Panitumumab, Regorafenib, or TAS-102, in Treating Patients With Metastatic and/or Unresectable RAS Wild-Type Colorectal Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Panitumumab, Regorafenib, or TAS-102, in Treating Patients With Metastatic and/or Unresectable RAS Wild-Type Colorectal Cancer (NCT03992456), a randomized phase 2 study. The primary outcome is Overall Survival (OS).

2. Clinical Question

This is a comparison of treatment strategies rather than a comparison of panitumumab against a single fixed drug. The treating physician selects regorafenib or trifluridine and tipiracil hydrochloride for patients in Arm B. The registry also lists progression-free survival, overall response rate, clinical benefit rate, adverse events, and patient-reported quality of life among the secondary outcomes. The primary outcome reported is median overall survival, with a stated timeframe of 2 years. The registry record does not report detailed eligibility criteria, including how RAS wild-type status or resistance mutations were assessed, or the reasons for selecting a particular standard-therapy option.

3. Trial Design

The registry describes PULSE as a randomized, phase II, open-label study with a parallel design. It compares panitumumab in Arm A with a standard-therapy strategy in Arm B. The comparator strategy consists of regorafenib or trifluridine and tipiracil hydrochloride, selected at the discretion of the treating physician. No masking is reported. The registry record does not report the allocation ratio or the method used to generate or conceal the randomization sequence.

Panitumumab is administered intravenously over 30–90 minutes on days 1 and 15, with treatment repeating every 28 days for a maximum of 24 cycles, unless disease progresses or toxicity is unacceptable. The standard-therapy options have different schedules. Trifluridine and tipiracil hydrochloride is taken by mouth twice daily on days 1–5 and 8–12; regorafenib is taken by mouth once daily on days 1–21. Treatment repeats every 28 days for a maximum of 24 cycles, subject to the same stated progression and toxicity conditions.

The registry reports enrollment of 12. The posted outcome groups list 5 participants for Arm A and 6 for Arm B. These are the outcome-group counts reported for the overall-survival results; the registry record does not explain the difference between those counts and the enrollment figure. The trial start date is 2020-04-24, and the primary completion date is 2023-12-01. The status is listed as unknown.

The primary outcome is overall survival, defined as the time from randomization to death from any cause. The registry specifies Kaplan-Meier estimation of the survival distribution and an un-stratified log-rank test to compare the 2 treatment arms. Secondary outcomes include progression-free survival, overall response rate, clinical benefit rate, incidence of adverse events, and patient-reported quality of life. Detailed analysis populations, missing-data procedures, and other elements of the statistical plan are not reported in 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

Reasons for not completing this period:

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 Overall Survival (OS), measured over 2 years.

Overall survival measures the time from randomization until death from any cause. It captures the length of time participants remain alive after assignment, without restricting the endpoint to deaths attributed to the cancer or to a particular cause. The registry identifies overall survival as the primary outcome and gives a timeframe of 2 years.

The planned summary is the median overall-survival time for each treatment arm, with a 95% confidence interval. The median is the point in the estimated survival distribution at which half of the participants are estimated to have experienced the event and half have not yet experienced it. This is a general interpretation of a survival median; it does not mean that every participant had follow-up lasting to that point, nor does it describe the experience of every participant.

Survival outcomes are commonly analyzed with methods that account for censoring: a participant may have a last known follow-up before death, so the exact later survival time is not observed. The registry states that the Kaplan-Meier method will estimate the survival distribution. The posted median estimates are 7.5 months for Arm A and 8.3 months for Arm B. The upper confidence limits are not reported in the registry record. A confidence interval describes uncertainty around an estimate under the assumptions of the analysis; it is not a range containing every participant’s survival time.

The registry names an un-stratified log-rank test for comparing the arms. The reported p-value is 0.513. A p-value summarizes how compatible the observed test statistic is with the test’s null hypothesis under its assumptions; it is not the probability that a treatment works, or the probability that the null hypothesis is true. The reported summary does not establish the size or precision of a treatment effect on its own.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry specifies a time-to-event primary outcome and names two methods: Kaplan-Meier estimation for the overall-survival distribution and an un-stratified log-rank test for comparison of the 2 treatment arms. The posted results give median survival estimates, partial confidence-interval information, group counts, and the log-rank p-value. No hazard ratio is reported. The concepts below are general explanations of the methods and their interpretation.

What does Kaplan-Meier estimation do?

Kaplan-Meier estimation describes the probability of remaining free of the event over time. For overall survival, the event is death from any cause. Unlike an analysis that needs every participant’s event time to be observed, this method can use information from participants whose follow-up ends before death is recorded. Such observations are censored: the analysis knows that the participant remained alive up to a last observed time, but does not know the later event time.

Censoring allows participants with incomplete follow-up to contribute information up to the point they were last observed. It does not make the later outcome known. The method relies on assumptions about censoring and observation, including that censoring does not systematically remove participants with different underlying survival prospects in a way that distorts the estimate. The registry record does not describe censoring patterns or provide the numbers at risk at different times. Consequently, readers cannot assess the amount of information remaining at later points in follow-up from the posted summaries alone.

A Kaplan-Meier curve can describe the estimated survival distribution across follow-up, while the median compresses that distribution into a single summary. The median is not a comparison of all survival times and does not reveal whether the curves differ at particular times or whether their shapes are similar. The registry states that the distribution will be estimated using Kaplan-Meier, but the posted numerical results list the median estimates rather than a complete set of curve coordinates or time-specific survival probabilities.

What does the log-rank test assess?

The log-rank test compares event experience over follow-up between groups. In broad terms, it contrasts the observed event pattern with what would be expected under a null hypothesis of no difference between the groups’ survival distributions. It uses information across event times rather than comparing only the two medians. The registry specifies an un-stratified test, meaning that the described comparison does not report adjustment or stratification by prognostic factors.

The reported p-value is 0.513. A p-value is a measure of how unusual the observed test result, or a more extreme result, would be under the test’s null hypothesis and assumptions. It is not a measure of effect size, and it does not state the probability that the null hypothesis is true. A non-small p-value cannot establish that the treatment strategies have identical outcomes. A small p-value, in general, would not by itself tell readers whether an estimated effect is large, precise, clinically important, or free of bias.

The registry record does not report a hazard ratio, a confidence interval for a hazard ratio, or a full test statistic. It is therefore not possible to use the posted results to describe a relative hazard estimate or its uncertainty. The p-value alone is not a substitute for that effect estimate. It also does not supply a complete account of absolute survival differences across follow-up.

How should the median and confidence interval be read?

The posted median estimates are 7.5 months for Arm A and 8.3 months for Arm B. A median is a location summary of the estimated distribution, not a guarantee about the survival time of any particular participant. The registry lists 95% confidence intervals, but only lower limits are present in the posted arm summaries: 7.2 months for Arm A and 4.9 months for Arm B. Upper limits are not reported in the registry record.

A confidence interval expresses uncertainty in an estimate using a method with stated coverage properties under its assumptions. It should not be interpreted as a range of individual participant outcomes. In survival analysis, uncertainty can be substantial when event information is limited or follow-up is incomplete. The listed group counts are 5 and 6. The registry does not provide the event counts for the primary outcome, the number censored, the confidence-interval calculation details, or the follow-up distribution, so the precision of the medians cannot be fully assessed from these results.

What does the comparator structure imply?

Arm B is not a single drug regimen. Participants receive trifluridine and tipiracil hydrochloride or regorafenib, with the choice left to the treating physician. The arm-level outcome therefore combines results across those treatment options. This can be a relevant comparison when the research question concerns panitumumab against a usual-care strategy, but the registry record does not report treatment-specific outcome results, selection criteria, or how many participants received either option. The posted overall-survival summary cannot answer separate questions about each comparator drug.

Randomization can support a causal comparison when it is properly implemented and when the analysis preserves the randomized groups, but the registry record does not describe sequence generation, allocation concealment, or the analysis population. Open-label status means masking is listed as none; it does not establish whether outcome ascertainment or follow-up could have been influenced. Death is a relatively objective event, but complete and comparable follow-up remains important. These are general considerations, not additional findings about this trial.

What can be concluded from the reported analysis?

The posted result establishes that the registry reports median overall survival estimates and an un-stratified log-rank p-value for the arm comparison. It does not establish a precise treatment effect or equivalence. The limited outcome-group counts, incomplete confidence limits, absent hazard-ratio estimate, and unreported details of follow-up and analysis populations restrict interpretation. No recalculated comparison is available because the posted summary statistics are missing or not estimable for that calculation. Any planning illustration based on an observed effect is not observed power and cannot replace the trial’s reported analysis or describe what the trial was designed to detect.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparison as an un-stratified log-rank test, with p = 0.513. It also reports median overall survival of 7.5 months for Arm A and 8.3 months for Arm B. The sponsor’s analysis is the reported inferential result for this endpoint.

A recalculated two-arm comparison is not available: the registry record identifies values as missing or not estimable for a median-time comparison. The posted arm summaries do not include the information needed to reconstruct a time-to-event comparison, such as the underlying event times or a reported hazard ratio. No independent recalculated p-value or effect estimate should therefore be substituted for the sponsor’s log-rank result.

The absence of a recalculation does not validate every element of the reported analysis. It means the posted summary statistics do not support that particular recalculation. The reported p-value should be interpreted together with the medians, the limited group counts, the incomplete confidence-interval information, and the missing follow-up and analysis details. The registry record does not report a hazard ratio or a full confidence interval for a comparative effect. It also does not report enough detail to assess the analysis population or censoring. Thus, the sponsor’s test is the available inferential comparison, while the precision and magnitude of the treatment difference remain difficult to characterize from the registry results.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists several secondary outcomes: progression-free survival, overall response rate, clinical benefit rate, incidence of adverse events, and patient-reported quality of life. It reports 9 outcome measures and 5 analyses, and states that 5 secondary outcomes are posted. The registry record does not report how the primary and secondary hypotheses were ordered, whether any outcomes were designated as key secondary endpoints, or whether testing was adjusted for multiple comparisons.

When many outcomes are examined, the chance of obtaining at least one seemingly unusual result can increase. Adjustment methods or a prespecified testing sequence can address some multiplicity concerns, but neither an adjustment approach nor a testing hierarchy is reported in the registry record. The primary log-rank p-value should therefore be read as the reported test result, without assuming that it is adjusted for other outcomes or that the broader family of comparisons was controlled.

The record does not report the analysis population for overall survival or the secondary outcomes. It does not state whether analyses followed randomized assignment, whether participants were excluded from particular analyses, or how missing outcome data were handled. For time-to-event outcomes, the treatment of censoring and loss to follow-up is important; the registry record does not provide detailed censoring information or numbers at risk over time. For patient-reported quality of life, missing questionnaires and the timing of assessments can affect interpretation, but the handling of those issues is not reported.

The registry’s planned primary analysis is a Kaplan-Meier estimate and un-stratified log-rank test. It does not report interim analyses, stopping rules, a detailed statistical analysis plan, or a sample-size justification. Those items are not reported in the registry record; their absence from the record does not show what was or was not considered outside it. The record also lists an enrollment of 12 and outcome-group counts of 5 and 6, without explaining the difference. That limits assessment of the population contributing to the primary analysis.

These gaps do not change the numerical results that are posted, but they constrain the assessment of their reliability and scope. In particular, readers cannot determine from the registry record whether multiplicity was controlled, how missing data were addressed, or which participants contributed to each analysis.

12. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: back pain (16.7%), muscle weakness lower limb (16.7%).

The posted safety figures list serious adverse events among 5 participants at risk in Arm A and 6 in Arm B. The affected counts are 0 for Arm A and 2 for Arm B. The figures also list other adverse events among 5 participants at risk in Arm A and 6 in Arm B, with affected counts of 5 and 6, respectively. The registry record lists deaths of 5 in Arm A and 5 in Arm B.

These are descriptive counts, not a randomized statistical comparison. The registry record does not provide a comparative test, confidence intervals, event timing, severity grading details beyond the serious-adverse-event grouping, or causal attribution for the listed deaths. The deaths should not be interpreted as necessarily caused by treatment, and the adverse-event counts do not establish a treatment-related difference. The safety figures also do not explain whether an affected participant could contribute to more than one event term.

The most frequently listed serious adverse-event terms are back pain and muscle weakness lower limb, each reported as 16.7%. The registry does not provide further context for those percentages in the posted summary, such as the number of events per participant or the full accounting rules for event terms. Small group counts mean that these descriptive figures can be sensitive to the experience of individual participants. They should be read as the registry’s reported safety summaries, not as precise estimates of comparative risk.

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

The registry results describe overall survival in two randomized treatment arms: panitumumab, and a standard-therapy strategy in which the treating physician selects regorafenib or trifluridine and tipiracil hydrochloride. The posted median estimates are 7.5 months and 8.3 months, respectively, with group counts of 5 and 6. The reported un-stratified log-rank p-value is 0.513. These registry entries describe the posted comparison; they do not establish that the strategies are equivalent or that one is superior.

The p-value is not a measure of the probability that a treatment works, nor does it quantify how large a difference may be. The median estimates alone do not show the full survival experience over time. Upper confidence limits for the medians are not reported in the registry record, and no hazard ratio or confidence interval for a relative effect is reported. The precision of the comparative effect therefore cannot be characterized fully from the posted primary results.

The standard-therapy group combines two options, chosen at the treating physician’s discretion. The arm-level result cannot be interpreted as a separate estimate for regorafenib or for trifluridine and tipiracil hydrochloride. The registry record does not report how many participants received each option or treatment-specific outcomes. The findings also should not be generalized beyond the population and setting described in the registry without additional evidence.

The trial is listed as phase II, randomized, open-label, and parallel. The registry reports enrollment of 12, while the outcome groups list 5 and 6. It does not explain this difference or provide detailed analysis-population and follow-up information. Those limits, together with the small reported groups and incomplete precision measures, make the result uncertain. The posted findings are a record of this trial’s reported outcomes; they do not resolve the broader clinical question or determine an individual treatment choice.

14. Why This Trial Matters Statistically

This study raises a statistical question about retreatment with panitumumab after progression on anti-EGFR therapy, compared with a standard-therapy strategy for patients with metastatic and/or unresectable RAS wild-type colorectal cancer. The endpoint is overall survival, a time-to-event outcome that requires attention to censoring, follow-up, and the pattern of events over time. The registry’s stated use of Kaplan-Meier estimation and the log-rank test reflects methods designed for survival data rather than a simple comparison of averages.

The result also illustrates the distinction between a test of evidence and an estimate of effect. The registry reports p = 0.513, but does not report a hazard ratio or its confidence interval. A p-value alone cannot convey the magnitude or precision of the treatment contrast. The median estimates add a descriptive summary, but the reported confidence limits are incomplete. A fuller statistical account would include comparative effect estimates and uncertainty, as well as information about follow-up and censoring.

The comparator structure is important to the estimand: Arm B includes regorafenib or trifluridine and tipiracil hydrochloride, selected by the treating physician. The comparison therefore concerns a treatment strategy, not a single fixed drug. If treatment-specific outcomes are not reported, an arm-level result cannot distinguish the effects of the options included in that strategy.

The registry lists enrollment of 12 and outcome-group counts of 5 and 6. The difference is unexplained in the record, and the small group counts limit the precision that can be expected from the posted summaries. This makes transparent reporting of analysis populations, missingness, event counts, and uncertainty especially important. The record’s limitations show why a p-value should be considered alongside design, endpoint, effect estimates, and data completeness rather than in isolation.

15. 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 posted primary outcome summaries have important limitations. Median overall survival is reported for each arm, but the upper confidence limits are not reported in the registry record. No hazard ratio or confidence interval for a comparative effect is posted. As a result, the magnitude and precision of the difference between the survival experiences cannot be fully characterized. The reported log-rank p-value does not fill that gap.

The outcome-group counts are 5 for Arm A and 6 for Arm B, while enrollment is listed as 12. The registry record does not explain this difference or specify the analysis population. It also does not report the number of deaths contributing to the primary endpoint, the number censored, numbers at risk during follow-up, or detailed follow-up duration. These omissions make it difficult to judge the amount and distribution of information underlying the survival estimates.

The comparator arm allows physician choice between regorafenib and trifluridine and tipiracil hydrochloride. The registry does not report treatment-specific counts, selection criteria, or separate outcome estimates for these options. Thus, the result describes the combined strategy and cannot be used to compare panitumumab separately with each comparator drug.

The record describes an open-label randomized design but does not report the randomization method, allocation concealment, analysis population, or missing-data procedures. It does not report a detailed statistical analysis plan, interim analyses, stopping rules, multiplicity handling, or sample-size justification. These details are not reported in the registry record, so it is not possible to assess them from the posted information.

The registry status is unknown. Although a primary completion date is listed, the record does not provide a complete account of follow-up or explain the status. The limitations do not negate the reported results, but they narrow the conclusions that can be drawn from them and leave substantial uncertainty about comparative effect and precision.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Panitumumab, Regorafenib, or TAS-102, in Treating Patients With Metastatic and/or Unresectable RAS Wild-Type Colorectal Cancer — NCT03992456.
• Registry record last updated 2024-05-21; status last verified 2024-04.
• Results first posted 2024-05-21.
• 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.

19. Record Summary

NCT03992456 is a randomized, phase II, open-label, parallel study of panitumumab versus standard therapy for metastatic and/or unresectable RAS wild-type colorectal cancer after progression on anti-EGFR therapy. The standard-therapy arm uses regorafenib or trifluridine and tipiracil hydrochloride at the treating physician’s discretion. The registry lists enrollment of 12, a start date of 2020-04-24, primary completion on 2023-12-01, and status as unknown.

The primary outcome is overall survival, defined as time from randomization to death from any cause, with a stated timeframe of 2 years. The reported median estimates are 7.5 months for panitumumab and 8.3 months for standard therapy. The un-stratified log-rank test reports p = 0.513. Outcome-group counts are 5 and 6; the upper confidence limits for the medians and a hazard ratio are not reported in the registry record.