Phase III Trial to Assess Efficacy and Safety of Cetuximab for the Treatment of Chinese Participants With Head and Neck Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase III Trial to Assess Efficacy and Safety of Cetuximab for the Treatment of Chinese Participants With Head and Neck Cancer (NCT02383966), a randomized phase 3 study. The primary outcome is Progression-free Survival (PFS) Time, as Assessed by an Independent Review Committee (IRC).

2. Clinical Question

3. Trial Design

This completed phase III trial was randomized, parallel-group, and open-label, with no masking reported. It enrolled 243 participants. The registry lists a start date of 2015-07-31 and a primary completion date of 2018-01-19. The study was designed to assess efficacy and safety of cetuximab combined with chemotherapy compared with chemotherapy alone as first-line treatment for Chinese participants with recurrent and/or metastatic SCCHN.

The intervention arm is labelled cetuximab plus cisplatin/carboplatin plus 5-fluorouracil; the comparator is labelled cisplatin/carboplatin plus 5-fluorouracil. The official title instead describes cisplatin in both regimens. The registry record does not clarify this difference in wording, so the arm labels are best read as recorded rather than as evidence of which platinum agent each participant received.

The primary endpoint was PFS assessed by an independent review committee (IRC). The primary analysis population is described as the intention-to-treat set, including all randomized participants. The reported group sizes for this endpoint are 164 and 79. Allocation concealment, randomization sequence generation, treatment adherence, and protocol deviations are not reported in the registry record.

The record also lists investigator-assessed PFS, overall survival, best overall response rate, disease control rate, duration of response, and treatment-emergent adverse events among its secondary outcomes. Quantitative results are posted for investigator-assessed PFS, overall survival, and best overall response rate. Numerical results for disease control rate and duration of response are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT analysis set included all participants who were randomized to 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: Intention-to-treat (ITT) analysis set included all participants who were randomized to study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Progression-free Survival (PFS) Time, as Assessed by an Independent Review Committee (IRC), measured over Every 6 weeks starting from the date of randomization until occurrence of PD, assessed up to data-cutoff (904 days).

The primary endpoint is progression-free survival, or PFS, assessed by an independent review committee. It measures the time from randomization until the first observation of progressive disease based on imaging, or death from any cause within 60 days after the last tumor assessment or randomization, whichever is later. The registry defines progressive disease using changes in target lesions, progression of non-target lesions, and/or new lesions, with an absolute increase requirement for the sum of target-lesion diameters. The posted definition refers to at least a 20 percent increase and an absolute increase of at least 5 millimeter.

The registry reports PFS as a median in months estimated with the Kaplan–Meier method. A median is the time by which the estimated survival curve reaches one half; it describes the middle of the estimated event-time distribution, not the time experienced by every participant. A hazard ratio is also reported for the comparison between groups. These two summaries answer related but different questions: the median describes a time point on each group’s curve, while the hazard ratio summarizes the relative event rate over follow-up under the model or analysis used.

PFS is used in oncology trials because it captures disease progression as well as death, rather than requiring death to occur before an event is counted. The registry states that assessments were scheduled every 6 weeks from randomization until progression, with follow-up assessed up to the data cutoff of 904 days. The analysis method beyond Kaplan–Meier estimation and the IRC assessment is not reported in the registry record.

7. Results Posted in the Registry

• Cetuximab + Cisplatin/Carboplatin + 5-Fluorouracil vs Cisplatin/Carboplatin + 5-Flurouracil: hazard ratio 0.57 (95% CI 0.40–0.80) (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.

Progression-free Survival (PFS) Time, as Assessed by the Investigator

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Cetuximab + Cisplatin/Carboplatin + 5-Fluorouracil vs Cisplatin/Carboplatin + 5-Flurouracil: hazard ratio 0.57 (95% CI 0.41–0.80) (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Cetuximab + Cisplatin/Carboplatin + 5-Fluorouracil vs Cisplatin/Carboplatin + 5-Flurouracil than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Overall Survival (OS) Time

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Cetuximab + Cisplatin/Carboplatin + 5-Fluorouracil vs Cisplatin/Carboplatin + 5-Flurouracil: hazard ratio 0.70 (95% CI 0.50–0.99) (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Cetuximab + Cisplatin/Carboplatin + 5-Fluorouracil vs Cisplatin/Carboplatin + 5-Flurouracil than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Best Overall Response Rate (ORR)

• 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:
  • Cisplatin/Carboplatin + 5-Flurouracil: 26.6% (21/79) versus 50.0% (82/164) on control; risk difference -23.4 percentage points (95% CI -35.8 to -11.0), risk ratio 0.53 (0.36–0.79); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Cisplatin/Carboplatin + 5-Flurouracil than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary results point in a direction broadly consistent with the primary endpoint’s point estimate: investigator-assessed PFS and overall survival have hazard ratios below 1, while best overall response rate is higher in the combination group. This directional pattern can provide context for the primary IRC-assessed PFS analysis, but it does not make the secondary endpoints equivalent to the primary endpoint. They use different definitions, assessment processes, or effect measures.

Investigator-assessed PFS is related to the primary endpoint but is not an independent replication of it, because both concern progression-free time in the same trial. Overall survival measures a different event and has its own follow-up and uncertainty. Response rate asks whether a qualifying response occurred rather than how long participants remained progression-free. Disease control rate and duration of response are also listed as outcomes, but their numerical results are not reported in the registry record.

Secondary results generally carry less confirmatory weight when several outcomes or comparisons are examined without a reported multiplicity plan. Repeated testing increases the opportunity to obtain apparently notable results by chance. The registry does not report whether an adjustment strategy, testing hierarchy, or other multiplicity control was prespecified. The unadjusted response-rate p-value should therefore be interpreted as an unadjusted result, not as a family-wise conclusion across all outcomes.

Consistency across related outcomes can be descriptively informative, but it does not resolve uncertainty about analysis methods, missing data, or the impact of open-label assessment. The registry record also does not provide the full statistical analysis plan needed to establish which secondary analyses were prespecified or how they were intended to support the primary question.

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

• Progression-free Survival (PFS) Time, as Assessed by the Investigator
• Overall Survival (OS) Time
• Best Overall Response Rate (ORR)
• Disease Control Rate (DCR)
• Duration of Response (DOR)
• Number of Participants With Treatment Emergent Adverse Events (TEAEs), Treatment Emergent Serious Adverse Events (TESAEs), Treatment Emergent Adverse Events Leading to Death and AEs Leading to Discontinuation

9. Statistical Interpretation

10. Statistical Methods Explained

What does a median PFS estimate describe?

Time-to-event outcomes measure the time until an event such as progression or death. Participants who have not had the event by the end of observed follow-up, or who leave follow-up before an event is known, are generally treated as censored in standard survival analysis. Censoring allows their observed time to contribute information without treating their later event status as known. The registry record does not describe its censoring rules or missing-data procedures in detail, so those aspects cannot be assessed from the posted information.

The Kaplan–Meier method estimates the proportion remaining event-free over time while accounting for censored observations. The median is the time at which the estimated curve reaches 50 percent. It is a summary of the curve, not an average duration and not a prediction for a particular participant. It can be sensitive to the shape of the estimated curve and to the amount of follow-up around the point where the curve crosses one half.

What does the hazard ratio add?

A hazard is the instantaneous event rate among people who have not yet experienced the event, considered over time. A hazard ratio compares this rate between groups. With the combination group as the numerator in the posted comparison, a ratio below 1 indicates a lower estimated hazard in that group under the conventional interpretation. It does not mean that the event risk at every point is reduced by the same amount, and it is not a direct comparison of medians.

A hazard ratio is often estimated under a proportional-hazards framework, in which the relative hazard is treated as constant over time. The registry does not report the method used to estimate these hazard ratios or whether proportional hazards were assessed. A single hazard ratio can be difficult to summarize if the relative hazards change over follow-up. The record also does not report a p-value for the primary hazard-ratio result; the confidence interval is therefore the posted measure of its statistical precision.

Why use independent review?

An IRC assesses progression using information independently from the treating investigator. In general, independent review can reduce the influence of knowledge of treatment assignment on judgments about progression, especially in an open-label trial. It does not remove every source of bias: interpretation still depends on the assessment rules, the available scans, and how missing or unevaluable assessments are handled. Details of these procedures beyond the endpoint definition and assessment schedule are not reported in the registry record.

The registry also reports investigator-assessed PFS. This is related to the primary endpoint but depends on the investigator’s assessment rather than the IRC’s. Agreement between the two measures can be informative, while differences may arise from the assessment process or the timing and interpretation of scans. Their estimates should not be combined or treated as interchangeable.

How should the response-rate comparison be read?

Best overall response rate is a binary outcome: participants are classified according to whether they achieved a qualifying response. The registry posts percentages and confidence intervals by group, a sponsor-reported odds ratio, and an unadjusted recalculation expressed as a risk difference with an unadjusted p-value. A risk difference compares the proportions directly; an odds ratio compares the odds of response. They use different scales and need not have the same numerical value or intuitive interpretation.

The recalculated comparison is explicitly unadjusted. It does not account for multiple outcomes or comparisons, and it is not a substitute for the sponsor’s analysis. The sponsor-reported odds ratio is the registry’s reported analysis for this outcome; its method and p-value are not reported in the registry record. A p-value describes compatibility of data with a specified null model under the test assumptions. It does not give the probability that a treatment effect is true or the probability that the result will replicate.

Why can over-dispersion matter in other settings?

As a general statistical point, over-dispersion occurs when observed variability is greater than a model assumes. If a model underestimates variability, its standard errors can be too small and its intervals too narrow. This issue can matter for count outcomes and some clustered or repeated-measure data. The posted primary endpoint is a time-to-event outcome, and the registry does not report a count model or an over-dispersion assessment. No conclusion about such a model is warranted for this trial.

What do the posted analysis details establish?

The record identifies the primary endpoint, its IRC assessment, the intention-to-treat population, the Kaplan–Meier method for estimating PFS, and the reported hazard ratio. It does not describe the hazard-ratio estimation method, covariate adjustment, treatment of protocol deviations, missing-data handling, interim analyses, or the full statistical analysis plan. Those details are not reported in the registry record. As a result, the posted summaries permit interpretation of the estimated effects and their intervals, but not a full audit of the analysis choices or assumptions.

11. Confidence Intervals and Precision

The primary PFS medians are accompanied by 95% confidence intervals: 5.4 to 5.6 months for the cetuximab-combination group and 3.0 to 5.3 months for chemotherapy alone. The reported hazard ratio is 0.566, with a 95% confidence interval of 0.4 to 0.803. A confidence interval expresses the precision of an estimate under the analysis and its assumptions. It is not a range containing 95 percent of individual participants’ outcomes, and it does not account for biases or unreported analytic choices.

The interval around the combination-group median is comparatively narrow, while the comparator median interval is wider. The hazard-ratio interval spans a range of estimates rather than identifying one exact effect. Its entire posted range is below 1, but the registry does not provide a p-value for this analysis. The lower and upper limits should be read as uncertainty bounds for the estimated relative hazard, not as alternative observed results.

Other outcomes also have confidence intervals. Investigator-assessed PFS and overall survival report hazard-ratio intervals, and the best overall response rates report intervals for each group. The response-rate comparison also has an unadjusted risk-difference interval in the recalculation. These intervals differ in the quantity being estimated, so their widths are not directly comparable without considering the scale and method.

Confidence intervals reflect sampling uncertainty under the statistical procedure; they do not by themselves address open-label conduct, measurement differences, missing information, or whether the enrolled participants represent other populations. The registry does not report all methods needed to evaluate those issues. Precision is therefore one component of interpretation, alongside the endpoint definition, analysis population, and design.

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 planning figures are illustrations based on the observed effect, not the trial’s original design assumptions. For time-to-event outcomes, the entries are total events required, not participants per arm. For the other outcomes, the entries are participants per arm. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The survival planning calculations use the Schoenfeld approach.

The 80% power column gives the illustrative number of events or participants needed for an 80% chance of detecting the specified effect under those assumptions, if the effect and model assumptions hold. The 90% power column gives the corresponding planning amount for a 90% chance. These are conditional planning probabilities, not guarantees about a future study and not descriptions of the results already observed.

Rows that plan for a smaller fraction of the observed effect require more information. For a binary outcome, that is expressed as more participants per arm; for time-to-event outcomes, it is expressed as more total events. The accompanying curve should be read as a visual representation of how the planning requirement changes with the effect being targeted, not as an additional observed result.

These figures are not observed power. The original trial’s sample-size justification and design assumptions are not reported in the registry record, and the planning illustrations should not be taken as a reconstruction of that original plan.

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 = 0.566, so ln(HR) = -0.5692.
2. d = 4 × (1.96 + 0.842)² / [ln(0.566)]² = 96.9, rounded up to 97 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 <- 0.566                              # 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(2383966)
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 = 97 (80% power), 130 (90% power)
 75% of observed effect: events = 173 (80% power), 231 (90% power)
 50% of observed effect: events = 388 (80% power), 519 (90% power)
Simulated power with 97 events: 0.763 (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 143 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 243 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 <- 4.2          # control-arm median, months (converted from the posted unit)
hr          <- 0.566          # 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: 86         p (1-sided)   0.0001
  Events: 44        ~HR at bound   0.3248
   Month: 14    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.57   0.0338
   IA 2: 67%                   Z   2.5114
      N: 138         p (1-sided)   0.0060
  Events: 88        ~HR at bound   0.5838
   Month: 23    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.57   0.5603
       Final                   Z   1.9930
      N: 144         p (1-sided)   0.0231
 Events: 131        ~HR at bound   0.7056
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.57   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 primary PFS result is reported by the sponsor as a hazard ratio of 0.566 with a 95% confidence interval of 0.400 to 0.803. The registry’s displayed primary comparison gives the same estimate and interval, with the interval limits expressed as 0.4 and 0.803. No discrepancy is flagged. For the time-to-event outcomes, the interpretation uses the sponsor-reported hazard ratios rather than a separate recalculation from summary statistics.

The sponsor’s reported analysis for best overall response rate is an odds ratio of 2.76 with a 95% confidence interval of 1.52 to 5.45. The registry does not report a p-value or method for that analysis. A separate unadjusted recalculation expresses the comparison as a risk difference and reports p < 0.001. Because an odds ratio and a risk difference are different effect measures, their values should not be compared as though they were estimates on the same scale.

The sponsor’s reported analysis takes precedence if a discrepancy is flagged. No discrepancy is flagged in this record. The registry does not provide enough detail to reconstruct the sponsor’s analytical method or assess all its assumptions, so the reported estimates should be read with that limitation in mind.

16. Multiplicity, Missing Data and Analysis Populations

The registry identifies an intention-to-treat analysis set for primary PFS: all participants randomized to study treatment. This definition is important because it preserves the randomized group assignment in the primary comparison. The record does not describe whether every randomized participant contributed usable endpoint data, how missing assessments were handled, or whether there were additional analysis populations for secondary or safety outcomes. Those details are not reported in the registry record.

Several secondary outcomes are listed, including investigator-assessed PFS, overall survival, best overall response rate, disease control rate, duration of response, and adverse-event outcomes. The registry does not report a multiplicity strategy, such as an ordering of hypotheses, adjustment of significance thresholds, or another procedure to manage the chance of false-positive findings across multiple tests. It is therefore not possible to determine from the record whether secondary comparisons were intended as confirmatory or exploratory.

The primary hazard-ratio result has a reported 95% confidence interval, but no p-value is posted. The response-rate recalculation has an unadjusted p-value; that does not establish control of error across the set of outcomes. Statistical significance for an individual unadjusted comparison, where reported, should not be interpreted as proof that all related outcomes support the same conclusion.

The registry also does not report interim analyses, stopping rules, covariate adjustment, detailed censoring procedures, or the full statistical analysis plan. Their absence from the registry record does not establish that such procedures were not used; it means their role cannot be evaluated from the posted information. A fuller assessment of multiplicity and missing-data risk would require those analysis details and the corresponding participant-level flow and outcome data.

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: lung infection (5.3%), tumor haemorrhage (3.9%), dyspnoea (3.1%), anaemia (2.6%), thrombocytopenia (2.6%).

The registry posts serious adverse-event figures by group. In the cetuximab-combination group, 46 participants are listed as experiencing a serious adverse event, with 163 at risk. In the chemotherapy-alone group, 21 are listed as experiencing a serious adverse event, with 76 at risk. The record also lists other adverse events for 163 at risk in the combination group and 76 at risk in the comparator group, with 163 and 73 respectively listed as affected. Death counts are posted as 105 and 54, respectively.

These are registry counts, not adjusted comparisons of safety. The record does not establish from these figures whether events were caused by treatment, how long each participant was followed for safety, or whether the groups differed in exposure time. A count alone does not account for event severity beyond the serious-event classification, timing, recurrence, or competing risks.

The most frequently listed serious-event terms include lung infection (5.3%), tumor haemorrhage (3.9%), dyspnoea (3.1%), anaemia (2.6%), thrombocytopenia (2.6%), and upper respiratory tract infection (2.6%). These terms and percentages should be read as registry-reported safety summaries, not as evidence of comparative causation. The registry record does not provide additional details on attribution or exposure-adjusted rates.

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

The posted results describe a randomized comparison of cetuximab combined with chemotherapy against chemotherapy alone in Chinese participants with recurrent and/or metastatic SCCHN receiving first-line treatment. For the primary endpoint, the registry reports IRC-assessed PFS medians and a hazard ratio whose confidence interval lies below 1. The posted point estimates therefore favor the combination on the measured time-to-event summaries, while the confidence interval conveys uncertainty about the relative hazard.

That statistical description is not a guarantee about an individual participant’s outcome, nor does it establish the same effect in populations not represented by the trial. A median is a group-level summary, and a hazard ratio is a relative time-to-event measure. Neither states the precise experience of any one participant or supplies a complete account of how benefits and harms compare for an individual.

The endpoint reflects progression based on imaging or qualifying death, assessed according to the registry definition. It does not by itself describe symptoms, quality of life, treatment burden, or all other outcomes that might matter to participants. Although the record lists adverse-event outcomes, the posted counts alone do not establish causality or provide a full comparative safety analysis.

Interpretation is also limited by the registry’s reporting gaps. The method used to estimate hazard ratios, the multiplicity strategy, missing-data procedures, interim analyses, and the full statistical analysis plan are not reported in the registry record. The trial was open-label, although the primary progression assessment was by an IRC. These features should inform how confidently the estimates can be generalized or treated as definitive.

19. Why This Trial Matters Statistically

This trial is statistically informative because it reports a randomized, parallel-group comparison with an independently reviewed time-to-event primary endpoint. It also identifies the primary analysis population and gives both Kaplan–Meier median estimates and a hazard ratio with a confidence interval. Considering these summaries together is more informative than relying on a single point estimate: the medians describe the estimated event-time curves at a particular point, while the hazard ratio summarizes a relative comparison across follow-up.

The trial also illustrates why endpoint assessment matters. The primary outcome used IRC assessment, while a secondary outcome used investigator assessment. Comparing these related measures can help readers see whether conclusions depend on who assessed progression, although the measures are not interchangeable and do not independently reproduce one another.

Its reporting also demonstrates why confidence intervals and analysis details matter. The primary hazard ratio is reported without a p-value or a method for estimating it, and the registry does not describe multiplicity or missing-data procedures. The direction of a point estimate is only part of the evidence; precision, assumptions, and the analysis plan all affect how the result should be understood.

Finally, the response-rate outcome has both a sponsor-reported odds ratio and an unadjusted recalculation on a different scale. This is a useful reminder that effect measures answer related but distinct questions and should not be compared by their numerical size alone. The record supports a careful statistical reading of the posted outcomes, while leaving important design and analysis details 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 are specific to the information recorded for this trial. The study was open-label, and the registry does not report procedures for allocation concealment or random sequence generation. The primary endpoint’s IRC assessment may reduce reliance on investigator judgment, but the record does not detail the review procedures, handling of unevaluable scans, or whether reviewers were masked to treatment assignment.

The treatment labels are not fully consistent: the official title describes cisplatin plus 5-fluorouracil, whereas the arm labels include cisplatin/carboplatin. The registry record does not clarify this wording. The endpoint analysis results also do not state the hazard-ratio estimation method, covariate adjustment, proportional-hazards assessment, or censoring details. A p-value for the primary hazard ratio is not reported.

There is no registry-reported multiplicity plan, missing-data approach, interim-analysis information, or full statistical analysis plan. Numerical results for disease control rate and duration of response are not reported in the registry record. The safety counts are not accompanied by exposure-adjusted comparisons or detailed attribution information. These gaps do not prove that relevant procedures were absent from the trial; they limit what can be assessed from the registry record.

The results apply directly to the trial’s enrolled population and setting. The registry does not provide information sufficient to assess how well participants represented other populations or how findings would translate to other treatment settings.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Phase III Trial to Assess Efficacy and Safety of Cetuximab for the Treatment of Chinese Participants With Head and Neck Cancer — NCT02383966.
• Registry record last updated 2022-05-13; status last verified 2022-04.
• Results first posted 2019-04-16.
• 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

NCT02383966 is a completed randomized, open-label, phase III parallel-group trial of cetuximab plus chemotherapy versus chemotherapy alone for first-line treatment of Chinese participants with recurrent and/or metastatic SCCHN. The registry lists 243 enrolled participants and identifies IRC-assessed PFS as the primary endpoint, analyzed in the intention-to-treat population with Kaplan–Meier estimates.

The primary result is reported as a hazard ratio of 0.566 with a 95% confidence interval of 0.4 to 0.803. The registry also posts results for investigator-assessed PFS, overall survival, and best overall response rate. Several analysis and design details, including the multiplicity strategy, missing-data handling, and full statistical analysis plan, are not reported in the registry record.