This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Study of the Safety, Tolerability, Pharmacokinetics and Efficacy of Pembrolizumab (MK-3475) in Chinese Participants With Non-Small-Cell Lung Cancer (MK-3475-032/KEYNOTE-032) (NCT02835690), a randomized phase 1 study. The primary outcome is Number of Participants Who Experienced an Adverse Event (AE).
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02835690 |
| Phase | Phase 1 |
| Condition | Non-Small-Cell Lung Cancer |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 44 |
| Arms in the primary outcome | 3 |
| Lead sponsor | Merck Sharp & Dohme LLC |
| Results posted | Yes |
2. Clinical Question
Population
Adult Chinese participants with locally advanced or metastatic non-small-cell lung cancer. The primary endpoint population comprised randomized participants who received at least one dose of study treatment.
Intervention
Pembrolizumab given intravenously in one of three regimens: 2 mg/kg, 10 mg/kg, or a 200 mg fixed dose. The regimens were administered every 3 weeks.
Comparator
The 2 mg/kg arm is the reference for the posted arm comparisons. The other arms received pembrolizumab as well, so these are comparisons among dose regimens, not comparisons against placebo or no treatment.
Primary question
How many participants in each dose arm experienced at least one adverse event during the first course of treatment, and how uncertain are the differences between arms? The registry also posts objective response rate results assessed by central radiologists under RECIST 1.1 and irRECIST.
The primary question concerns the frequency of participants with an adverse event, not whether an event was caused by pembrolizumab. The registry definition includes an untoward medical occurrence whether or not a causal relationship with study treatment was considered. It also includes worsening of a pre-existing condition when temporally associated with the sponsor's product.
The design asks whether the observed safety and response patterns differ across the listed dose regimens in this study population. Because every arm received pembrolizumab, the results do not estimate the effect of pembrolizumab versus no pembrolizumab. The posted primary outcome is a count of participants with at least one adverse event; it does not measure the number, duration, or severity of every event. The registry's response outcomes add a separate measure of tumor response, but those results do not change the definition of the primary safety question.
3. Trial Design
The registry describes an open-label, randomized phase I study with a parallel-group design and no masking. It enrolled participants with non-small-cell lung cancer and compared three pembrolizumab dose regimens: 2 mg/kg, 10 mg/kg, and a 200 mg fixed dose. The registry lists an enrollment of 44. The primary endpoint results are reported for arm populations of 14, 13, and 15, respectively; the primary endpoint description specifies that the population includes randomized participants who received at least one dose of study treatment.
Pembrolizumab was administered intravenously every 3 weeks. Cycle 1 was 28 days long, and subsequent cycles were 21 days long. The arm descriptions state that treatment could continue for up to 35 administrations. The registry's brief summary describes the study purpose as assessment of safety, tolerability, pharmacokinetics, and efficacy in adult Chinese participants with locally advanced or metastatic disease. Further details of the pharmacokinetic analyses are not reported in the registry record summarized for these results.
The study started on 2016-08-04, had primary completion on 2017-09-19, and is listed as completed. The primary adverse-event timeframe is up to approximately 13 months, through the final-analysis database cutoff dated 19-Sept-2017. The response outcomes use the same stated timeframe.
Randomization can help make treatment groups comparable on average, but it does not ensure that a small trial's arms have identical participant characteristics. The registry record does not report the randomization procedure, allocation concealment, stratification factors, or baseline balance details. As an open-label study, participants and study staff were not masked to assigned regimen; the response outcomes were assessed by central radiologists according to the outcome descriptions. These design features matter when interpreting small between-arm differences, especially when the main endpoint is an event count and the arms are modest in size.
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.
Before assignment: Of 44 participants randomized, 42 received treatment. Per protocol, response/progression or adverse events (AEs) that occurred during the second course were not counted towards efficacy outcome measures or safety outcome measures, respectively.
Overall Study
| Stage | Pembrolizumab 2 mg/kg | Pembrolizumab 10 mg/kg | Pembrolizumab 200 mg Fixed Dose |
|---|---|---|---|
| STARTED | 14 | 15 | 15 |
| Received First Course of Pembrolizumab | 14 | 13 | 15 |
| Received Second Course of Pembrolizumab | 1 | 0 | 0 |
| COMPLETED | 0 | 0 | 0 |
| NOT COMPLETED | 14 | 15 | 15 |
Reasons for not completing this period:
| Reason | Pembrolizumab 2 mg/kg | Pembrolizumab 10 mg/kg | Pembrolizumab 200 mg Fixed Dose |
|---|---|---|---|
| Death | 12 | 11 | 13 |
| Transferred to Extension Study | 2 | 2 | 1 |
| Did Not Continue on Extension Study | 0 | 0 | 1 |
| Not Treated | 0 | 2 | 0 |
Analysis population for the primary outcome, as defined in the registry: All randomized participants that received at least one dose of study treatment.
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: All randomized participants that received at least one dose of study treatment.
| Characteristic | Unit | Pembrolizumab 2 mg/kg | Pembrolizumab 10 mg/kg | Pembrolizumab 200 mg Fixed Dose | Total |
|---|---|---|---|---|---|
| Participants analysed | 14 | 13 | 15 | 42 | |
| Age, Continuous | years | 53.2 (SD 9.5) | 56.2 (SD 11.5) | 56.7 (SD 8.8) | 55.4 (SD 9.8) |
| Sex: Female, Male: Female | Participants | 5 | 4 | 6 | 15 |
| Sex: Female, Male: Male | Participants | 9 | 9 | 9 | 27 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 0 | 0 | 0 | 0 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 14 | 13 | 15 | 42 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 0 | 0 | 0 | 0 |
| Race/Ethnicity, Customized: Chinese | Participants | 14 | 13 | 15 | 42 |
6. Primary Endpoint
The registry lists the primary outcome as Number of Participants Who Experienced an Adverse Event (AE), measured over Up to ~13 months (through Final Analysis database cut-off date of 19-Sept-2017).
| Element | Registry entry |
|---|---|
| Result type | Count Of Participants |
| Measure of spread | not applicable |
| Unit | Participants |
| Endpoint classification | Binary Endpoint |
The primary endpoint is the number of participants who experienced at least one adverse event during the first course of treatment. The registry describes an adverse event broadly: an unfavorable or unintended sign, symptom, or disease temporally associated with use of a medicinal product or a protocol-specified procedure, whether or not it was considered related to the product or procedure. Worsening of a pre-existing condition can also qualify when temporally associated with the sponsor's product.
This endpoint is used to describe the proportion of participants with any event, providing a broad summary of observed safety experience in each arm. It is not a count of all event episodes: a participant with multiple events still contributes once to the endpoint. It also does not, by itself, distinguish mild events from serious events or establish causation. The registry reports serious adverse events separately, which supplies additional descriptive safety information but does not alter the primary endpoint definition.
In the posted results, all participants in the 2 mg/kg and 10 mg/kg arms experienced at least one adverse event, while one participant in the 200 mg fixed-dose arm did not. These arm-level results describe what was recorded in the endpoint population during the specified course and timeframe. They should not be read as a causal comparison of treatment-related harm: the endpoint includes events regardless of whether investigators considered them related to treatment, and the registry record does not report an adjusted causal analysis.
7. Results Posted in the Registry
- Pembrolizumab 10 mg/kg: 100.0% (13/13) versus 100.0% (14/14) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0), risk ratio 1.00 (1.00–1.00); unadjusted two-proportion z-test p = n/a.
- Pembrolizumab 200 mg Fixed Dose: 93.3% (14/15) versus 100.0% (14/14) on control; risk difference -6.7 percentage points (95% CI -19.3 to 6.0), risk ratio 0.93 (0.82–1.07); unadjusted two-proportion z-test p = 0.326.
Risk difference: Pembrolizumab 10 mg/kg
95% CI: 0.0 to 0.0 percentage points
Recalculated from the posted summary statistics
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.
| Secondary outcome | Type | Effect measure | Estimate (95% CI) | Planning size, 80% power |
|---|---|---|---|---|
| Objective Response Rate (ORR) Per RECIST 1.1 as Assessed by Central Radiologists' Review | Binary Endpoint | Risk difference | +23.1 pts (0.2 to 46.0) | n/a |
| ORR Per Immune-related RECIST (irRECIST) as Assessed by Central Radiologists' Review | Binary Endpoint | Risk difference | +23.1 pts (0.2 to 46.0) | n/a |
Objective Response Rate (ORR) Per RECIST 1.1 as Assessed by Central Radiologists' Review
- Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
- Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
- Result:
- Pembrolizumab 10 mg/kg: 23.1% (3/13) versus 0.0% (0/14) on control; risk difference 23.1 percentage points (95% CI 0.2 to 46.0); unadjusted two-proportion z-test p = 0.057.
- Pembrolizumab 200 mg Fixed Dose: 20.0% (3/15) versus 0.0% (0/14) on control; risk difference 20.0 percentage points (95% CI -0.2 to 40.2); unadjusted two-proportion z-test p = 0.077.
- Interpretation:
- The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pembrolizumab 10 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
- The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pembrolizumab 200 mg Fixed Dose and the control. The interval gives the range of effects that remain compatible with the data.
ORR Per Immune-related RECIST (irRECIST) as Assessed by Central Radiologists' Review
- Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
- Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
- Result:
- Pembrolizumab 10 mg/kg: 23.1% (3/13) versus 0.0% (0/14) on control; risk difference 23.1 percentage points (95% CI 0.2 to 46.0); unadjusted two-proportion z-test p = 0.057.
- Pembrolizumab 200 mg Fixed Dose: 20.0% (3/15) versus 0.0% (0/14) on control; risk difference 20.0 percentage points (95% CI -0.2 to 40.2); unadjusted two-proportion z-test p = 0.077.
- Interpretation:
- The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pembrolizumab 10 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
- The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pembrolizumab 200 mg Fixed Dose and the control. The interval gives the range of effects that remain compatible with the data.
The secondary results posted for objective response rate present the same arm estimates under RECIST 1.1 and irRECIST. The 2 mg/kg arm has no observed responses, while both other dose regimens have observed responses. Their confidence intervals are broad, reflecting substantial uncertainty around the arm-specific response estimates. The unadjusted comparisons are also uncertain: the reported p-values are 0.057 and 0.077. These findings can be read as descriptive evidence that responses were observed in the higher-dose arms in this study, but they do not establish a reliable dose-related difference.
These response outcomes address a different question from the primary safety endpoint. A participant can have an adverse event and also have a tumor response; one result does not negate or explain the other. The primary endpoint records whether at least one adverse event occurred, while objective response rate concerns whether a defined tumor-response criterion was met. The registry's list includes duration of response, progression-free survival, and overall survival, but numerical results for those outcomes are not reported in the registry record. No broader efficacy pattern across those outcomes can therefore be inferred.
The response comparisons carry less inferential weight because the registry reports them as unadjusted and does not report a multiplicity plan or testing hierarchy. In general, when multiple outcomes and comparisons are examined, isolated nominal results can arise by chance, and unadjusted p-values do not control the overall false-positive risk. The posted p-values should not be interpreted as definitive confirmation of a dose difference, particularly alongside broad confidence intervals and the small arm populations.
The primary safety result also offers limited evidence for distinguishing the regimens: the endpoint occurred in nearly every participant in each arm. Taken together, the posted outcomes describe observed safety and response patterns, not a definitive comparison of overall clinical benefit. The registry record does not report an integrated benefit-risk analysis or a prespecified framework for weighing these outcomes against one another.
Secondary outcomes registered for this trial, for reference (only those above were recalculated):
- Objective Response Rate (ORR) Per RECIST 1.1 as Assessed by Central Radiologists' Review
- ORR Per Immune-related RECIST (irRECIST) as Assessed by Central Radiologists' Review
- Duration of Response (DOR) Per RECIST 1.1 as Assessed by Central Radiologists' Review
- DOR Per irRECIST as Assessed by Central Radiologists' Review
- Progression-Free Survival (PFS) Per RECIST 1.1 as Assessed by Central Radiologists' Review
- PFS Per irRECIST as Assessed by Central Radiologists' Review
- Overall Survival (OS)
9. Statistical Interpretation
For the primary adverse-event outcome, the posted counts are 14 of 14 participants in the 2 mg/kg arm, 13 of 13 in the 10 mg/kg arm, and 14 of 15 in the 200 mg fixed-dose arm. Thus, the endpoint was observed for every participant in the first two arms and for all but one participant in the fixed-dose arm. This is a descriptive finding about at least one adverse event during the first course, not a judgment about event severity, causality, or overall benefit and risk.
The registry's unadjusted arm comparison for 10 mg/kg versus 2 mg/kg reports a risk difference of 0.0, with limits of 0.0 and 0.0. The comparison for the 200 mg fixed dose versus 2 mg/kg reports a risk difference of -0.067, with a 95% confidence interval from -0.193 to 0.060 and an unadjusted p-value of 0.326. The estimate for the fixed-dose comparison is consistent with a modestly lower observed event proportion, but the interval also includes no difference and values in the opposite direction. The trial's small arm sizes make the estimates imprecise, and the primary endpoint's near-universal occurrence leaves little separation among regimens in these data.
The posted objective response rate results show no responses in the 2 mg/kg arm and observed responses in the other arms under both RECIST 1.1 and irRECIST. For 10 mg/kg versus 2 mg/kg, the reported unadjusted p-value is 0.057; for the 200 mg fixed dose versus 2 mg/kg, it is 0.077. These are unadjusted comparisons, and the registry record does not report a multiplicity adjustment or a prespecified inferential hierarchy for interpreting them. They should therefore be treated as exploratory signals rather than decisive evidence of a dose difference.
There is no placebo or untreated comparator, and the primary endpoint is not a treatment-attributable event endpoint. The study can describe observed safety and response outcomes across these regimens in its enrolled population, but it cannot isolate the effect of pembrolizumab relative to no treatment. The posted results also do not establish that the regimens are equivalent: absence of a clear difference in a small study is not proof of equivalence, and the primary comparison intervals do not establish a narrow range of clinically acceptable differences.
The numerical results are specific to the registry's stated population, outcome definitions, and follow-up. The record does not report enough information to assess how missing observations, treatment discontinuation, or participant characteristics may have influenced these estimates. Interpretation should remain centered on the observed counts, the width of the uncertainty intervals where reported, and the limited scope of the randomized dose-regimen comparisons.
10. Statistical Methods Explained
The primary endpoint is binary at the participant level: each person is classified as having experienced at least one adverse event or not. The posted arm results give the number of participants with the event and the arm population used for that outcome. This permits a direct description of event proportions. For an outcome with this definition, the count is not the number of individual adverse-event episodes; recurring or multiple events in one participant do not create multiple participants in the endpoint count.
The registry reports risk differences for the primary arm comparisons, using the 2 mg/kg arm as the reference. A risk difference subtracts the event proportion in the reference group from the event proportion in the comparison group. In general, a risk difference of zero means the observed proportions are equal, a negative value means the comparison group's observed proportion is lower, and a positive value means it is higher. The estimate of 0.0 for 10 mg/kg indicates equal observed proportions. The estimate of -0.067 for the fixed-dose comparison indicates a lower observed proportion in that arm, but its confidence interval includes zero, so the data are compatible with no difference as well as with differences in either direction.
A confidence interval describes the uncertainty of an estimated quantity under the statistical method used. In general, an interval is not a guarantee that the true value lies inside it, nor is it a probability statement about the particular interval after it has been calculated. The primary comparison interval from -0.193 to 0.060 spans both a possible lower event proportion and a possible higher one in the fixed-dose arm relative to the reference. This breadth matters: the point estimate alone does not capture how much uncertainty remains.
The unadjusted p-value is a measure of how compatible the observed data are with a specified null hypothesis under a statistical model. In general, a p-value is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of an effect. The reported p-value of 0.326 for the fixed-dose primary comparison does not establish that the groups are the same. The registry reports no unadjusted p-value for the 10 mg/kg primary comparison, so none should be inferred from the risk difference or its limits.
For the response outcomes, the registry reports percentages and 95% confidence intervals for objective response rate, as well as risk-difference comparisons against the 2 mg/kg arm. The posted values are the same under RECIST 1.1 and irRECIST. In general, response rate estimates the proportion of participants meeting the defined response criterion; it does not describe how long a response lasted. Duration of response is separately listed as a secondary outcome, but results for that outcome are not reported in the registry record. The same is true for progression-free survival and overall survival: they appear among listed outcomes, but no numerical results for them are included in the posted results.
Small event counts require particular care. When a binary outcome occurs in almost everyone, as with the primary endpoint, the estimated proportions can cluster near the upper boundary. In general, this may make differences between arms difficult to estimate precisely, and conventional large-sample approximations can be less reliable when counts are sparse or proportions are near a boundary. The registry record does not identify the precise method used for the primary confidence limits or report a detailed statistical analysis plan. Accordingly, the posted interval and p-value should be treated as the registry's reported statistics, not as independently verified results from a specified model.
Multiplicity arises when several outcomes or comparisons are examined. In general, conducting many tests increases the chance that at least one result appears unusual by chance if each is interpreted without adjustment. This study lists multiple secondary outcomes and posts response results for RECIST 1.1 and irRECIST. The registry record does not report whether a multiplicity procedure was planned or applied, nor does it report a hierarchy that would determine which comparisons are confirmatory. The reported response p-values are explicitly unadjusted, so they merit less inferential weight than results from a clearly prespecified, multiplicity-controlled analysis.
Sample-size planning is distinct from observed power. In general, a planning calculation asks how many participants per arm might be needed to detect a chosen effect under specified assumptions; it does not show the probability that an observed result is true or establish that a completed study had adequate power. The registry's planning illustration is described as using the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation, with effects expressed as fractions of the observed effect. The record does not report a sample-size justification for the trial's own design. No planning counts are included in the record summarized for these results, so no numerical planning requirement should be inferred.
The analysis population also affects interpretation. The primary endpoint description includes randomized participants who received at least one dose, while the arm results give the endpoint populations used. The record does not report a full intention-to-treat analysis, details of exclusions after randomization, missing-data procedures, or sensitivity analyses. General statistical practice distinguishes an analysis restricted to treated participants from one retaining all randomized participants; the choice can affect estimates if treatment initiation or follow-up differs by arm. Without those details, the reported endpoint population should be described as stated rather than assumed to represent every randomized participant.
Over-dispersion is mainly a concern for count outcomes modeled as repeated event counts, where observed variability exceeds that expected under a simple model. General explanation: allowing for over-dispersion can prevent uncertainty from being understated when event counts are more variable than a basic model assumes. The primary endpoint is a count of participants who experienced an event, not a total count of event episodes; no count-model specification or over-dispersion assessment is reported in the registry record. It would therefore be inappropriate to claim that such a model was used or that over-dispersion was tested.
Randomization supports comparisons between assigned regimens, but it does not eliminate uncertainty or solve limitations in outcome measurement and follow-up. The open-label design may affect reporting or assessment of some outcomes, although the response outcomes specify central radiologist review. The registry record does not report the radiology review procedures in enough detail to assess them further. These general considerations do not change the posted estimates; they define the care needed when interpreting them.
11. Confidence Intervals and Precision
The primary event counts are near the top of the possible range: every participant in the 2 mg/kg and 10 mg/kg arms, and all but one in the fixed-dose arm, had at least one adverse event. When an outcome is so common in each group, the observed proportions provide little separation between regimens. The estimate for 10 mg/kg is exactly 0.0 as a risk difference, while the fixed-dose estimate is -0.067. Neither point estimate should be read without the uncertainty around it.
The 95% confidence interval for the fixed-dose comparison extends from -0.193 to 0.060. In general, this interval represents a range of values reasonably compatible with the data and the method used, under repeated-sampling assumptions. Here it includes zero and spans effects in both directions. It therefore does not establish a clear reduction or increase in the primary event proportion for the fixed-dose arm. The reported p-value of 0.326 is also not evidence that the arms are equivalent; a failure to distinguish groups is not the same as demonstrating that any difference is sufficiently small.
The reported interval for the 10 mg/kg primary comparison has limits of 0.0 and 0.0, alongside an estimated risk difference of 0.0. This reflects the reported result, but the registry record does not provide details of the calculation method or a separate p-value for this comparison. The displayed limits should not be taken to imply certainty about the underlying population difference; the arm populations remain small, and the endpoint is observed in every participant in both groups.
The response estimates are also imprecise. For 10 mg/kg, the posted response percentage is 23.1 with a 95% confidence interval from 5.0 to 53.8. For the fixed dose, it is 20.0 with an interval from 4.3 to 48.1. These wide intervals indicate substantial uncertainty in the response proportions. The 2 mg/kg arm has a reported response estimate of 0.0, with an interval from 0.0 to 23.2. In general, intervals around proportions based on small groups can be wide, especially when observed counts are at or near a boundary.
Precision is not the same as validity or clinical importance. A narrow interval can still surround a biased estimate, while a broad interval can include both little effect and effects that matter. The registry record does not define a clinically important margin for the primary endpoint or response outcomes. Therefore, the interval widths describe statistical uncertainty but do not determine whether any possible difference would be important in practice.
12. Multiplicity, Missing Data and Analysis Populations
The registry lists several secondary outcomes, including objective response rate, duration of response, progression-free survival, and overall survival, with response and time-to-event outcomes defined under different assessment frameworks. Numerical secondary results posted in the record include objective response rate under RECIST 1.1 and irRECIST. Results for the other listed secondary outcomes are not reported in the registry record. The record also does not report whether all planned analyses were completed or why numerical results are absent for those outcomes.
Multiplicity is relevant because several outcomes and dose comparisons can produce multiple opportunities for apparently notable findings. General statistical explanation: a multiplicity adjustment or prespecified hierarchy can help control the chance of false-positive conclusions across a family of tests. The registry record does not report an adjustment method, a formal testing hierarchy, or which comparisons were intended to be confirmatory. The response p-values are unadjusted; they should be interpreted as exploratory rather than as a family-wise controlled assessment.
The record does not report how missing outcome data were handled, whether any participants were excluded after randomization, or whether sensitivity analyses assessed alternative assumptions. The primary endpoint population is described as randomized participants who received at least one dose, but a complete account of all randomized participants and their outcome availability is not reported. This matters because analyses restricted to treated participants may differ from analyses that retain everyone as randomized if treatment initiation or follow-up is related to prognosis or outcome.
The registry record does not report imputation procedures, censoring rules for time-to-event outcomes, or a detailed statistical analysis plan. It also does not report interim analyses or stopping rules. These details should not be inferred from the phase, randomization, or final-analysis timeframe. Their absence limits assessment of whether the posted analyses follow a prespecified plan and how missingness or multiple testing may affect the conclusions.
In general, missing-data assumptions and multiplicity control affect the reliability of statistical inference, but the lack of reported detail does not itself establish that a method was inappropriate. It means that readers cannot evaluate those methods from this record. The most defensible interpretation is therefore descriptive: report the observed arm results and their stated uncertainty, and avoid treating the secondary comparisons as conclusive evidence.
13. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Pembrolizumab 2 mg/kg First Course | 2 | 14 | 14.3% |
| Pembrolizumab 10 mg/kg First Course | 1 | 13 | 7.7% |
| Pembrolizumab 200 mg Fixed Dose First Course | 3 | 15 | 20.0% |
| Pembrolizumab 2 mg/kg Second Course | 0 | 1 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Pembrolizumab 2 mg/kg First Course | 11 of 14 |
| Pembrolizumab 10 mg/kg First Course | 12 of 15 |
| Pembrolizumab 200 mg Fixed Dose First Course | 13 of 15 |
| Pembrolizumab 2 mg/kg Second Course | 1 of 1 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Pembrolizumab 2 mg/kg First Course | 14 of 14 |
| Pembrolizumab 10 mg/kg First Course | 13 of 13 |
| Pembrolizumab 200 mg Fixed Dose First Course | 14 of 15 |
| Pembrolizumab 2 mg/kg Second Course | 0 of 1 |
Most frequent serious adverse event terms
| Event term | Organ system | Pembrolizumab 2 mg/kg First Course | Pembrolizumab 10 mg/kg First Course | Pembrolizumab 200 mg Fixed Dose First Course | Pembrolizumab 2 mg/kg Second Course |
|---|---|---|---|---|---|
| Pneumonia | Infections and infestations | 2 of 14 | 0 of 13 | 0 of 15 | 0 of 1 |
| Eyelid ptosis | Eye disorders | 0 of 14 | 0 of 13 | 1 of 15 | 0 of 1 |
| Abdominal pain upper | Gastrointestinal disorders | 0 of 14 | 1 of 13 | 0 of 15 | 0 of 1 |
| Cerebral infarction | Nervous system disorders | 0 of 14 | 0 of 13 | 1 of 15 | 0 of 1 |
| Dermatitis acneiform | Skin and subcutaneous tissue disorders | 0 of 14 | 0 of 13 | 1 of 15 | 0 of 1 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Pembrolizumab 2 mg/kg First Course | Pembrolizumab 10 mg/kg First Course | Pembrolizumab 200 mg Fixed Dose First Course | Pembrolizumab 2 mg/kg Second Course |
|---|---|---|---|---|---|
| Fatigue | General disorders | 5 of 14 | 4 of 13 | 4 of 15 | 0 of 1 |
| Rash | Skin and subcutaneous tissue disorders | 5 of 14 | 2 of 13 | 4 of 15 | 0 of 1 |
| Decreased appetite | Metabolism and nutrition disorders | 0 of 14 | 4 of 13 | 5 of 15 | 0 of 1 |
| Dyspnoea | Respiratory, thoracic and mediastinal disorders | 4 of 14 | 1 of 13 | 4 of 15 | 0 of 1 |
| Upper respiratory tract infection | Infections and infestations | 4 of 14 | 3 of 13 | 1 of 15 | 0 of 1 |
| Alanine aminotransferase increased | Investigations | 4 of 14 | 1 of 13 | 1 of 15 | 0 of 1 |
| Haemoglobin decreased | Investigations | 3 of 14 | 1 of 13 | 2 of 15 | 0 of 1 |
| Constipation | Gastrointestinal disorders | 2 of 14 | 0 of 13 | 3 of 15 | 0 of 1 |
| Pyrexia | General disorders | 0 of 14 | 3 of 13 | 2 of 15 | 0 of 1 |
| Aspartate aminotransferase increased | Investigations | 3 of 14 | 2 of 13 | 0 of 15 | 0 of 1 |
Serious events with the highest arm-level rate: pneumonia (14.3%), abdominal pain upper (7.7%), eyelid ptosis (6.7%), cerebral infarction (6.7%), dermatitis acneiform (6.7%).
The registry reports serious adverse events separately from the primary endpoint of participants with at least one adverse event. During the first course, serious adverse events were reported for 2 of 14 participants in the 2 mg/kg arm, 1 of 13 in the 10 mg/kg arm, and 3 of 15 in the 200 mg fixed-dose arm. The registry also lists a second course for the 2 mg/kg regimen, with 0 of 1 participants recorded as having a serious adverse event. These figures are descriptive and involve small arm populations.
The listed serious-event terms include Pneumonia at 14.3%, Abdominal pain upper at 7.7%, Eyelid ptosis at 6.7%, Cerebral infarction at 6.7%, and Dermatitis acneiform at 6.7%. The registry record does not give further detail in the information about the event-term denominators, overlap among participants, timing, or attribution to treatment. A listed event term should not be assumed to have been caused by pembrolizumab.
The record also reports deaths of 11, 12, and 13 in the first-course groups, respectively, and 1 in the second-course group. The information does not report causes of death or a causal assessment. Serious adverse events and deaths are not interchangeable with the primary endpoint, and these counts alone do not establish comparative safety or treatment causation.
14. What the Result Does, and Does Not, Mean
The results establish what was recorded in this study population under the registry's endpoint definitions: at least one adverse event was reported for every participant in the 2 mg/kg and 10 mg/kg groups and for all but one participant in the 200 mg fixed-dose group. The primary risk-difference estimate for the fixed-dose regimen is slightly below zero, but its confidence interval includes zero and extends in both directions. The data do not provide a precise estimate of a difference between those regimens.
The findings do not show that pembrolizumab caused every recorded adverse event. The primary endpoint definition explicitly includes events whether or not they were considered related to treatment. Nor do the results establish that any regimen is safer than another: the comparison groups are small, the event outcome is nearly universal, and the reported interval for the fixed-dose comparison remains compatible with different directions of effect.
The response findings describe observed objective responses under the registry's stated assessment criteria. They do not establish that one dose is superior, because the estimates are imprecise, the reported comparisons are unadjusted, and the record does not provide a multiplicity plan. They also do not establish a survival benefit. Numerical results for duration of response, progression-free survival, and overall survival are not reported in the registry record.
Because the study compares three pembrolizumab regimens rather than pembrolizumab with placebo or no treatment, it cannot estimate the effect of receiving pembrolizumab versus not receiving it. The results are specific to the enrolled adult Chinese participants with locally advanced or metastatic non-small-cell lung cancer and to the stated treatment and follow-up context. The registry record does not provide enough information to generalize beyond that context with confidence.
In general, statistical evidence is strongest when estimates are precise, methods are clearly specified, and results align with a prespecified analysis plan. Here, the arm counts and posted intervals support a cautious description of observed outcomes, not a definitive claim of equivalence, superiority, or causal safety difference. The record's missing details about analysis populations, missing data, and multiplicity further limit the strength of conclusions that can be drawn.
15. Why This Trial Matters Statistically
This trial is statistically informative as an example of the difference between observing an outcome and precisely comparing groups. The primary endpoint occurred in nearly every participant in every arm, so the observed counts offer a clear description of the study experience but little separation among dose regimens. A point estimate alone—especially one based on small groups—cannot resolve whether underlying event proportions differ.
The registry also demonstrates why confidence intervals matter alongside estimates and p-values. The fixed-dose primary comparison has a negative point estimate, but its interval reaches across zero and into positive values. The estimate therefore does not settle the direction of a population difference. Similarly, the response results include arm estimates and unadjusted comparisons, but broad intervals and multiple outcomes make a single nominal result weak evidence for a confirmed dose effect.
Another statistical lesson is that outcome definitions shape interpretation. The primary outcome counts participants with at least one adverse event, irrespective of causal attribution; it is not a count of all events or a treatment-related-event rate. Serious adverse-event counts provide a separate safety description and should not be conflated with the primary measure. The response outcomes address efficacy-related observations but do not substitute for survival results.
Finally, a completed randomized study does not necessarily resolve every comparison of interest. The registry record identifies the design and results, but does not report the trial's sample-size justification, detailed analysis plan, missing-data methods, or multiplicity strategy. These omissions constrain inferential strength. The practical statistical contribution is therefore a transparent account of observed arm-level outcomes and their uncertainty, while underscoring the distinction between descriptive evidence and conclusive comparative evidence.
16. 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 arm populations are small, and the primary adverse-event outcome is observed in almost everyone. This leaves little room for the endpoint to distinguish dose regimens and limits the precision of comparative estimates. The confidence interval for the fixed-dose comparison is broad and includes values on both sides of no difference. The results should not be taken as evidence that the regimens are equivalent.
The trial is open-label. Although the response outcomes specify central radiologists' review, the registry record does not report review procedures or further safeguards against assessment bias. The primary endpoint includes adverse events irrespective of whether they were judged related to treatment, so it describes overall reported events rather than treatment-attributable harms. The registry record does not provide enough detail to assess event collection or attribution practices.
The comparison is among pembrolizumab regimens, with no placebo or untreated arm. It cannot estimate the effect of pembrolizumab versus no pembrolizumab. The response estimates are imprecise and the p-values are unadjusted; a multiplicity plan and testing hierarchy are not reported. Numerical results for several listed secondary outcomes, including duration of response, progression-free survival, and overall survival, are not reported in the registry record.
Further limitations arise from reporting detail. The record does not report the randomization procedure, allocation concealment, baseline balance, a full account of post-randomization exclusions, missing-data methods, sensitivity analyses, or a detailed statistical analysis plan. It also does not report the trial's own sample-size justification or interim-analysis procedures. These absences do not establish that particular methods were omitted in conduct, but they prevent evaluation of those methods from the registry record.
The enrollment figure is 44, while the primary endpoint arm populations are reported separately as 14, 13, and 15. The registry information does not explain the difference between the enrollment figure and these outcome populations. The primary endpoint population is defined as randomized participants who received at least one dose, but additional participant-flow detail is not reported.
17. Related Tutorials
18. Related Statistical Calculators
19. Sources
- ClinicalTrials.gov: Study of the Safety, Tolerability, Pharmacokinetics and Efficacy of Pembrolizumab (MK-3475) in Chinese Participants With Non-Small-Cell Lung Cancer (MK-3475-032/KEYNOTE-032) — NCT02835690.
- Registry record last updated 2022-10-26; status last verified 2022-09.
- Results first posted 2019-10-07.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
20. Record Summary
The registry describes a completed, open-label, randomized phase I parallel-group study of pembrolizumab in adult Chinese participants with locally advanced or metastatic non-small-cell lung cancer. It lists an enrollment of 44 and three intravenous regimens administered every 3 weeks: 2 mg/kg, 10 mg/kg, and a 200 mg fixed dose. Cycle 1 was 28 days, and subsequent cycles were 21 days.
The primary outcome was the number of participants with at least one adverse event during the first course. The posted counts were 14 of 14, 13 of 13, and 14 of 15 across the respective arms. The fixed-dose comparison against 2 mg/kg had a risk difference of -0.067, a 95% confidence interval from -0.193 to 0.060, and an unadjusted p-value of 0.326.
Objective response rate results are posted under RECIST 1.1 and irRECIST; the registry record does not report numerical results for several other listed secondary outcomes. The study started on 2016-08-04, reached primary completion on 2017-09-19, and is listed as completed. Unreported analysis details and broad uncertainty limit the strength of comparative conclusions.