Evaluating the Combination of MK-3475 and Sterotactic Body Radiotherapy in Patients With Metastatic Melanoma or NSCLC: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Evaluating the Combination of MK-3475 and Sterotactic Body Radiotherapy in Patients With Metastatic Melanoma or NSCLC (NCT02407171), a non-randomized phase 1, phase 2 study. The primary outcome is Phase 2, Overall Response.

2. Clinical Question

The clinical question combines two different settings. In the melanoma group, patients had already progressed on anti-PD-1 therapy before enrollment. In the NSCLC group, patients began MK-3475 without prior anti-PD-1 or anti-PD-L1 treatment and proceeded to SBRT after progression. These groups differ in cancer type and treatment history, so their results would need to be read in their respective contexts rather than treated as interchangeable estimates.

The proposed sequence makes the endpoint a description of response after local radiation and renewed MK-3475, not a direct randomized test of SBRT against no SBRT. The registry record does not report a control group receiving the same treatment sequence without radiation. It therefore does not establish what response would have occurred in the absence of SBRT.

3. Trial Design

This was a prospective, nonrandomized, unmasked Phase I/II trial with a parallel model. Yale University was the sponsor. The record lists an enrollment of 61 and describes the study as completed. The recorded start date is 2015-04-01 and the primary completion date is 2019-02-19.

The trial had two parts. Phase I was a radiation dose-escalation study conducted with a constant MK-3475 dose. The record describes dose-escalation cohorts for lung and non-lung targets. It also describes Phase 2a expansion cohorts for melanoma and NSCLC, treated at the maximum tolerated dose identified in Phase I. Details of the dose-escalation decision process and the final selected radiation dose are not reported in the registry record.

The treatment sequence differed according to cancer type and prior treatment. Patients with metastatic melanoma were enrolled after progression on anti-PD-1 therapy. Patients with metastatic NSCLC who had not previously received anti-PD-1 or anti-PD-L1 therapy were treated with MK-3475 until progression. At progression, one target lesion was selected for SBRT; MK-3475 was then restarted and continued until further progression.

Because treatment allocation was not randomized and masking was absent, the study is not a blinded comparison of assigned treatments. Its cohorts reflect the trial’s dose-escalation and expansion structure and differing clinical histories. The record does not describe a concurrent control group, random assignment, or a prespecified between-group treatment contrast. Consequently, comparisons across cohorts would be descriptive and potentially affected by differences in cancer type, prior immunotherapy, disease course, and treatment phase.

The primary endpoint was a Phase 2 response count among participants who received SBRT. The stated assessment window was up to 12 months. The registry record lists two outcome measures, but details of the other measure are not reported in the registry record. The response category counts, analysis methods, and the number assessed for the primary endpoint are also 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.

Phase 1 Dose Escalation

Phase 2 Treatment

Analysis population for the primary outcome, as defined in the registry: Only those that received SBRT.

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.

Population: Baseline characteristics are presented for participants in the study arm in which they began the study. Phase 1 participants included in the Phase 2 portion are presented in a single arm as the enrollment numbers are presented combined overall for both Phases.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Phase 2, Overall Response, measured over up to 12 months.

The Phase 2 primary endpoint records the number of participants in each response category after SBRT and MK-3475. The categories are complete response, partial response, stable disease, progressive disease, and not evaluable. The registry describes response assessment using Response Evaluation Criteria in Solid Tumors (RECIST), with a timeframe of up to 12 months. The endpoint population is limited to those who received SBRT.

A categorical response endpoint can describe the range of observed disease status rather than reducing every participant to a single continuous measurement. Complete and partial response represent response categories; stable disease and progressive disease distinguish other recorded outcomes. Not evaluable is important because it separates participants without an assessable response classification from those assigned one of the disease-status categories. The registry record does not give the criteria used to assign this category or explain how such participants were handled in any analysis.

In general, response categories are used in oncology trials because they provide a standardized way to classify observed tumor response over a defined assessment period. They can support a clear descriptive account of outcomes, but interpretation depends on who was assessed, when assessments occurred, and how missing or unevaluable observations were treated. Those details are not reported in the registry record.

The registry record describes the endpoint and its population but does not report the response counts. Thus, the endpoint definition can be explained, but the proportion or distribution of observed responses cannot be determined from the registry record. No response estimate or precision interval is reported there.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The trial has a dose-escalation component and expansion cohorts, but the registry record does not report the full statistical analysis plan. The following explains the statistical ideas relevant to interpreting the design and endpoint; general explanations are not claims about analyses actually conducted.

What does a nonrandomized design imply?

In a randomized trial, assignment by chance can help balance measured and unmeasured characteristics between treatment groups. This trial is recorded as nonrandomized, and no masking was used. Its cohorts were organized around cancer type, treatment history, phase, and radiation target. Those features mean group outcomes can differ for reasons other than the treatment sequence under study. A comparison between cohorts would not, by itself, estimate the causal effect of adding SBRT.

The trial’s sequence also complicates attribution. Participants received MK-3475 before SBRT in the NSCLC description, then resumed it after radiation. An observed response after the combined sequence could reflect several influences, including the ongoing treatment and the underlying disease course. A control group undergoing the same sequence without SBRT would be needed for a direct comparison of that component; such a control is not described in the registry record.

What does a response count represent?

The primary outcome is categorical: participants are classified as having complete response, partial response, stable disease, progressive disease, or being not evaluable. Counts in these categories describe how many participants fall into each classification. They are not, without further analysis, a measure of the difference between treatments. The endpoint population is those who received SBRT, which is narrower than a general statement about everyone enrolled.

In general, a binary summary may combine categories into an event and no-event outcome, but the choice of which categories count as an event must be defined in advance. The registry record does not define such a binary recoding for this primary endpoint. It describes category counts instead. No event rate, confidence interval, or hypothesis test for the response endpoint is reported in the registry record.

Why does the analysis population matter?

Analysis populations determine which participants contribute to an estimate. The stated primary endpoint population is limited to participants who received SBRT. The record does not report how many participants received SBRT, how many had response assessments, whether all treated participants were included in the endpoint counts, or how participants with missing assessments were handled. These details matter because an estimate among treated and evaluable participants can differ from an estimate defined for everyone enrolled.

The category not evaluable signals that some assessments may not have resulted in an assignable response category, but its count is not reported. In general, excluding unevaluable participants can change the composition of the analyzed group; treating them as a separate outcome preserves their visibility but does not resolve why they were unevaluable. The registry record does not describe the reasons or analysis handling.

What does precision mean?

In general, a point estimate describes the observed data, while a confidence interval communicates uncertainty around that estimate under a statistical model. Wider intervals indicate less precision. Precision depends on the number of observations and the amount of variation in the data, among other factors. Because the primary response counts and confidence intervals are not reported in the registry record, neither the response estimate nor its precision can be evaluated from this record.

A p-value, in general, measures how incompatible the observed data are with a specified null hypothesis under the assumptions of a test. It is not the probability that a treatment works, and it does not measure the size or clinical importance of an effect. The registry record has no posted analysis records for the primary outcome and does not report a p-value or a null hypothesis for that endpoint.

How should planning illustrations be read?

The planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the intended effect is a stated fraction of the observed effect. These are planning assumptions, not evidence that the trial used that design or analysis. The record does not describe a randomized two-arm comparison for the primary response endpoint. Sample-size illustrations based on an observed effect are hypothetical planning calculations; they are not observed power and do not establish what the trial could detect unless the design, effect, and analysis assumptions are appropriate.

In general, sample-size calculations depend on the outcome definition, target effect, variability or event rate, significance level, allocation, and analysis method. Changing any of these assumptions changes the planning result. The registry record does not report the trial’s sample-size justification, and no planning value should be treated as a description of the observed response outcome.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists two outcome measures, while the Phase 2 primary endpoint is described as response counts in several categories. The other outcome measure’s details are not reported in the registry record. The record does not report a multiplicity strategy for handling several endpoints, response categories, cohorts, or possible comparisons.

In general, multiplicity arises when several statistical tests or estimates are considered together. If many tests are conducted, the chance of finding at least one apparently notable result by chance can increase. Prespecified priorities or adjustment methods can help readers understand how such testing was handled. The registry record does not describe whether any such approach was planned or used, so no adjustment or unadjusted interpretation can be attributed to this study.

The record also does not report the response counts or a statistical analysis of the primary endpoint. It therefore does not indicate whether all response categories were analyzed separately, whether categories were combined for any analysis, or whether any subgroup results were planned. The distinction between planned analyses and analyses actually conducted is not available in the registry record.

Missing-data handling is likewise not reported in the registry record. The endpoint includes a not-evaluable category, but there are no counts or definitions explaining when that classification was assigned. The record does not state whether missing assessments were counted as not evaluable, excluded from an analysis, or handled by another method. In general, these choices affect the population represented by an estimate and should be described alongside the outcome results.

The primary endpoint population is described as those who received SBRT, but the record does not give the number assessed or the number with missing outcome information. It also does not describe analysis populations such as an intention-to-treat set or a per-protocol set. Without these details, readers cannot assess how enrollment, treatment receipt, and outcome availability relate to the response analysis.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: gastric hemorrhage (20.0%), nausea (16.7%), vomiting (16.7%), device related infection (10.0%), confusion (10.0%).

The registry record presents safety counts by cohort. For the group titled Phase 1: Lung Targets + MK-3475 / Phase 2 NSCLC, no Prior PD-1 Therapy, 10 participants were at risk for serious adverse events and 5 were affected; 10 were at risk for other adverse events and 10 were affected; deaths were 0. For Phase 1: Lung Targets / Phase 2: NSCLC, Previously Treated With Immunotherapy/irPD, the corresponding counts were 3 and 0 for serious events, 3 and 3 for other events, and 0 deaths.

For Phase 1: Non-Lung Target + MK-3475 / Phase 2 Melanoma, no Prior PD-1 Therapy, the counts were 5 at risk and 1 affected for serious adverse events, 5 at risk and 5 affected for other adverse events, and 1 death. For Phase 1: Non-Lung Target / Phase 2: Melanoma, Previously Treated With Immunotherapy/irPD, the counts were 6 and 1 for serious events, 6 and 6 for other events, and 0 deaths. For Phase 2: NSCLC With no Prior PD-1 Therapy, the counts were 35 and 13 for serious events, 35 and 35 for other events, and 3 deaths. For Phase 2: Melanoma, Previously Treated With Immunotherpay, With irPD, the counts were 2 and 0 for serious events, 2 and 2 for other events, and 0 deaths.

The listed serious adverse event terms include Gastric hemorrhage (20.0%), Nausea (16.7%), Vomiting (16.7%), Device related infection (10.0%), Confusion (10.0%), and Pneumonitis (10.0%). The denominator and method used for these percentages are not reported in the registry record. Counts of affected participants are not necessarily counts of individual events, and these cohort records do not establish causality or comparative safety.

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

The registry record establishes that a prospective Phase I/II study was conducted to examine SBRT in combination with MK-3475 in metastatic melanoma and NSCLC. It describes a dose-escalation phase and expansion cohorts, a treatment sequence involving progression and radiation to a selected lesion, and a Phase 2 response endpoint among participants who received SBRT. It also provides cohort-level adverse event counts.

It does not report the primary response category counts. As a result, the record does not show how many participants had complete response, partial response, stable disease, progressive disease, or were not evaluable. It also does not report an estimate of response with a precision measure. No conclusion about the size, certainty, or statistical significance of a response can be drawn from the endpoint description alone.

The design does not establish a randomized comparison of SBRT plus MK-3475 against MK-3475 alone or another control. The treatment sequence and cohort structure mean that outcomes, even if fully reported, would be descriptive of the study population and treatment course. They would not isolate the effect of SBRT without a suitable comparison and analysis. Differences between melanoma and NSCLC cohorts also cannot be treated as direct treatment contrasts because the populations and treatment histories differ.

The safety entries describe counts recorded in the registry, not proof that an event was caused by a study intervention. The record does not provide comparative safety estimates or details sufficient to establish how the selected adverse event percentages were calculated. The deaths recorded by cohort should likewise be read as descriptive counts, not as evidence of cause.

In general, a trial record can document what investigators intended to measure, how cohorts were organized, and which outcomes were posted. Those elements do not replace outcome results or a transparent statistical analysis. For this trial, the registry record supports a description of the design and the reported safety entries, while leaving the primary response result and its statistical interpretation unreported.

13. Why This Trial Matters Statistically

This trial raises a useful statistical question about treatment sequences rather than a simple comparison of two assigned therapies. Participants with NSCLC began MK-3475 and proceeded to SBRT after progression, while the melanoma cohort was enrolled after progression on anti-PD-1 therapy. The intended response assessment followed radiation and renewed MK-3475. The timing and treatment history are therefore essential to defining what the endpoint would represent.

The design also illustrates the distinction between a dose-escalation study and an expansion cohort. Dose escalation is intended to explore treatment dose levels and identify a dose for further study. Expansion cohorts then describe outcomes in selected populations at that dose. These purposes are not the same as estimating a randomized treatment contrast. The registry record does not report the dose-selection outcome or the analysis plan in enough detail to evaluate those decisions statistically.

Another important issue is the endpoint population. The primary response measure applies only to people who received SBRT, not automatically to every person enrolled. Interpreting a response count requires knowing who received treatment, who had an assessment, and how unevaluable or missing outcomes were handled. Those details are not reported in the registry record, and the response counts are absent.

Finally, the record demonstrates why design information and result information must be considered separately. A clear endpoint definition does not reveal the observed outcome; a completed trial status does not provide an effect estimate; and adverse event counts do not answer the efficacy question. The nonrandomized structure also limits causal interpretation. These distinctions are broadly relevant when reading early-phase and cohort-based oncology studies, particularly when populations differ by cancer type and prior treatment.

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 most direct limitation for statistical interpretation is that the registry record does not report the primary response counts. Without those counts, no response distribution or effect estimate can be described. No confidence interval, p-value, or other inferential result for the primary endpoint is reported either.

The study was nonrandomized and unmasked, and the registry record does not describe a concurrent control group for the primary response question. The design therefore does not support a direct randomized comparison of the combined treatment sequence against an alternative. Cohorts differ in cancer type, prior treatment, and study phase, so cross-cohort contrasts would be difficult to interpret as treatment effects.

The primary endpoint applies to participants who received SBRT. The number who received SBRT, the number assessed for response, and the handling of missing or unevaluable assessments are not reported in the registry record. Although not evaluable is an endpoint category, its count and definition are absent. This leaves uncertainty about the endpoint denominator and the analysis population.

The record lists two outcome measures, but the other outcome measure’s details are not reported. Multiplicity procedures, statistical methods for the primary endpoint, and the analysis plan are not reported in the registry record. The record also does not report the sample-size justification, interim analyses, or the rationale for the planned enrollment.

The safety data are descriptive cohort counts. The registry record does not provide the denominator or calculation details for percentages attached to selected serious adverse event terms, and the entries do not establish causality. In addition, some cohort descriptions combine Phase I and Phase 2 labels, making it important not to assume that all participants within a named group had identical treatment histories or study roles beyond what the record states.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Evaluating the Combination of MK-3475 and Sterotactic Body Radiotherapy in Patients With Metastatic Melanoma or NSCLC — NCT02407171.
• Registry record last updated 2023-10-31; status last verified 2023-10.
• Results first posted 2023-10-31.
• 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

NCT02407171 is a completed, nonrandomized, unmasked Phase I/II trial sponsored by Yale University. The record lists enrollment of 61 and a parallel model. The recorded start date is 2015-04-01, and the primary completion date is 2019-02-19.

The study examined MK-3475 with SBRT in metastatic melanoma and NSCLC. It included Phase I radiation dose escalation and Phase 2a expansion cohorts. In the described treatment sequence, a single target lesion was treated with SBRT after progression, followed by restarting MK-3475.

The Phase 2 primary endpoint was response-category counts among participants who received SBRT, assessed using RECIST over up to 12 months. The response counts and statistical analysis are not reported in the registry record. Cohort-level adverse event counts and selected serious adverse event terms are recorded, but these are descriptive and do not provide a controlled comparison.