Study of MK-3475 in Patients With Microsatellite Unstable (MSI) Tumors (Cohorts A, B and C): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of MK-3475 in Patients With Microsatellite Unstable (MSI) Tumors (Cohorts A, B and C) (NCT01876511), a non-randomized phase 2 study. The primary outcome is Immune-related Progression Free Survival (irPFS) at 20 Weeks in MSI Positive and Negative Colorectal Adenocarcinoma Participants Using Immune Related Response Criteria (irRC) During Stages 1 and 2.

2. Clinical Question

The clinical question is thus a cohort-specific assessment, rather than a randomized test of MK-3475 against another treatment. Cohort C broadens the study to MSI-positive non-colorectal cancers, but it is not part of the stated primary endpoint population. Secondary outcomes include overall survival, response and disease-control measures, progression-free survival at 28 weeks, toxicity, and whether MSI predicts treatment response. Those outcomes address distinct aspects of activity and safety; they do not replace the prespecified primary question. The registry record does not report results for these outcomes.

3. Trial Design

This was a phase 2, non-randomized, unmasked, parallel-cohort study of MK-3475. The registry lists an enrollment of 113 and describes three cohorts: MSI-positive colorectal cancer (Cohort A), MSI-negative colorectal cancer (Cohort B), and MSI-positive non-colorectal cancer (Cohort C). The sponsor was Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins. The listed study start is 2013-09, primary completion is 2019-08, and status is completed.

Non-randomized allocation means that the registry does not describe random assignment between treatment strategies. The parallel label identifies distinct cohorts studied alongside one another; it does not establish that participants in one cohort serve as a valid control group for another. In particular, MSI status and cancer type distinguish the groups. Differences in their outcomes could reflect those underlying population differences as well as other factors not detailed in the record. The record lists MK-3475 as the intervention and does not describe a comparator intervention.

The primary endpoint was specified for Cohorts A and B only. The registry states that a pre-specified objective was to complete this endpoint in those cohorts and that the enrollment goal during stages 1 and 2 was 25 for Cohort A and 25 for Cohort B. It also says that 24 subjects were enrolled in Cohort A and that additional subjects were enrolled during the second expansion, but the registry text ends before completing that explanation. The endpoint’s stages are mentioned, but their decision rules, stopping boundaries, and rationale are not reported in the registry record.

The registry lists 12 outcome measures and reports 0 analyses. A completed study status does not itself provide an outcome estimate. The posted record does not give a primary endpoint value, confidence interval, or treatment-effect estimate, so the design information alone cannot establish what was observed.

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:

Analysis population for the primary outcome, as defined in the registry: It was a pre-specified objective to complete this objective in Cohort A and B only. Per protocol during stages 1 and 2, the study enrollment goal was 25 for Cohort A and B. However, only 24 subjects were enrolled in Cohort A. Additional subjects for this study were enrolled during the second expansion portion of this protocol.

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 Immune-related Progression Free Survival (irPFS) at 20 Weeks in MSI Positive and Negative Colorectal Adenocarcinoma Participants Using Immune Related Response Criteria (irRC) During Stages 1 and 2, measured over 20 weeks.

The primary endpoint is titled immune-related progression-free survival (irPFS) at 20 weeks in MSI-positive and MSI-negative colorectal adenocarcinoma participants, assessed using immune-related response criteria (irRC) during stages 1 and 2. The record says the endpoint applies to Cohorts A and B and is expressed as a percentage of participants. It specifies estimation from a Kaplan-Meier curve and a 95% confidence interval.

The description defines response categories using changes in tumor burden and whether disease meets criteria for complete response, partial response, stable disease, or progressive disease. It defines complete response as disappearance of all target lesions; partial response as a decrease in tumor burden by 50% or greater at a consecutive assessment at least 4 weeks after first documentation; stable disease as not meeting the criteria for complete or partial response in the absence of progressive disease; and progressive disease as at least 25% increase in tumor burden relative to nadir.

There is an important wording ambiguity. The endpoint title says progression-free survival, but its description calls the irPFS rate the percentage of patients with disease progression or relapse from complete response, or death, at 20 weeks. That is phrased as the occurrence of an event rather than being free of an event. The registry record does not clarify the discrepancy, so readers should not silently reinterpret the stated definition. The primary endpoint result is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The record identifies several statistical features of the planned primary endpoint: a time point of 20 weeks, a percentage scale, a 95% confidence interval, and estimation using a Kaplan-Meier curve. It does not report an endpoint estimate or a statistical analysis. The distinctions below explain what these methods generally mean; they do not fill in missing details about how this study was analyzed.

What does a Kaplan-Meier estimate represent?

In general, Kaplan-Meier methods estimate the probability of remaining free of a defined event over time while accounting for participants whose event status is not observed through the full follow-up period. Such observations are commonly called censored: the participant’s subsequent event time is unknown, rather than known to be event-free indefinitely. A Kaplan-Meier curve is built from event times and the participants still at risk at each event time. Its interpretation depends on a clearly specified event, appropriate follow-up information, and assumptions about censoring.

The registry says that the primary estimate is based on a Kaplan-Meier curve, but does not report the event counts, numbers at risk over time, censoring details, curve, or method used to obtain the confidence interval. It also does not explain how the curve was converted into the stated percentage at the time point. Those details are not reported in the registry record.

Why does the endpoint definition matter?

A time-to-event measure needs a precise definition of what counts as an event. The record describes progression, relapse from complete response, and death as events, but the wording says the rate is the percentage of patients with these events. The title, irPFS, ordinarily signals a measure of being free of progression or death, while the description’s phrasing signals the percentage experiencing them. Since those interpretations differ, the event definition should be resolved before interpreting an estimate. The record does not resolve it.

The response criteria also matter because the classification of response or progression determines when an event might be recorded. The description specifies irRC categories and tumor-burden thresholds, including a 50% or greater decrease for partial response and an increase of at least 25% relative to nadir for progressive disease. It does not provide participant-level assessments, adjudication procedures, or details about handling assessments that are missing or occur outside the planned schedule.

What does a confidence interval add?

In general, a confidence interval gives a range of values compatible with the data and the method used, under the model’s assumptions. It is not a guarantee that the true value lies in that particular interval, and it does not by itself establish clinical importance. A point estimate and interval are read together: the point estimate gives the central summary, and the interval shows its statistical uncertainty. Here, the record names a 95% confidence interval but does not report its endpoints, so the degree of precision cannot be assessed.

How should statistical significance be understood?

In general, a significance test evaluates data against a specified null hypothesis using a chosen test and assumptions. Its interpretation depends on the analysis plan, the outcome, the comparison, and the handling of multiple tests. The registry record does not report a primary hypothesis test, a p-value, a null value, or a decision threshold for this endpoint. It would therefore be inappropriate to infer statistical significance from the study’s phase, completion status, or enrollment.

What can a sample-size planning illustration tell readers?

In general, a sample-size calculation describes how many participants or events a study might require under stated assumptions about an effect, variability, allocation, significance level, and desired power. It is a planning exercise, not evidence that an effect occurred and not a measure of the achieved power after results are observed. Any illustration based on an observed effect is especially dependent on that estimate and should not be confused with the trial’s own design rationale.

The registry record gives no sample-size justification, planned power, or statistical design calculation for the primary endpoint. It does mention enrollment goals for Cohorts A and B during stages 1 and 2, but does not state the rationale or the operating rules for those stages. Those missing details prevent a determination of whether the observed enrollment was considered adequate under the trial’s original design assumptions.

More broadly, statistical methods are not fully described by naming an endpoint and an estimation method. The analysis population, treatment of missing assessments, censoring rules, confidence-interval construction, and any adjustments for multiple outcomes can affect interpretation. These details are not reported in the registry record. General descriptions of Kaplan-Meier estimation and confidence intervals explain the concepts, but they cannot be used to reconstruct this study’s analysis.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Cohort A: MSI Positive Colorectal Cancer", "Cohort B: MSI Negative Colorectal Cancer")
n      <- c(24, 25)
events <- c(21, 3)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Cohort A: MSI Positive Colorectal Cancer  24  21  87.5  67.6  97.3  69.0  95.7  68.8  95.5
ARM  Cohort B: MSI Negative Colorectal Cancer  25  3  12.0  2.5  31.2  4.2  30.0  4.4  30.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 12 outcome measures, including overall survival, immune-related and conventional progression-free survival, objective response, disease control, treatment-related toxicity, and whether MSI predicts response. Testing or estimating many outcomes raises a general multiplicity concern: when numerous statistical tests are conducted, the chance of at least one apparently notable result can increase unless the analysis plan addresses the issue. Different endpoints may also answer different questions and use different criteria, so their results should not be treated as interchangeable.

The registry record does not report a multiplicity strategy, a hierarchy among the outcomes, or which analyses were considered confirmatory versus exploratory. It also lists 0 analyses. No adjustment method or interpretation plan can therefore be attributed to the study from this record.

Missing-data handling is not reported in the registry record. For a time-to-event analysis, incomplete follow-up and censoring need to be distinguished from documented absence of an event. For response assessments, missing visits or unevaluable scans may affect how participants are classified. The record does not say how either situation was handled, nor does it provide a sensitivity analysis.

Analysis populations are also not reported in the registry record. The endpoint is specified for Cohorts A and B, but the record does not explain which enrolled participants were included in the primary analysis, how eligibility for analysis was determined, or whether any participants were excluded. The enrollment goal and actual Cohort A count do not establish the analysis denominator. These omissions limit assessment of completeness and potential bias.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: diabetic ketoacidosis (4.3%), acute kidney injury (2.4%), arthritis (2.4%), pancreatitis (2.4%), myocarditis (2.4%).

The registry reports serious adverse event summaries by cohort. For Cohort A, 41 were at risk for serious adverse events and 4 were reported as affected; for Cohort B, 25 were at risk and 0 were affected; for Cohort C, 47 were at risk and 7 were affected. These are cohort-specific counts. They should be read with their stated at-risk denominators rather than compared as if the cohorts were randomized or otherwise directly comparable.

The record also lists other adverse events separately: 41 at risk and 41 affected in Cohort A, 25 at risk and 3 affected in Cohort B, and 47 at risk and 47 affected in Cohort C. The record does not give further definitions or event-level detail in these summary fields, so the counts should not be taken to describe severity, duration, or causality beyond the labels used.

Deaths are listed as 3 for Cohort A, 0 for Cohort B, and 2 for Cohort C. The registry summary does not state that these deaths were caused by treatment, nor does it give a cause-of-death breakdown in these figures. The listed serious terms include diabetic ketoacidosis (4.3%), acute kidney injury (2.4%), arthritis (2.4%), pancreatitis (2.4%), myocarditis (2.4%), and infusion reaction (2.4%). The denominator and calculation basis for these percentages are not reported in the registry record. These summaries do not establish comparative safety or causation.

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

The registry record establishes that a phase 2 study evaluated MK-3475 in three MSI-defined cancer cohorts and that a primary 20-week irPFS objective was specified for the two colorectal cancer cohorts. It also records secondary outcomes and cohort-level adverse-event summaries. These are descriptions of study design and posted registry information, not a complete report of the trial’s efficacy findings.

The record does not report a primary endpoint estimate, its 95% confidence interval, or results for the listed secondary outcomes. It also lists 0 analyses. As a result, readers cannot determine from this record the observed primary endpoint value, the precision of that value, or whether any stated statistical criterion was met. The absence of posted analyses should not be interpreted as evidence of either benefit or lack of benefit.

The design does not provide a randomized comparison. Cohorts were organized by cancer type and MSI status, and no like-for-like control comparison is described. Even if cohort outcome estimates were available, differences across these groups could not automatically be attributed to treatment because the groups differ in their defining clinical characteristics. The record does not report a causal comparison that would separate treatment effects from these differences.

The primary endpoint description itself leaves an unresolved interpretive issue: it labels the measure progression-free survival while defining the rate as the percentage with progression or death. Without clarification, the direction and meaning of a hypothetical estimate would be uncertain. The record also does not report analysis-population rules, missing-data handling, or details of the staged design that would help assess the result.

The safety summaries report numbers of participants affected and deaths by cohort, along with several serious adverse event terms and percentages. They do not, on their own, establish whether an event was caused by MK-3475 or whether safety differs between cohorts. Overall, the registry record supports a description of the study’s intended questions and selected safety counts, but not a judgment about treatment effectiveness, comparative benefit, or causal safety effects.

14. Why This Trial Matters Statistically

This trial raises several statistical issues that are important when reading clinical research. First, the study separates participants into groups defined by MSI status and cancer type. Such cohorts may be useful for describing outcomes in distinct populations, but a non-randomized design does not make those groups equivalent. A direct difference between cohorts would mix treatment experience with population differences unless an appropriate analysis and design supported a stronger interpretation.

Second, the primary endpoint combines a time point, an immune-related response framework, and a time-to-event estimation method. Each piece has to align: the event definition determines what is counted, response criteria determine how progression is classified, and the survival method determines how incomplete follow-up is handled. The registry’s mismatch between the endpoint title and its event-rate wording shows why precise endpoint definitions are not merely technical details. They determine what an estimate means.

Third, the record lists multiple outcomes spanning survival, tumor response, toxicity, and a potential marker question. These outcomes answer different scientific questions and may involve different analysis methods. Without a reported multiplicity plan or posted analyses, readers cannot tell which findings were designated as primary evidence or how uncertainty across outcomes was managed.

Finally, registration and completion status do not replace a results report. The record provides design information and some safety summaries, but not the primary estimate, interval, or analysis details needed to assess precision and validity. The distinction between a prespecified question and a reported result is central to statistical interpretation: knowing what was intended to be measured does not reveal what was observed.

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 most consequential limitation for interpreting outcomes is that the primary endpoint estimate and its 95% confidence interval are not reported in the registry record. Although the study is marked completed, no analysis results are listed. The record therefore cannot support an assessment of the magnitude or precision of the primary outcome.

The endpoint description is internally ambiguous. It names irPFS but defines the rate as the percentage experiencing progression, relapse from complete response, or death at 20 weeks. The record does not clarify whether this is intended as an event proportion or a progression-free estimate. Its description of the primary population also ends mid-sentence after noting additional enrollment during the second expansion, leaving that explanation incomplete.

The design is non-randomized and does not identify a concurrent comparator. Cohort membership reflects MSI status and cancer type, so the cohorts are not like-for-like controls. The registry’s enrollment count and cohort-specific safety counts cannot resolve confounding or establish comparative treatment effects.

Important analysis details are absent, including the analysis population, missing-data rules, censoring details, confidence-interval method, multiplicity handling, and the staged-design rules. The registry record also does not provide the denominator or calculation basis for the percentages listed with serious adverse event terms. These gaps limit reproducibility and make it difficult to judge bias, precision, or the appropriate interpretation of any outcome.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study of MK-3475 in Patients With Microsatellite Unstable (MSI) Tumors (Cohorts A, B and C) — NCT01876511.
• Registry record last updated 2020-02-06; status last verified 2020-02.
• Results first posted 2020-01-07.
• 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

NCT01876511 is a completed, non-randomized phase 2 study of MK-3475 in MSI-positive colorectal cancer, MSI-negative colorectal cancer, and MSI-positive non-colorectal cancer. The registry lists enrollment of 113, with the primary endpoint specified for Cohorts A and B: irPFS at 20 weeks using irRC and Kaplan-Meier estimation. It names a 95% confidence interval.

The registry lists 12 outcome measures and 0 analyses. It does not report the primary endpoint result, its confidence interval, or results for the secondary outcomes. It reports cohort-level adverse-event counts and deaths, but these summaries do not establish causality or comparative safety. Interpretation is further limited by the non-randomized cohort design, missing analysis details, and ambiguity in the primary endpoint wording.