Pembrolizumab (MK-3475) Versus Placebo Following Surgery and Radiation in Participants With Locally Advanced Cutaneous Squamous Cell Carcinoma (MK-3475-630/KEYNOTE-630): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pembrolizumab (MK-3475) Versus Placebo Following Surgery and Radiation in Participants With Locally Advanced Cutaneous Squamous Cell Carcinoma (MK-3475-630/KEYNOTE-630) (NCT03833167), a randomized phase 3 study. The primary outcome is Recurrence-Free Survival (RFS) as Assessed by the Investigator and Confirmed by Biopsy.

2. Clinical Question

The clinical question is whether the posted time-to-event results favor pembrolizumab over placebo in this post-surgery and post-radiation setting. The registry record reports a hazard ratio for the comparison and median recurrence-free survival in each randomized group. Those statistics describe the observed trial results; they do not by themselves establish the size or certainty of a treatment effect for every participant or setting.

3. Trial Design

This was a Phase 3, randomized, double-blind, parallel-group, placebo-controlled study. The registry record lists enrollment of 450 and reports 225 participants in each randomized group. Participants were assigned to pembrolizumab or placebo after surgery and radiotherapy for high-risk locally advanced cutaneous squamous cell carcinoma.

The planned treatment was intravenous infusion on Day 1 of each 42-day cycle, for up to 9 cycles. The pembrolizumab arm description specifies 400 mg per infusion. The registry also describes possible pembrolizumab treatment after biopsy-proven recurrence for some participants, including an open-label option after recurrence in the placebo group. The details of eligibility for retreatment are not fully reported in the registry record.

The primary endpoint was investigator-assessed recurrence-free survival confirmed by biopsy, analyzed in the group to which each participant was randomized. The stated follow-up timeframe was up to approximately 62 months. The primary completion date is 2024-06-28, and the study status is terminated. The reason for termination is not reported in the registry record.

Secondary outcomes listed include overall survival, quality-of-life measures, adverse events, and discontinuation due to adverse events. The registry record does not report the original sample-size justification or the assumptions used to design the trial. It also does not report details of the allocation procedure beyond randomized assignment, such as the method used to generate the allocation sequence.

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: All randomized participants

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 Recurrence-Free Survival (RFS) as Assessed by the Investigator and Confirmed by Biopsy, measured over Up to approximately 62 months.

Recurrence-free survival measures the time from randomization until the first qualifying event. In this study, the qualifying events were local or regional recurrence of the index lesion, distant metastasis, or death from any cause, whichever occurred first. The registry describes assessment by the investigator, with recurrence confirmed by biopsy.

A time-to-event endpoint retains information about when events occur, rather than reducing each participant’s course to a simple yes-or-no event count at one fixed time. This can be useful when participants have different lengths of follow-up. The analysis compares the event experience over follow-up between the randomized groups and is summarized by a hazard ratio, alongside group medians and confidence intervals.

Recurrence-free survival is relevant to an adjuvant study because the treatment is given after surgery and radiotherapy, when the primary focus is whether and when recurrence, metastasis, or death occurs. Including death from any cause means that death is counted as an event even when a cancer recurrence has not been recorded. The registry record names all randomized participants as the primary analysis population and states that participants were analyzed in their assigned treatment group.

The posted median estimates are 53.3 months for pembrolizumab and 53.7 months for placebo. The registry record reports lower confidence limits of 51.3 and 51.8 months, respectively, but does not report the upper confidence limits for these medians.

7. Results Posted in the Registry

• Pembrolizumab vs Placebo: hazard ratio 0.76 (95% CI 0.53–1.10), reported p = 0.072 (as reported in the registry).

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Overall Survival (OS)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Pembrolizumab vs Placebo: hazard ratio 1.47 (95% CI 0.87–2.48) (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Pembrolizumab vs Placebo and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists secondary outcomes spanning overall survival, quality of life, adverse events, and discontinuation due to adverse events. Those outcomes address different aspects of follow-up: survival beyond recurrence, participants’ reported health and functioning, and treatment tolerability. However, results for most listed secondary outcomes are not reported in the registry record. The overall pattern across those outcomes therefore cannot be characterized from the posted results.

An overall-survival comparison is posted, but the registry does not report median overall survival estimates or a p-value. The hazard-ratio estimate and interval are available in the outcome results, but they should be interpreted as a secondary finding rather than as a substitute for the primary recurrence-free survival analysis. Other secondary outcome results, including the listed quality-of-life and adverse-event outcomes, are not reported in the registry record in a form that permits a complete synthesis.

Secondary results can help describe whether findings are consistent across different outcomes, but they do not automatically confirm or overturn the primary result. Their interpretation depends on the prespecified testing strategy, the completeness of reporting, and the number of comparisons considered. When many outcomes are tested, chance findings become more likely unless the analysis accounts for multiplicity. The registry record does not report whether secondary comparisons were adjusted for multiplicity or how they were prioritized. Any unadjusted secondary findings should consequently carry less weight than a clearly prespecified primary analysis, especially when the secondary results are incomplete.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Overall Survival (OS)
• Change From Baseline in European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30) Global Health Status (Item 29) and Quality of Life (Item 30) Combined Score
• Change From Baseline in Physical Functioning Using the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) Items 1-5 Score
• Percentage of Participants Who Experience an Adverse Event (AE)
• Percentage of Participants Who Discontinue Study Treatment Due to an Adverse Event (AE)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary result is a time-to-event comparison. The registry reports a hazard ratio and a Log Rank analysis, as well as median recurrence-free survival estimates for each randomized group. These statistics answer related but different questions: the medians describe the time point at which an estimated group’s event-free experience reaches its midpoint, while the hazard ratio summarizes relative event rates across follow-up under the analysis model.

What does a hazard ratio describe?

In general, a hazard ratio compares the event hazard between groups over time. It is not a ratio of the number of participants who experienced an event, and it is not a direct comparison of median survival times. A hazard ratio of 1 represents equal hazards under the model; a value below 1 favors the first named group for an endpoint where events are undesirable, while a value above 1 favors the comparator. The reported estimate for pembrolizumab versus placebo is 0.76. That estimate alone does not capture its uncertainty, so it must be interpreted alongside the confidence interval.

The hazard-ratio interpretation is most straightforward when the relative event pattern between groups is reasonably stable over time. If that assumption does not hold, a single hazard ratio may not represent the full pattern of group differences. The registry record does not report a proportional-hazards assessment or other diagnostics, so that aspect of the analysis cannot be evaluated from the posted record.

What does the Log Rank method do?

In general, the Log Rank test compares observed and expected event experience between groups across follow-up, taking account of the timing of events and censored observations. It is a common approach for randomized time-to-event comparisons. The registry identifies Log Rank as the method for the reported primary comparison. It does not give further details about the test implementation, stratification, or any covariate adjustment.

A Log Rank p-value addresses how compatible the observed group separation is with a no-difference hypothesis under the test’s assumptions. The posted p-value is 0.072. The registry’s planning illustration specifies a two-sided significance level of 0.05, but the original trial’s significance threshold and decision rules are not reported in the registry record. The planning assumption should not be treated as proof that it was the trial’s prespecified inferential rule.

Why report medians as well?

A median time-to-event estimate gives a recognizable time summary for each group, but it does not show the entire distribution of event times or the shape of the survival curves. In this study, the median recurrence-free survival estimates are 53.3 months for pembrolizumab and 53.7 months for placebo. The registry lists 95% confidence intervals but only the lower limits, 51.3 and 51.8 months. The upper limits are not reported in the registry record. The medians and hazard ratio should be considered together rather than treated as interchangeable summaries.

How should the confidence interval and p-value be read together?

In general, a confidence interval describes the range of effect estimates reasonably compatible with the data and the statistical model at the stated confidence level. A 95% confidence interval is not a probability statement that the particular interval contains the true value after it has been calculated. Here, the reported hazard-ratio interval is 0.53 to 1.10. Its inclusion of 1 indicates that the data are compatible with no relative hazard difference as well as with effects in either direction within the interval. The point estimate is only one value within this uncertainty range.

The p-value and confidence interval are related but convey different aspects of the result. The p-value summarizes compatibility with a specified null model; the interval shows uncertainty around the estimated effect and the range of effect sizes supported by the data and model. Neither quantity addresses the importance of a potential effect by itself. Clinical relevance requires context that cannot be established from the hazard ratio alone.

What does the analysis population contribute?

The registry states that the primary endpoint analysis included all randomized participants and that participants were analyzed in the treatment group to which they were assigned. This preserves the randomized comparison and avoids redefining the primary groups according to treatment received after randomization. The arm counts listed for the primary result are 225 in each group. The adverse-event record lists 224 participants at risk in each main treatment group; the registry record does not explain the difference between those safety denominators and the randomized group counts.

The record does not report how incomplete follow-up, censoring, or missing endpoint assessments were handled. It also does not report an interim-analysis plan, stopping boundaries, the original sample-size calculation, or whether the primary analysis was stratified or adjusted. Those details can affect interpretation of a time-to-event analysis and should not be inferred from the posted summary.

11. Confidence Intervals and Precision

The reported hazard ratio for recurrence-free survival is 0.76, with a 95% confidence interval from 0.53 to 1.10. This interval gives a sense of the statistical precision: the point estimate favors pembrolizumab, but the plausible values under the analysis include both a lower hazard and a value above 1. The interval therefore does not rule out no relative difference or a result in the opposite direction. The p-value of 0.072 likewise indicates that the result is not sharply separated from the no-difference model under the reported analysis.

Confidence intervals are useful because a point estimate alone can make an uncertain result appear more definite than it is. In general, the interval reflects both the amount of information in the data and the variability expected under the statistical model. A narrower interval would indicate greater precision, while a wider interval would indicate less precision. Precision is not the same as freedom from bias, and a confidence interval does not address whether the analysis assumptions are appropriate.

The median estimates are 53.3 months for pembrolizumab and 53.7 months for placebo. Their reported lower confidence limits are 51.3 months and 51.8 months. The upper confidence limits for the medians are not reported in the registry record, so the uncertainty around those median estimates cannot be fully assessed from the posted summaries.

The registry describes follow-up up to approximately 62 months. That timeframe and the listed arm sizes do not, by themselves, establish how much event information contributed to the analysis. The number of primary events and the detailed follow-up distribution are not reported in the registry record. Consequently, precision is best judged from the reported hazard-ratio interval and the incomplete median intervals, with the limitations of the available reporting kept in view.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size illustration is a planning exercise based on the observed effect, not a report of the trial’s original design assumptions. For the time-to-event endpoint, the figures are total event counts needed across the study, not participants per arm. The illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation, with the effect to detect set to the stated fraction of the observed effect. The original trial’s own sample-size assumptions are not reported in the registry record.

The 80% power column describes the event target for a hypothetical study designed to have an 80% chance of detecting the specified effect under those assumptions, if that effect is truly present. The 90% power column uses a higher chance of detection and therefore gives a larger event target for the same effect assumption. For example, at 100% of the observed effect, the illustration lists 417 total events for 80% power and 559 for 90% power. At 75% of the observed effect, it lists 742 and 993 total events, respectively; at 50%, it lists 1,668 and 2,233.

Smaller effects are harder to distinguish from random variation, so the illustration requires more events as the assumed effect gets smaller. These figures are not participants per arm, because the time-to-event calculation is expressed in total events. They are not observed power; they are planning illustrations based on the observed effect and stated assumptions. The results should not be read as a claim about the trial’s design adequacy or its original target.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.760, so ln(HR) = -0.2744.
2. d = 4 × (1.96 + 0.842)² / [ln(0.760)]² = 416.9, rounded up to 417 events in total for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
hr <- 0.76                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(3833167)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 417 (80% power), 559 (90% power)
 75% of observed effect: events = 742 (80% power), 993 (90% power)
 50% of observed effect: events = 1,668 (80% power), 2,233 (90% power)
Simulated power with 417 events: 0.793 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 1,666 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 450 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 53.3          # control-arm median, months (converted from the posted unit)
hr          <- 0.76          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
    Analysis               Value Efficacy
   IA 1: 33%                   Z   3.7103
     N: 1642         p (1-sided)   0.0001
 Events: 189        ~HR at bound   0.5826
   Month: 24    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.76   0.0338
   IA 2: 67%                   Z   2.5114
     N: 1666         p (1-sided)   0.0060
 Events: 378        ~HR at bound   0.7721
   Month: 38    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.76   0.5603
       Final                   Z   1.9930
     N: 1666         p (1-sided)   0.0231
 Events: 566        ~HR at bound   0.8457
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.76   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

The sponsor-reported primary analysis gives a hazard ratio of 0.76, a 95% confidence interval of 0.53 to 1.10, and p = 0.072, using the Log Rank method. The comparison summary lists the same hazard-ratio estimate and interval. No discrepancy is flagged between the reported analysis and the comparison summary, so the sponsor’s reported analysis is the appropriate result to interpret rather than treating a separate unadjusted recalculation as a replacement.

The reported analysis is a time-to-event comparison and should be read as a whole: the hazard-ratio estimate, its confidence interval, the p-value, and the group median estimates provide complementary information. The point estimate favors pembrolizumab, while the confidence interval includes 1. The p-value is 0.072. The medians are 53.3 months for pembrolizumab and 53.7 months for placebo, with upper confidence limits not reported in the registry record.

The registry record does not provide the full statistical analysis plan or details such as stratification, covariate adjustment, or censoring rules. It is therefore not possible to reconstruct the sponsor’s analysis completely from the posted summary or to determine whether the reported result followed a particular testing hierarchy. The registry’s planning illustration is separate from the sponsor’s reported efficacy analysis and should not be used to reinterpret it.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates more than one endpoint, comparison, subgroup, or analysis time. In general, each additional statistical test creates another opportunity for a chance finding, so a prespecified strategy may control the overall error rate, order tests hierarchically, or identify analyses as exploratory. The registry lists multiple secondary outcomes as well as the primary recurrence-free survival endpoint, but the multiplicity strategy is not reported in the registry record.

The posted primary analysis has a Log Rank method and a p-value of 0.072. The registry does not state whether the primary test was part of a hierarchical plan, whether any alpha-spending procedure was used, or whether confirmatory claims depended on preceding outcomes. A two-sided significance level of 0.05 appears in the planning illustration, but that illustration is not evidence of the trial’s original testing plan. The registry record also does not report an interim-analysis plan or stopping rules.

Missing-data and censoring details are important for time-to-event outcomes because participants may have different follow-up and not every event will necessarily be observed. The primary definition and population are reported, but the registry record does not describe censoring rules, procedures for incomplete assessments, or sensitivity analyses for missing outcomes. It also does not provide enough detail to judge whether follow-up was complete or balanced between groups.

The primary analysis population is described as all randomized participants, with each analyzed in the group to which they were assigned. That statement is useful, but it does not settle all questions about missingness or post-randomization events. The handling of treatment discontinuation, protocol deviations, and any data collected after recurrence is not fully reported in the registry record. These omissions limit how closely an independent reader can reproduce or audit the analysis.

17. 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: basal cell carcinoma (10.0%), hyponatraemia (6.7%), acute myocardial infarction (3.3%), myocarditis (3.3%), pericarditis (3.3%).

For the main safety groups, the registry lists 224 participants at risk in each arm. Serious adverse events were reported for 55 participants in the pembrolizumab group and 43 in the placebo group. Other adverse events were reported for 175 participants in the pembrolizumab group and 143 in the placebo group. The record also lists 34 deaths in the pembrolizumab group and 14 in the placebo group. These are posted counts, not adjusted comparisons, and they should not be interpreted as rates or as evidence of causation.

The registry separately lists participants in a pembrolizumab retreatment phase and a placebo crossover group. Those groups reflect treatment after the original randomized comparison and should not be combined with the primary randomized-arm safety summaries as though they represented the same comparison. The listed serious adverse-event terms include basal cell carcinoma, hyponatraemia, acute myocardial infarction, myocarditis, pericarditis, and melaena, with percentages reported for those terms in the safety results.

The registry record does not provide enough detail to evaluate event timing, exposure-adjusted rates, attribution, or statistical comparisons between groups. The safety counts describe what was posted; they do not establish that treatment caused a particular event or that observed counts reflect comparative risk.

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

The primary result is an estimated hazard ratio of 0.76 for recurrence-free survival with pembrolizumab versus placebo. Its 95% confidence interval, 0.53 to 1.10, includes 1, and the reported p-value is 0.072. The estimate points toward a lower event hazard in the pembrolizumab group, but the uncertainty interval does not exclude no relative difference or a result in the opposite direction. The group medians, 53.3 months and 53.7 months, are also close, and the registry record does not report their upper confidence limits.

This result does not establish that pembrolizumab was superior for the primary endpoint. It also does not establish equivalence or prove that the treatments have identical effects. A failure to show a clearly different result is not the same as evidence that meaningful differences have been ruled out; that would require an appropriate design and analysis, which are not reported as an equivalence or non-inferiority framework in the registry record.

The findings apply to the study population and the treatment comparison described in the registry: high-risk locally advanced cutaneous squamous cell carcinoma after surgery and radiotherapy, with assignment to pembrolizumab or placebo. They should not automatically be generalized to other diseases, treatment settings, or patient groups. The posted analysis summarizes the randomized groups and cannot determine an individual participant’s outcome.

The record is marked terminated, but the reason is not reported in the registry record. The status alone does not explain the result or establish why the study stopped. Interpretation is also constrained by missing information about the original statistical plan, censoring, missing data, multiplicity, and the full confidence intervals for the medians. The statistics describe the posted evidence and its uncertainty; they are not medical advice or a treatment recommendation.

19. Why This Trial Matters Statistically

This trial matters statistically because it uses a randomized, double-blind, placebo-controlled design to examine a time-to-event endpoint after surgery and radiotherapy. Random assignment is intended to make the groups comparable at baseline, while blinding can reduce some forms of bias in treatment delivery and outcome assessment. The primary endpoint also combines recurrence, distant metastasis, and death, so its definition is consequential for how the result is interpreted.

The reported hazard ratio points toward pembrolizumab, but the confidence interval includes 1 and the p-value is 0.072. This is a useful example of why an estimated direction should not be confused with a definitive conclusion. The confidence interval communicates uncertainty in effect size, while the medians give a separate time-based summary. Neither one alone describes the full event pattern over follow-up.

The study also illustrates the importance of distinguishing the primary outcome from secondary outcomes, and observed results from planning calculations. The sample-size illustration uses the observed effect and specified assumptions; it is not the original trial design calculation and is not observed power. The registry record’s gaps about multiplicity, missing data, and the trial’s original analysis plan limit how fully the reported result can be audited.

Finally, the terminated status and incomplete secondary reporting make careful registry-based interpretation especially important. A result should be described in terms of its estimate, uncertainty, and available methods rather than reduced to a simple success-or-failure label. The posted record supports a cautious account of the primary comparison, while several questions about design and analysis remain unreported.

20. 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 specific to this record affect interpretation. The registry status is terminated, but the reason for termination is not reported in the registry record. The primary completion date is 2024-06-28, and the stated timeframe is up to approximately 62 months; details of the actual follow-up distribution are not reported.

The primary endpoint results include median estimates and lower confidence limits, but the upper confidence limits for both medians are not reported. The registry does provide a hazard-ratio interval, but it does not report the number of primary endpoint events in the posted summary. Further time-to-event details, including censoring rules, proportional-hazards checks, stratification, or covariate adjustment, are not reported in the registry record.

The listed secondary outcomes are not accompanied by a complete set of results. The record does not report the multiplicity plan, the original sample-size justification, or an interim-analysis plan. The stated planning illustration is based on the observed effect and should not be mistaken for the original trial design assumptions. The registry also does not report enough information to assess missing outcome data or sensitivity analyses.

Safety summaries include counts for the main randomized groups and separate post-randomization groups, but the record does not provide exposure-adjusted rates, detailed event timing, or comparative statistical analyses. The difference between randomized group counts and the safety denominators is not explained in the registry record. These reporting gaps do not negate the posted findings, but they limit independent evaluation and the scope of conclusions that can be drawn.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Pembrolizumab (MK-3475) Versus Placebo Following Surgery and Radiation in Participants With Locally Advanced Cutaneous Squamous Cell Carcinoma (MK-3475-630/KEYNOTE-630) — NCT03833167.
• Registry record last updated 2026-09-28; status last verified 2026-08.
• Results first posted 2025-07-22.
• 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.

24. Record Summary

NCT03833167 is a Phase 3 randomized, double-blind, parallel-group study of pembrolizumab versus placebo as adjuvant therapy after surgery and radiotherapy for high-risk locally advanced cutaneous squamous cell carcinoma. The registry lists enrollment of 450, with 225 participants in each randomized group. The primary endpoint was investigator-assessed, biopsy-confirmed recurrence-free survival, defined to include local or regional recurrence, distant metastasis, or death from any cause.

The sponsor-reported primary hazard ratio was 0.76, with a 95% confidence interval of 0.53 to 1.10 and p = 0.072. Median recurrence-free survival was 53.3 months with pembrolizumab and 53.7 months with placebo; upper confidence limits for the medians are not reported. The study status is terminated, and the reason is not reported in the registry record.

Key details about the original design assumptions, multiplicity, missing-data handling, and parts of the secondary results are not reported in the registry record. The posted result therefore supports a description of the estimate and its uncertainty, but not a definitive claim of superiority or equivalence.