Treatment of Patients With Metastatic or Unresectable Gastrointestinal Stromal Tumors in First Line With Nilotinib: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Treatment of Patients With Metastatic or Unresectable Gastrointestinal Stromal Tumors in First Line With Nilotinib (NCT00756509), a na phase 4 study. The primary outcome is Proportion of Participants With Best Overall Response at Month 6 (Core Phase) Determined According to the RECIST v1.0..

2. Clinical Question

The study concerns response during the core phase, with the primary assessment timeframe described as baseline to month 6. Its central statistical question is descriptive: how many participants met the specified response-or-stability criteria in this group? Without a concurrent control group, the study cannot estimate a comparative difference between nilotinib and another treatment or between nilotinib and no treatment.

The registry record labels the study Phase 4, while its brief summary calls it a phase II trial. That difference in phase descriptions is not resolved in the record. The stated objective is to evaluate nilotinib in first-line treatment, but the posted results do not give the primary endpoint estimate or its confidence interval. The clinical question can therefore be described from the protocol summary, but the magnitude and precision of the observed primary outcome cannot be assessed from the registry information available.

3. Trial Design

This is an open-label, multicenter, single-arm study of nilotinib in adults with metastatic or unresectable GIST. The registry record lists 34 enrolled participants and marks the study as completed. It gives a start date of 2008-08-29 and a primary completion date of 2024-12-18. The listed phase is PHASE4, although the brief summary describes the study as phase II; the record does not explain the difference.

Because there is one treatment group and no comparator, the design is suited to describing outcomes among participants treated in the study, not to estimating a randomized treatment contrast. Allocation is listed as not applicable, and masking is listed as none. An open-label design means treatment assignment was not masked; the record does not detail whether outcome assessments were independently reviewed or otherwise protected from knowledge of treatment.

The primary efficacy endpoint concerns response status by month 6 in the core phase. The record also lists secondary outcomes covering objective response rate at month 6, time to response at month 6, duration of response, progression-free survival, overall survival, and treatment-emergent adverse events during the core and follow-up phases. The record states that there are 7 outcome measures, but it contains no analysis records. Numerical estimates for the primary efficacy outcome and the listed secondary outcomes are not reported in the registry record.

Safety summaries distinguish a core phase from a follow-up phase. The listed numbers at risk are 41 for the core safety group and 31 for the follow-up safety group. These figures do not match the listed enrollment of 34, and the record does not explain the difference. They should be read as the denominators attached to the corresponding safety summaries, not treated as a clarification of enrollment or as evidence that all records refer to the same analysis population.

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.

Before assignment: All enrolled participants received 400 mg bid dose of nilotinib.

Core Phase

Reasons for not completing this period:

Follow-Up Phase

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The intent-to-treat (ITT) population consisted of all participants who received at least one dose of the study drug and had at least one post-baseline assessment of the primary efficacy variable (assessment according to RECIST). Participants without any post-baseline assessment of tumour were included if they were defined as progressive disease based on clinical evaluation.

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 Proportion of Participants With Best Overall Response at Month 6 (Core Phase) Determined According to the RECIST v1.0., measured over from baseline to month 6, core phase.

The primary endpoint is described as the proportion of participants with a best overall response of complete response (CR), partial response (PR), or stable disease (SD) by month 6, assessed according to RECIST v1.0. The description gives specific rules for classifying response: CR and PR require confirmation on assessments separated by at least four weeks, while SD requires an assessment more than six weeks after treatment begins. Progressive disease (PD) includes progression or cancer-related death within 12 weeks of starting treatment, subject to the stated classification rules. UNK is used for cases that do not meet the listed criteria, including some cases without confirmed response or qualifying stable disease.

Combining CR, PR, and SD into one endpoint captures participants who meet any of those three categories by the specified assessment period. It therefore describes a broader outcome than objective response alone, because SD is included alongside CR and PR. The record’s endpoint title refers to best overall response, while its description identifies the proportion reaching CR, PR, or SD; the description supplies the operational categories used for the primary measure.

For a binary endpoint, the observed proportion would describe the share of the analysis population classified in the selected categories. A confidence interval would communicate the uncertainty around that share. The record specifies a one-sided exact 90% confidence interval, or an 80% two-sided interval, for the ITT F group. It does not report the resulting proportion, interval limits, or participant counts by response category. The endpoint definition is therefore available, but the observed primary result and its precision are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The methods described in the registry record are limited. The endpoint definitions and intended confidence-interval format are available, but numerical efficacy estimates and analysis records are not. The following concepts explain how the stated methods are generally interpreted; they do not imply that unreported procedures were used in this study.

What does a proportion measure?

A binary outcome assigns each participant to a category, such as meeting or not meeting a specified endpoint definition. The proportion is the count meeting the definition relative to the analysis denominator. In this study, the primary description groups CR, PR, and SD as favorable classifications for the endpoint. Interpreting a proportion requires knowing both the classification rule and which participants were included in its denominator. The registry defines the response categories but does not report the primary count, proportion, or analysis denominator.

What does a confidence interval add?

A confidence interval expresses statistical uncertainty around an estimate. A narrower interval indicates greater precision under the method’s assumptions; a wider interval indicates less precision. The registry description says that the proportion of participants with CR, PR, or SD would be presented with a one-sided exact 90% confidence interval, or an 80% two-sided confidence interval, for the ITT F group. These are alternative ways of expressing the intended interval convention. No interval limits are reported, so the precision of the primary estimate cannot be evaluated.

An exact interval is a method for estimating uncertainty around a proportion that does not rely on the same large-sample normal approximation used by some simpler methods. This can be useful when counts are limited or proportions are near the boundaries of their possible range. The record’s mention of an exact interval does not supply the underlying count, denominator, or resulting interval, and therefore cannot substitute for the missing result.

Why does the analysis population matter?

The endpoint description identifies an intent-to-treat population, abbreviated in the record as ITT F. It says this population includes participants who received at least one dose and had at least one post-baseline assessment of the primary efficacy variable. It also says participants without a post-baseline tumor assessment were included. These statements conflict as written: one condition requires a post-baseline assessment, while the other describes inclusion without one. The registry record does not clarify which rule governed the analysis.

Population rules matter because the denominator determines the interpretation of a proportion. Excluding participants without an assessment, counting them as unknown, or assigning them another outcome can lead to different estimates. Those are general possibilities, not a description of what was done in this study. The record does not specify how missing primary assessments were handled, how many participants had missing assessments, or which denominator was used for the primary analysis.

How should a single-group result be understood?

In a single-group study, an estimate describes the observed outcomes in the treated group under the study’s eligibility rules, follow-up, and measurement procedures. It does not provide a direct treatment contrast. A randomized comparison can balance measured and unmeasured factors in expectation through random allocation; this study has no randomized comparator. Consequently, a single-group proportion cannot establish how the same participants would have fared under an alternative treatment. This is a general limitation of the design, not a judgment about the clinical value of the study.

Open-label conduct means that masking was absent. For outcomes involving assessment or categorization, knowledge of treatment can be relevant to how measurements are made or interpreted. The registry record does not say whether response assessments were centrally reviewed, whether assessors were otherwise masked, or what procedures were used to reduce assessment bias. It would be inappropriate to assume either that such safeguards were present or that they were absent beyond the listed lack of masking.

What can be inferred from the safety summaries?

Safety outcomes are commonly presented as counts of participants with events and the denominator observed or considered at risk. A serious adverse event count is not interchangeable with a count of all adverse events: a participant may experience more than one event, and a summary may count participants rather than event episodes. The registry labels the displayed serious and other affected values as participant counts. It does not provide enough detail to infer event duration, recurrence, timing, causal attribution, or the number of event episodes.

The core and follow-up safety groups have different listed numbers at risk. A reader should retain each denominator with its own phase-specific count rather than combine the phases or assume that the groups are independent. The record does not explain overlap between phases, changes in observation time, or the relationship between safety denominators and the enrollment figure. The percentages listed for serious-event terms should be read as recorded; the registry does not provide term counts alongside those percentages in this summary.

What does the absence of analysis records mean?

The registry lists 0 analysis records and does not give numerical results for the primary endpoint. An empty analysis listing does not prove that investigators conducted no analyses. It does mean that model details, estimates, and inferential results cannot be checked from the analysis records. Similarly, a stated confidence-interval plan is not evidence that the interval was calculated or posted in the registry. No multiplicity procedure, interim analysis, or other statistical plan is reported in the registry record.

Planning calculations, when presented, are distinct from observed outcomes. A sample-size illustration based on an observed effect would be a planning exercise, not observed power and not evidence that the original study had a particular probability of detecting an effect. The registry information available does not give numerical planning estimates or a sample-size justification. The design rationale and any formal power calculation are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists 7 outcome measures, including the primary response endpoint and secondary measures of objective response rate, time to response, duration of response, progression-free survival, overall survival, and treatment-emergent adverse events. It does not report a multiplicity strategy for interpreting results across these outcomes. When several outcomes are examined, general statistical practice recognizes that the chance of at least one apparently notable finding can increase if each is assessed separately. Whether an adjustment, hierarchy, or other interpretive plan was used is not reported in the registry record.

The record lists 0 analysis records. It also does not report numerical primary or secondary outcome results in the available information. This prevents assessment of the estimates, uncertainty intervals, analysis methods, and any outcome-specific population sizes. The lack of analysis entries should not be taken as proof that analyses were not conducted; their details are not reported in the registry record.

The primary endpoint population description is unclear. It first requires at least one post-baseline assessment, then says that participants without any post-baseline tumor assessment were included. The record does not resolve that contradiction or state how missing data were handled. It also does not report the number of missing assessments or whether missing status was treated as UNK, excluded from a denominator, or handled another way. The definition of UNK identifies circumstances that may fall into that category but does not by itself explain the analysis rule for missing data.

The registry does not report whether there were interim analyses, how the sample size was justified, or whether any prespecified statistical analysis plan governed the outcomes. These absences limit the ability to evaluate how outcome selection, missingness, and multiple testing were handled.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: abdominal pain (6.5%), pyrexia (6.5%), cardiac failure (6.5%), vertigo (6.5%), myocardial infarction (3.2%).

The registry separates safety summaries into a core phase and a follow-up phase. For the core group, it lists 41 participants at risk, 10 affected by serious events, and 39 affected by other events; deaths are listed as 0. For the follow-up group, it lists 31 participants at risk, 14 affected by serious events, and 28 affected by other events; deaths are listed as 0. These are phase-specific registry counts and should be read with their own denominators.

The listed number at risk for each safety group is not the same as the enrollment of 34. The record does not explain this discrepancy, and it does not specify how the groups relate to the primary efficacy population. The figures should therefore not be used to reconstruct an overall participant count or to assume the same people were observed in both phases.

The serious-event terms listed most often by percentage are abdominal pain, pyrexia, cardiac failure, and vertigo, each recorded as 6.5%. Myocardial infarction and cardiac failure chronic are each recorded as 3.2%. These percentages are reported without term counts in the available summary, and the record does not describe the selection rule for this list. The summaries do not establish causation, event severity beyond the serious-event classification, timing, or duration. They describe what the registry records, not a comparison of safety between treatments.

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

The registry record establishes the study’s intended question and design: an open-label, single-arm evaluation of nilotinib in adults with metastatic or unresectable GIST receiving first-line treatment. It also gives a detailed definition of the month 6 primary endpoint, including criteria for CR, PR, SD, PD, and UNK. These details make clear what the study intended to classify, but they are not numerical evidence about how participants were classified.

The primary efficacy result is not reported in the registry record. No primary response count, proportion, confidence interval, or analysis denominator is available in the information summarized for this study. Therefore, the record does not permit a quantitative judgment about the observed response-or-stability proportion or the precision of that estimate. It would be incorrect to infer a result from the trial objective or from the outcome definitions alone.

Even if a single-group outcome estimate were available, its interpretation would remain descriptive. Without a concurrent control group, the study cannot directly compare outcomes under nilotinib with outcomes under another treatment. The design cannot by itself establish that a difference in outcome was caused by treatment, because it does not provide the randomized counterfactual comparison needed for that claim. This limitation applies generally to uncontrolled studies.

The safety summaries provide recorded counts for serious and other adverse events in the core and follow-up phases, as well as selected serious-event terms. They should be interpreted as descriptive registry summaries. The discrepancy between the listed enrollment and the safety denominators is not explained, and the record does not provide enough information to reconcile those populations. The counts also do not establish whether events were caused by nilotinib.

In short, the record supports a description of the trial question, design, endpoint criteria, and selected safety counts. It does not support a numerical conclusion about primary efficacy, a comparison with an alternative treatment, or a causal conclusion about treatment effects. The distinction between an endpoint definition and its observed result is essential when interpreting this record.

13. Why This Trial Matters Statistically

This study is statistically notable because its primary endpoint combines response and stable disease into a single month 6 classification. That choice asks a broader descriptive question than an endpoint restricted to CR and PR. Understanding the endpoint requires attention to its operational definitions, confirmation rules, time window, and handling of cases classified as PD or UNK. The registry record documents these definitions, allowing readers to see what the intended outcome represents.

The study also illustrates the limits of interpreting a single-arm design. A proportion observed in one treated group can describe outcomes in that group, but without a comparator it cannot provide a treatment contrast. This distinction matters in clinical research because an outcome count is not, on its own, an estimate of what would have happened under another treatment. Design governs which statistical questions the evidence can answer.

Registry completeness is also important. Here, the intended confidence-interval approach is stated, but the primary estimate and its interval are not reported in the available record. The analysis population description contains conflicting statements about participants without post-baseline assessments, and the safety denominators do not match the enrollment figure. Those issues make it harder to connect outcome definitions, denominators, and estimates.

The record further distinguishes core and follow-up safety summaries and lists adverse-event counts and selected terms. These are useful descriptive elements, but interpreting them requires clarity about the populations at risk, observation periods, and whether the same participants contribute to both phases. Those details are not reported. The study therefore provides a useful example of how statistical interpretation depends not only on methods and definitions, but also on complete reporting of results and populations.

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.

Several limitations are specific to the information in this registry record. The most consequential is the absence of numerical results for the primary efficacy endpoint. Although the endpoint definition and intended confidence-interval approach are described, the response count, proportion, interval, and analysis denominator are not reported. The listed secondary outcomes likewise have no numerical estimates in the available record.

The analysis population description is incomplete and internally inconsistent. It states that the ITT F population includes participants with at least one post-baseline assessment, while also saying participants without such an assessment were included. The record does not explain how this was operationalized or how missing assessments were handled. The number enrolled is listed as 34, but no primary endpoint analysis count is given.

The safety summaries use denominators of 41 for the core phase and 31 for follow-up, both of which differ from the enrollment figure. No explanation is reported. The record does not clarify whether these values reflect a different safety population, a record-specific counting convention, or another issue; no explanation should be assumed. The relationship between these phase-specific safety groups and the efficacy population is also not described.

The study has no comparison group and is open label. These design features limit causal interpretation and leave the possibility that outcome assessment could be affected by knowledge of treatment, depending on assessment procedures. The registry does not report whether independent or masked assessment procedures were used.

There is also a phase-label discrepancy: the formal phase field is PHASE4, while the brief summary calls the trial phase II. The record does not resolve this. It does not report a sample-size justification, interim analyses, a multiplicity plan, or a statistical analysis plan. These omissions limit evaluation of the prespecified inferential framework.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Treatment of Patients With Metastatic or Unresectable Gastrointestinal Stromal Tumors in First Line With Nilotinib — NCT00756509.
• Registry record last updated 2026-07-16; status last verified 2026-07.
• Results first posted 2026-01-20.
• 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

The registry record describes NCT00756509, an open-label, multicenter, single-arm study of nilotinib for adults with metastatic or unresectable GIST in first-line treatment. It lists 34 enrolled participants, a completed status, and a PHASE4 designation, while the brief summary calls the study phase II.

The primary endpoint is the proportion with CR, PR, or SD by month 6 according to RECIST v1.0. The record specifies an exact confidence-interval approach but does not report the primary estimate, interval, or analysis denominator. It lists 7 outcome measures and 0 analysis records.

Safety summaries report phase-specific numbers at risk and counts of participants affected by serious and other events, plus selected serious-event terms. The safety denominators do not match the listed enrollment, and the record does not explain the discrepancy. The available information supports description of the design and endpoint definitions, but not a numerical efficacy conclusion or a comparison with another treatment.