Adjuvant Nivolumab After Salvage Resection in Head and Neck Cancer Patients Previously Treated With Definitive Therapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Adjuvant Nivolumab After Salvage Resection in Head and Neck Cancer Patients Previously Treated With Definitive Therapy (NCT03355560), a na phase 2 study. The primary outcome is Percentage of Patients With Grade 3 and 4 Adverse Events of Nivolumab.

2. Clinical Question

The clinical question is primarily about safety in a defined post-surgical setting. It concerns patients who had already received definitive therapy and subsequently underwent salvage surgery, rather than all patients with head and neck cancer. The registry’s stated purpose also includes testing efficacy, and it lists disease-free survival among secondary outcomes. However, the primary outcome is the count of participants with a higher-grade adverse event, not a measure of disease control.

Because this is a single-group study, the main descriptive question is what events were recorded among participants in the nivolumab group. The study does not include a concurrent group that would show what event counts might have occurred under a different treatment strategy. Any interpretation of the observed safety experience must therefore stay within that descriptive scope.

3. Trial Design

The registry classifies NCT03355560 as a Phase 2, single-group study of adjuvant nivolumab after salvage resection in patients with head and neck squamous cell carcinoma previously treated with definitive therapy. It records enrollment of 39 and lists nivolumab as the intervention. The study is marked completed. The registry dates are 2017-12-06 for the start and 2024-01-01 for primary completion.

The planned treatment was nivolumab starting 4-11 weeks after surgery for 6 doses. This describes the regimen in the record; it does not establish that every enrolled participant started treatment, received every dose, or remained under observation for the same period. Those details are not reported in the registry record summarized for this analysis.

There was no random allocation, masking, or control group. The record identifies allocation as not applicable, masking as none, and the model as single group. As a result, comparisons between nivolumab and another treatment are not part of this design. The observed event counts describe participants in the intervention group and cannot isolate the effect of nivolumab from the effects of surgery, prior treatment, underlying illness, or other factors.

The primary outcome is the number of participants with at least one Grade 3 or Grade 4 adverse event during nivolumab treatment. The record says adverse events were collected during treatment and for a minimum of 30 days after the last dose, up to 7 months. Secondary outcomes include any-grade adverse events and disease-free survival; disease-free survival appears twice in the registry’s secondary-outcome list. The record lists 4 outcome measures and 0 analyses.

The registry does not report a statistical analysis plan, detailed eligibility criteria, or the number of participants who completed treatment or follow-up. Those details should not be inferred from the single-group label or the planned regimen.

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

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: 39 subjects that received at least one dose of Nivolumab

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Patients With Grade 3 and 4 Adverse Events of Nivolumab, measured over AEs were collected during the treatment period and for a minimum of 30 days following the last dose of study treatment, up to 7 months..

The primary endpoint is the count of participants who experienced at least one Grade 3 or Grade 4 adverse event while receiving nivolumab. The registry says events were assessed using CTCAE v5.0. It describes the endpoint as a participant count, so the intended unit is people with an event rather than the total number of separate event episodes. A participant with multiple qualifying events would still meet the endpoint definition as a participant with at least one such event.

The collection window covered the treatment period and a minimum of 30 days following the last dose, up to 7 months. The registry does not explain how follow-up varied across participants or how events after the stated window were handled. The primary outcome is a safety measure: it captures whether participants experienced higher-grade toxicity during the defined observation period. It does not by itself describe the cause of an event, its duration, whether it resolved, or whether it was judged related to nivolumab.

Using a participant-level count can make the endpoint easier to interpret than a count of event episodes when people may experience more than one event. It still depends on consistent event assessment, severity grading, and follow-up. No result for this primary endpoint is reported in the registry record, so the count and its precision cannot be described from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation starts by matching each result to the question and endpoint it was designed to address. Here, the primary endpoint is a participant count: whether a participant had at least one Grade 3 or Grade 4 adverse event during treatment. A count of participants is not interchangeable with the number of individual event episodes. The registry also reports separate adverse-event summaries, including serious adverse events and other adverse events. Their categories should be kept distinct unless the record explicitly defines them as overlapping or equivalent; it does not do so.

What does a single-group design allow?

The study has one treatment group and no randomized or concurrent comparator. Its event counts describe what was recorded among participants in that group. They do not estimate a treatment contrast. In a comparative trial, a difference between groups can help quantify how outcomes varied between assigned strategies, subject to the design and analysis. A single-group study has no such internal contrast, so it cannot distinguish the treatment’s contribution from other explanations for observed outcomes. This is a general design principle, not a claim about the cause of any event in this study.

The lack of a comparator also matters for efficacy. An observed disease-free survival outcome, if reported, would describe participants’ experience over follow-up. Without a concurrent comparison group, it would not by itself show what would have happened under an alternative strategy. The registry does not report disease-free survival estimates or analyses, so even that descriptive result cannot be assessed from this record.

How should counts and percentages be read?

A count gives the number of participants recorded with a specified event. A percentage expresses a count relative to a denominator, but interpretation depends on knowing what the denominator represents and how missing observations were handled. The adverse-event summary lists 9 participants affected by serious adverse events among 39 at risk and 37 affected by other adverse events among 39 at risk. It separately reports 10 deaths. These summaries should be read as descriptive data for the listed group, not as comparative effects or proof of causation.

Percentages listed for serious adverse-event terms are also descriptive. The record names respiratory failure at 5.1%, abdominal wall cellulitis at 2.6%, acute kidney injury at 2.6%, aspiration at 2.6%, bone infection at 2.6%, and cardiac arrest at 2.6%. The percentages do not supply a complete account of the serious-event profile, and the record does not specify term-level counts or the denominator for those percentages. They should not be converted into counts or used to infer how many participants had multiple events.

What would a confidence interval tell the reader?

General explanation: a confidence interval expresses the uncertainty around an estimated quantity under the statistical model used. A narrower interval indicates greater precision under that model; a wider interval indicates less precision. For a participant-level event proportion, an interval can show how uncertain the underlying event frequency may be given the observed data. It does not remove possible bias, establish causation, or compensate for the absence of a control group. No confidence interval for the primary endpoint is reported in the registry record.

General explanation: statistical significance and clinical importance are different ideas. A p-value, when reported, summarizes how compatible data are with a specified statistical model or null hypothesis. It is not the probability that a treatment works, and it does not convey the size or precision of an effect on its own. The registry record lists 0 analyses, and no p-value or inferential result for the primary endpoint is reported. There is therefore no basis in this record for a claim of statistical significance.

Why can event data be difficult to summarize?

General explanation: adverse-event records may include participants with no event, one event, or multiple events. Analyses must distinguish participant-level incidence from event-level counts and specify the observation period. They must also state how events are graded, whether seriousness is assessed separately from severity, and how attribution to treatment is determined. Seriousness is a regulatory classification, while severity describes the intensity of an event; the terms are not interchangeable. The registry identifies the primary endpoint’s grading approach as CTCAE v5.0, but does not provide a crosswalk between that endpoint and the separate serious adverse-event summary.

General explanation: over-dispersion matters when event counts vary more than a model assumes. A model that assumes too little variability can produce uncertainty estimates that are too small. This is relevant in some settings involving repeated events or heterogeneous event counts, but the registry does not report a model for repeated adverse events. Because the primary endpoint is defined as a count of participants with at least one event, event-level modeling should not be assumed.

How should planning calculations be distinguished from trial results?

General explanation: sample-size calculations are prospective planning exercises based on assumptions about the effect to detect, variability, allocation, significance level, and analysis method. They do not describe what the trial observed and are not observed power. The planning assumptions associated with the posted calculations specify a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They also describe effects as a stated fraction of the observed effect; time-to-event calculations use a total event count under the Schoenfeld approach, while other calculations are expressed as participants per arm.

Those planning assumptions do not match this study’s recorded single-group design, which has no two-arm allocation. They should not be treated as the protocol’s design or as evidence that the study was planned or analyzed under those assumptions. The registry record does not report the trial’s sample-size justification, a statistical analysis plan, or interim analyses. It also does not report the primary endpoint estimate needed to assess a planning illustration based on an observed effect.

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("Nivolumab")
n      <- c(39)
events <- c(17)
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  Nivolumab  39  17  43.6  27.8  60.4  29.3  59.0  29.3  59.1

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 4 outcome measures, including a primary safety endpoint, an any-grade adverse-event outcome, and disease-free survival, with disease-free survival appearing twice in the secondary-outcome list. The record lists 0 analyses. It does not report a multiplicity strategy, such as adjustment for testing several outcomes, a hierarchy for interpreting outcomes, or a distinction between confirmatory and exploratory analyses. Without such information, the set of outcomes should not be treated as a collection of independent confirmatory tests.

Missing-data methods are not reported in the registry record. The record does not explain how participants with incomplete adverse-event observation, missing severity grades, or incomplete follow-up were handled. It also does not give the number of participants who began treatment, completed the planned doses, or had complete outcome assessment. The 39 participants listed at risk in the adverse-event summary do not establish that all enrolled participants had the same treatment exposure or complete follow-up.

Analysis populations are not defined in the available record. There is no reported intention-to-treat, treated, or evaluable population definition, and no account of exclusions from particular outcome analyses. The planned treatment window and adverse-event collection period are stated, but the handling of deviations from those periods is not reported.

These omissions limit assessment of how complete and comparable the outcome observations were. They also mean that the reported adverse-event counts should be interpreted as registry summaries, without assumptions about imputation, censoring, or participant inclusion rules that are not documented.

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: respiratory failure (5.1%), abdominal wall cellulitis (2.6%), acute kidney injury (2.6%), aspiration (2.6%), bone infection (2.6%).

The registry’s adverse-event summary reports 9 participants affected by serious adverse events among 39 at risk, 37 participants affected by other adverse events among 39 at risk, and 10 deaths. These are separate descriptive entries. The record does not explain whether the serious and other adverse-event categories overlap, whether deaths occurred during treatment or follow-up, or whether any death was attributed to nivolumab. The figures should not be used to infer those details.

The primary endpoint is specifically the number of participants with at least one Grade 3 or Grade 4 adverse event during treatment. That outcome is not reported in the registry record. Serious adverse events are not automatically the same as Grade 3 or Grade 4 events: seriousness and severity are distinct classifications. The summary of serious events therefore cannot substitute for the missing primary endpoint result.

The registry lists respiratory failure at 5.1%, abdominal wall cellulitis at 2.6%, acute kidney injury at 2.6%, aspiration at 2.6%, bone infection at 2.6%, and cardiac arrest at 2.6% among serious adverse-event terms. These percentages are term-level summaries; no corresponding counts, denominator definition, or participant-level overlap information is reported. They do not indicate how many people experienced multiple terms or establish treatment causality.

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

The record describes a Phase 2 single-group study of nivolumab after salvage resection and supplies some descriptive adverse-event information. It establishes the primary safety question and its intended observation period, but does not report the number of participants who met the Grade 3 or Grade 4 adverse-event definition. Without that result, the primary endpoint cannot be evaluated quantitatively from the registry record.

The serious adverse-event and other adverse-event summaries give information about reported events in the nivolumab group. They do not by themselves show whether nivolumab caused those events, whether the observed counts are higher or lower than under another strategy, or how the event experience compares with that of similar patients who did not receive nivolumab after surgery. There is no control group to answer those comparative questions.

The record’s stated purpose includes safety and efficacy, and disease-free survival is listed as a secondary outcome. However, the record does not provide disease-free survival results, the primary endpoint result, or analyses for the outcomes. A completed status indicates the study is marked completed in the registry; it does not establish that a particular safety or efficacy result was favorable, unfavorable, statistically significant, or clinically important.

The appropriate scope of interpretation is therefore limited: the registry describes the treatment plan, endpoint definition, and selected event summaries, but does not provide the analyses needed to assess the primary endpoint or efficacy. No conclusion about benefit, harm relative to another treatment, or causation can be drawn from these entries alone.

14. Why This Trial Matters Statistically

This study raises a statistically important question in a specific setting: how to describe safety after salvage resection among patients previously treated with definitive therapy. The primary endpoint is participant-based, which focuses on whether a person experienced at least one higher-grade adverse event rather than on how many event episodes accumulated. Clear definitions of the event grade, treatment exposure, and observation window are essential for interpreting such a measure.

The registry also illustrates why endpoint labels and safety summaries must be kept separate. Its primary endpoint concerns Grade 3 or Grade 4 events, while the adverse-event summary separately reports serious adverse events, other adverse events, and deaths. Those measures answer related but non-identical questions. Treating them as interchangeable could misstate what the record actually reports.

Design is equally important. A single-group study can describe outcomes among participants who received the intervention, but it cannot provide a randomized comparison with an alternative strategy. That constraint matters when interpreting adverse events in a population with prior treatment and salvage surgery, because the record cannot separate the contribution of those circumstances from that of the study treatment.

Finally, a trial’s completed status and its stated goals should not be mistaken for reported evidence. The registry lists a primary safety endpoint and secondary outcomes, but it does not report the primary endpoint result or analyses. Transparent reporting of outcome estimates, uncertainty, analysis populations, and missing-data handling would be needed for a fuller statistical assessment.

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 main limitation for interpreting this specific record is that it does not report the result for the primary endpoint: the number of participants with at least one Grade 3 or Grade 4 adverse event during nivolumab treatment. It also provides no confidence interval or statistical analysis for that endpoint. The reported serious adverse-event count cannot fill this gap because seriousness is not the same classification as event grade.

The single-group design limits comparative interpretation. There is no randomized control arm, so the record cannot estimate a treatment contrast or establish that reported events were caused by nivolumab. The registry gives no disease-free survival estimates or analyses, despite listing disease-free survival as a secondary outcome. It lists disease-free survival twice among the secondary outcomes, without explaining the duplication.

Important details for evaluating the summaries are not reported in the registry record: analysis-population definitions, missing-data procedures, participant-level follow-up completeness, treatment initiation and completion, and the basis for the term-level serious adverse-event percentages. The record also does not provide a statistical analysis plan, multiplicity approach, or sample-size justification. These omissions make it difficult to assess precision, selection into analyses, or the handling of incomplete observation.

The registry records enrollment of 39 and 39 at risk for the adverse-event summaries, but does not establish from these entries that all participants had the same exposure or complete outcome ascertainment. The planned regimen and collection window should therefore be read as protocol descriptions, not as proof that each participant followed them.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Adjuvant Nivolumab After Salvage Resection in Head and Neck Cancer Patients Previously Treated With Definitive Therapy — NCT03355560.
• Registry record last updated 2026-05-06; status last verified 2026-04.
• Results first posted 2026-05-06.
• 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

NCT03355560 is recorded as a completed Phase 2, single-group study of adjuvant nivolumab after salvage resection in head and neck squamous cell carcinoma previously treated with definitive therapy. The registry lists enrollment of 39 and a planned regimen starting 4-11 weeks after surgery for 6 doses.

The primary endpoint is the count of participants with at least one Grade 3 or Grade 4 adverse event during treatment, assessed using CTCAE v5.0 and collected during treatment and for a minimum of 30 days following the last dose, up to 7 months. The primary endpoint result is not reported in the registry record.

The adverse-event summary reports 9 participants affected by serious adverse events, 37 affected by other adverse events, and 10 deaths. The record lists 0 analyses and does not report disease-free survival results, a statistical analysis plan, missing-data methods, or analysis populations. Because there is no control group, the reported event summaries are descriptive rather than comparative.