, the randomization procedure and intended contrast are not described. The group estimates cannot be interpreted as a randomized comparison of sunitinib against another treatment.

The registry does not report a hazard ratio, p-value, between-group test, adjusted analysis, number of progression events, censoring details, or follow-up distribution. It also does not provide the statistical analysis plan, sample-size justification, or interim-analysis information. Without these details, the precision and analytical assumptions underlying the posted estimates cannot be fully evaluated. The confidence intervals are reported, but their estimation method is not described.

For safety, the at-risk denominators differ from the primary outcome group sizes, and the record does not explain the relationship between these populations. The event observation period, attribution rules, and term-specific percentage denominators are not reported. The study status is terminated, but the reason for termination is not given. These limitations call for cautious, descriptive interpretation rather than assumptions about missing information.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Sunitinib to Treat Recurrent Brain Cancer — NCT00923117.
• Registry record last updated 2015-09-30; status last verified 2014-07.
• Results first posted 2013-06-25.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT00923117 is a phase II, randomized, open-label, parallel trial titled “Sunitinib to Treat Recurrent Brain Cancer.” The official title is “A Phase II Trial of Sunitinib in the Treatment of Recurrent Malignant Gliomas.” The National Cancer Institute (NCI) is listed as sponsor. Enrollment was 87, and the record lists the trial status as terminated.

The primary outcome concerns progression-free survival in 63 patients with GBM, split by bevacizumab history. The posted median time to progression was 0.92 months for bevacizumab-resistant patients and 1.08 months for bevacizumab-naive patients, each with a reported 95% confidence interval. Sunitinib is the listed intervention for both groups. The registry also reports adverse-event participant counts and selected serious-event terms.

The record does not report a separate treatment comparator, a formal group comparison, a hazard ratio, or the statistical analysis plan. The reason for termination, missing-data handling, and explanation for the difference between total enrollment and the primary outcome population are not reported in the registry record.