Chemotherapy and Radiation in Treating Participants With Stage 3 Non-Small Cell Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Chemotherapy and Radiation in Treating Participants With Stage 3 Non-Small Cell Lung Cancer (NCT00686959), a randomized phase 3 study. The primary outcome is Overall Survival.

2. Clinical Question

This is a comparison of treatment strategies across both the concurrent chemoradiotherapy and consolidation phases, rather than a comparison of a single drug in isolation. The strategies differ in the concurrent chemotherapy agents, the stated cycle schedules, and the consolidation treatment. Consequently, the posted comparison cannot identify which component, if any, accounts for a difference between arms.

The registry reports overall survival as its primary outcome and describes progression-free survival, response, survival rates, disease-failure location, and swallowing diary scores among the secondary outcomes. The central question is whether the posted survival estimates and their uncertainty indicate a clear separation between strategies. They do not show a statistically clear difference in overall survival: the reported hazard ratio is 0.98, its 95% confidence interval is 0.79 to 1.20, and the reported p-value is 0.831. These results describe the registry analysis; they do not establish that the strategies are identical or determine which is preferable for an individual.

3. Trial Design

The registry identifies this as a randomized, parallel-group Phase 3 trial with no masking. It is marked completed. Enrollment is listed as 598. The registry record gives a start date of 2008-09 and a primary completion date of 2014-10. The sponsor is Eli Lilly and Company.

Participants were assigned to one of two multi-part treatment strategies. Arm A combined pemetrexed and cisplatin with concurrent thoracic radiation therapy, followed by a recovery period and consolidation pemetrexed. Arm B combined etoposide and cisplatin with concurrent thoracic radiation therapy, followed by a recovery period and consolidation cytotoxic chemotherapy of choice. The registry gives the concurrent schedules as three 21-day cycles for Arm A and two 28-day cycles for Arm B; it describes consolidation as up to four 21-day cycles for Arm A and up to two cycles for Arm B. These are descriptions of the treatment arms, not evidence that every participant completed every planned phase.

The primary analysis population is described as all randomized participants. The registry lists 301.0 participants in Arm A and 297.0 in Arm B for overall survival. The endpoint is time-to-event, and the registry describes Kaplan-Meier estimates for summarizing overall survival. Participants not known to have died by the data cut-off were censored at their last contact date when known alive. Arm A had 124 participants censored and Arm B had 117 participants censored.

The registry record does not report the original sample-size justification, the intended statistical plan in full, or whether interim analyses were planned or performed. It also does not report enough detail to assess adherence to assigned treatment or the balance of other prognostic factors. Randomization supports a comparative interpretation of the assigned strategies, while the unmasked design and the multi-component regimens are relevant context for interpreting treatment delivery and outcome assessment.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants. Arm A had 124 participants censored and Arm B had 117 participants censored.

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

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival, measured over Baseline to Date of Death from Any Cause (Up to 71.4 Months).

Overall survival is the time from baseline to death from any cause. For participants not known to have died by the data cut-off, the registry says the time was censored at the last contact date they were known to be alive. The stated timeframe extends from baseline to death from any cause, up to 71.4 Months.

Overall survival is used because it captures death from any cause rather than requiring a judgment about whether a death was caused by cancer or treatment. It is a time-to-event endpoint: the analysis considers both whether an event occurred and when it occurred, while accounting for participants whose event time is not observed before follow-up ends. This makes censoring details important to interpretation; censoring does not mean that a participant had the event, nor does it establish what happened after the last known contact.

The registry reports median overall survival for each arm using Kaplan-Meier estimates, alongside 95% confidence intervals. It also reports a hazard ratio from a Log Rank analysis. The median describes the estimated time by which half of the survival distribution has reached the event, while the hazard ratio summarizes the relative event rate over follow-up under the analysis. These are different summaries: neither should be read as a direct measure of an individual participant’s outcome.

Overall survival is clinically consequential, but a registry summary cannot by itself explain the causes of any observed pattern, characterize every aspect of quality of life, or establish which parts of a multi-component strategy contributed to the result. The reported analysis is evidence about the trial’s defined endpoint and follow-up, interpreted with its confidence interval and analysis context.

7. Results Posted in the Registry

• Arm A: Pemetrexed + Cisplatin and TRT vs Arm B: Etoposide + Cisplatin and TRT: hazard ratio 0.98 (95% CI 0.79–1.20), reported p = 0.831 (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)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Arm A: Pemetrexed + Cisplatin and TRT vs Arm B: Etoposide + Cisplatin and TRT: hazard ratio 0.86 (95% CI 0.71–1.04), reported p = 0.130 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Arm A: Pemetrexed + Cisplatin and TRT vs Arm B: Etoposide + Cisplatin and TRT and the control. The interval gives the range of effects that remain compatible with the data.

Objective Response Rate (Complete Response [CR] + Partial Response [PR])

• 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:
  • Arm B: Etoposide + Cisplatin and TRT: 33.0% (98/297) versus 35.9% (108/301) on control; risk difference -2.9 percentage points (95% CI -10.5 to 4.7), risk ratio 0.92 (0.74–1.15); unadjusted two-proportion z-test p = 0.458.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm B: Etoposide + Cisplatin and TRT and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With a Post Baseline Swallowing Diary Score >=4

• 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:
  • Arm B: Etoposide + Cisplatin and TRT: 29.0% (78/269) versus 33.8% (96/284) on control; risk difference -4.8 percentage points (95% CI -12.5 to 2.9), risk ratio 0.86 (0.67–1.10); unadjusted two-proportion z-test p = 0.224.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm B: Etoposide + Cisplatin and TRT and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes give additional views of disease control, response, survival over fixed periods, disease-failure patterns, and swallowing symptoms. The posted results include progression-free survival, objective response rate, and the swallowing diary outcome. Although the registry lists survival rates at 1, 2, and 3 Years and first site of disease failure, results for those outcomes are not reported in the registry record.

Across the secondary endpoints with posted comparisons, the point estimates generally lean in the direction of Arm A, while the reported p-values do not show a statistically clear between-arm difference. Progression-free survival is the closest of the posted secondary comparisons to the conventional null reference based on its reported interval and p-value, but its confidence interval includes 1. The response and swallowing diary comparisons also do not provide clear evidence of a between-arm difference under their reported tests. This overall pattern is broadly consistent with the primary overall-survival result, which likewise does not show a statistically clear difference.

Secondary outcomes are useful because they describe aspects of treatment outcomes that overall survival does not capture. However, they are not a substitute for the primary endpoint. Multiple outcomes create multiple opportunities for apparently unusual results, especially when no multiplicity adjustment is reported. The registry record does not say whether secondary hypotheses were prespecified, whether a formal hierarchy was used, or whether any adjustment was applied. Their p-values should therefore be treated as exploratory or supportive rather than as independent confirmation of a treatment effect.

Different outcomes also have different definitions, time windows, and analysis populations. The registry gives time-to-event summaries for progression-free survival and percentage outcomes for response and swallowing diary score. These measures do not all answer the same question, and similar directional estimates across several outcomes do not establish a common mechanism. The secondary findings add context, but the primary comparison remains the main basis for interpreting the trial’s central survival question.

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

• Progression-free Survival (PFS)
• Objective Response Rate (Complete Response [CR] + Partial Response [PR])
• Survival Rates at 1, 2, and 3 Years
• First Site of Disease Failure in Terms of Relapse
• Percentage of Participants With a Post Baseline Swallowing Diary Score >=4

9. Statistical Interpretation

10. Statistical Methods Explained

The methods help distinguish what each reported statistic says from what it cannot establish. The primary endpoint is time-to-event, so the registry uses survival-analysis summaries rather than treating every participant’s follow-up as though it ended at the same time.

What does Kaplan-Meier estimation describe?

In general, Kaplan-Meier estimation uses the observed event times and incorporates participants who are censored, meaning their event was not observed before the relevant follow-up ended. It estimates the probability of remaining event-free over time under assumptions about censoring. The registry states that overall survival time was censored at the last contact date when a participant was known to be alive before the data cut-off. It reports the primary analysis population as all randomized participants. This description does not supply the full details needed to assess censoring patterns or whether censoring was independent of prognosis.

The reported medians are points on the estimated survival distributions. A median is not an average and does not describe the outcome of each participant. The 95% confidence interval around a median communicates uncertainty in its estimate. Intervals for the two arm medians overlap, but overlap between separate arm-specific intervals is not itself the formal test of the between-arm comparison.

What does the hazard ratio represent?

In general, a hazard is an event rate at a particular time among participants still at risk. A hazard ratio compares those rates between groups over follow-up. A value of 1 is the reference for equal hazards; values below or above 1 indicate the direction of the relative comparison according to the order of the arms. The reported overall-survival hazard ratio is for Arm A versus Arm B. A hazard ratio is not a ratio of median survival times, a probability of survival, or a measure of how many months an individual gains.

The registry reports a Log Rank method for the primary comparison. In general, the Log Rank test compares event-time patterns between groups while taking account of censoring. The reported p-value is 0.831. A p-value summarizes how compatible the data are with a specified no-difference null model, given the test assumptions; it is not the probability that the null hypothesis is true. A large p-value is not proof of no effect. It means that this analysis did not identify a statistically clear departure from the null at commonly used thresholds. The registry record does not report the full analysis plan, including any prespecified threshold or interim-monitoring approach.

How should the confidence interval be read?

In general, a confidence interval gives a range of estimates compatible with the data and the method’s assumptions, with a stated long-run coverage property. The primary hazard-ratio interval is 0.79 to 1.20. Since it includes 1, the estimate is compatible with the no-difference reference as well as with relative effects on either side. The interval is not a range that guarantees the true effect lies within it, and it does not measure the probability that any particular effect is true.

Why do binary outcomes use different summaries?

Response and swallowing diary outcomes are reported as percentages of participants. For such outcomes, the registry also gives arm-specific confidence intervals and a comparison p-value. In general, a percentage is a proportion meeting the outcome definition within the analysis population; it does not retain timing information in the way a time-to-event endpoint does. The posted response comparison reports an unadjusted p-value of 0.458. The swallowing diary comparison reports a Fisher Exact p-value of 0.150. Fisher Exact is a method for testing association between categorical outcomes and groups, particularly useful when applying an exact calculation rather than relying solely on a large-sample approximation.

The reported binary p-values do not explain the clinical importance of any difference, and a p-value should not be read as an effect size. The registry does not report adjustment for multiple secondary comparisons. When many outcomes are examined, the chance of at least one apparently unusual result can increase; multiplicity methods can address this, but the registry record does not say whether such methods were planned or used.

Why can over-dispersion matter in other analyses?

In general, over-dispersion occurs when observed variability exceeds the amount expected under a model. If a model assumes less variability than the data contain, standard errors can be too small and confidence intervals too narrow. This is a general statistical consideration, not a reported feature of this trial’s primary analysis. The registry identifies a Log Rank analysis and Kaplan-Meier estimates for overall survival; it does not report an over-dispersion model for this endpoint.

What is not established by the registry description?

The registry does not report the original sample-size justification, detailed covariate adjustment plans, interim analyses, or complete handling rules for missing secondary outcome data. It does state that the primary endpoint population was all randomized participants and describes censoring for overall survival. These details should not be expanded into assumptions about unreported procedures. Interpretation should therefore stay close to the reported estimates, intervals, p-values, population description, and endpoint definitions.

11. Confidence Intervals and Precision

Precision concerns how tightly an estimate is determined by the data and analysis. The overall-survival median estimate is 26.81 months for Arm A, with a 95% confidence interval from 20.4 to 30.92 months. For Arm B, the median is 24.97 months, with a 95% confidence interval from 22.18 to 29.83 months. These intervals reflect uncertainty in each arm’s median; their widths are not a direct test of the difference between arms.

The between-arm hazard ratio is 0.98, with a 95% confidence interval from 0.79 to 1.20. That range includes 1, the no-difference reference, and extends to both lower and higher relative hazards. Thus, the point estimate alone should not be taken as a precise statement of the comparative effect. The p-value of 0.831 is consistent with the interval’s lack of a clear separation from the reference, but neither statistic proves equivalence.

Precision also depends on the amount and timing of event information, not just the number enrolled. For overall survival, the registry reports censoring counts of 124 in Arm A and 117 in Arm B, but does not provide a complete account of follow-up distribution or the information available at each time. The stated timeframe is up to 71.4 Months. A single follow-up limit does not mean every participant was followed for that length of time.

For secondary endpoints, the intervals and estimates differ by outcome. Those results should be interpreted using their own definitions and comparison methods rather than borrowing precision conclusions from overall survival. The registry record does not report all the information needed to assess precision under alternative models or assumptions. The posted confidence intervals are therefore best read as uncertainty under the reported analysis, not as guarantees about the range of true effects.

12. Planning a Similar Trial: Sample Size

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

The sample-size illustrations use the observed effect as a planning reference, then show the number of events or participants that would be needed under stated assumptions to target a chosen power. For overall survival and progression-free survival, the figures are total events, not participants per arm. For the binary outcomes, figures are participants per arm.

The planning 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 time-to-event illustrations use the Schoenfeld method. These are planning illustrations based on observed results, not the trial’s original design assumptions, which are not reported in the registry record.

In plain language, the 80% power column gives the event or enrollment target estimated to provide an 80% chance of detecting the assumed effect under those planning assumptions, if that effect is real. The 90% power column gives the corresponding target for a higher chance of detecting it. These are design probabilities under a model, not guarantees for a future study. Smaller assumed effects require more information and therefore more events or participants. For example, the planning rows for 50% of the observed effect call for larger targets than the rows for 100% of the observed effect.

The event totals shown for the time-to-event endpoints should not be confused with total enrollment, and the binary targets should not be read as total sample sizes: they are per arm. These figures are not observed power. They do not describe the probability that this completed trial detected its actual effect, nor do they replace the trial’s own sample-size justification, which is not reported in the registry record.

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 = 0.980, so ln(HR) = -0.0202.
2. d = 4 × (1.96 + 0.842)² / [ln(0.980)]² = 76,921.6, rounded up to 76,922 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 <- 0.98                              # 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(686959)
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 = 76,922 (80% power), 102,977 (90% power)
 75% of observed effect: events = 136,750 (80% power), 183,069 (90% power)
 50% of observed effect: events = 307,687 (80% power), 411,905 (90% power)
Simulated power with 76,922 events: 0.798 (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 171,439 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 598 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 <- 24.97          # control-arm median, months (converted from the posted unit)
hr          <- 0.98          # 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: 152196         p (1-sided)   0.0001
  Events: 34731        ~HR at bound   0.9610
      Month: 21    P(Cross) if HR=1   0.0001
                P(Cross) if HR=0.98   0.0338
      IA 2: 67%                   Z   2.5114
      N: 171440         p (1-sided)   0.0060
  Events: 69461        ~HR at bound   0.9811
      Month: 34    P(Cross) if HR=1   0.0060
                P(Cross) if HR=0.98   0.5603
          Final                   Z   1.9930
      N: 171440         p (1-sided)   0.0231
 Events: 104192        ~HR at bound   0.9877
      Month: 60    P(Cross) if HR=1   0.0250
                P(Cross) if HR=0.98   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 registry reports the sponsor’s overall-survival comparison as a Log Rank analysis with a hazard ratio of 0.98, a 95% confidence interval of 0.79 to 1.20, and p = 0.831. The statistical display uses this reported hazard ratio rather than creating a separate hazard-ratio estimate from arm-level medians. A median alone is not sufficient to reconstruct a time-to-event hazard ratio.

The registry record does not flag a discrepancy between the reported result and the analysis used for the statistical display. Accordingly, the interpretation is anchored to the sponsor-reported hazard ratio and p-value. If there were a discrepancy between a sponsor analysis and a simple recalculation from posted summaries, the sponsor’s own analysis would take precedence; no such discrepancy is flagged for this comparison.

The same distinction matters for other outcomes. Arm-level percentages can be described directly, but they do not replace the reported comparison analysis. The response result has a posted p-value, and the swallowing diary result has a posted Fisher Exact p-value. Those values should be treated as the registry’s analyses, not as independently adjusted or confirmatory findings. The registry record does not provide a full statistical analysis plan with which to evaluate every modeling or testing choice.

Thus, the sponsor result is the formal source for the reported time-to-event comparison, while the arm-level estimates and confidence intervals provide useful descriptive context. The reported analysis does not establish equivalence or prove the absence of an important effect.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, comparisons, time points, or subgroups. In general, testing more hypotheses increases the chance that at least one result will appear unusual by chance alone. Prespecified procedures such as a testing hierarchy or adjustment of significance levels can help control that risk, but the registry record does not report whether such a procedure was planned or used for the secondary outcomes.

The primary endpoint is overall survival, and the record reports its hazard ratio, confidence interval, and Log Rank p-value. The registry also lists multiple secondary outcomes, while posted results are available for only some of them. It is not reported in the registry record whether all secondary analyses were prespecified, whether subgroup analyses were conducted, or how the reported p-values should be interpreted in a testing hierarchy. The secondary findings should consequently be treated with more caution than the primary comparison, rather than as multiple independent confirmations.

Missing-data interpretation also differs by endpoint. For overall survival, the registry describes censoring participants not known to have died by the data cut-off at their last contact date when they were known to be alive. This is a standard way to represent incomplete observation in a time-to-event analysis, but interpretation relies on assumptions about the relationship between censoring and future event risk. The registry does not report enough detail to evaluate those assumptions or the pattern of follow-up loss.

The primary population is described as all randomized participants. For the secondary percentage outcomes, arm-specific denominators are reported, but the registry record does not fully explain missing-data handling or why the analysis populations differ from the primary endpoint population. It also does not report imputation methods or sensitivity analyses for missing outcomes. These omissions limit assessment of whether missingness could affect the estimates.

The registry record does not report the full statistical analysis plan, original sample-size justification, or interim-analysis procedures. Those details should not be inferred from the final results. The most defensible reading is therefore tied to the stated analysis population and endpoint definitions, the reported censoring description, and the available estimates and intervals, with appropriate caution for unreported multiplicity and missing-data procedures.

17. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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: oesophagitis (10.7%), febrile neutropenia (8.8%), pneumonia (6.6%), vomiting (4.9%), neutropenia (4.8%).

The registry lists serious adverse events by arm using participants at risk and participants affected. For the serious-event summary, Arm A lists 283 at risk and 134 affected; Arm B lists 272 at risk and 145 affected. These counts describe the registry’s safety summary and should be interpreted using the listed denominators rather than the overall randomized population. The record also lists other adverse events, with 279 affected in Arm A among 283 at risk and 269 affected in Arm B among 272 at risk.

The most frequently listed serious adverse-event terms are oesophagitis at 10.7%, febrile neutropenia at 8.8%, pneumonia at 6.6%, vomiting at 4.9%, neutropenia at 4.8%, and radiation oesophagitis at 4.8%. The registry does not provide arm-specific percentages for these terms in the listed summary, so they should not be attributed to one treatment arm or used to compare arm-specific risks.

These figures are descriptive safety information, not proof that a particular treatment caused an event. The registry record does not provide details on event severity beyond the serious-event classification, timing, duration, or treatment attribution. Nor does it report deaths in the safety summary. The listed counts and terms therefore provide a limited overview and should not be treated as a complete safety profile or a direct benefit-risk comparison.

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

The primary result describes a randomized comparison of two complete treatment strategies for locally advanced, unresectable Stage III non-small cell lung cancer with nonsquamous histology. The reported median overall survival estimates are 26.81 months for Arm A and 24.97 months for Arm B. The sponsor-reported hazard ratio is 0.98, with a 95% confidence interval of 0.79 to 1.20 and p = 0.831. Taken together, these results do not show a statistically clear difference in overall survival under the reported analysis.

That finding is not the same as demonstrating that the strategies are equivalent. The confidence interval includes the no-difference reference and also allows relative effects on either side. A non-significant p-value does not establish that there is no treatment effect, and the trial’s original equivalence or non-inferiority margins are not reported in the registry record. No claim of equivalence can be made from these results alone.

The comparison concerns treatment packages, not isolated components. The concurrent chemotherapy agents and consolidation therapies differ between arms, alongside the stated cycle schedules. The observed comparison cannot determine whether any outcome pattern is attributable to a particular drug, radiation, treatment schedule, or consolidation choice. It also cannot directly answer questions about unreported outcomes or populations outside the registry’s eligibility description.

Secondary results add information about progression, response, and swallowing diary scores, but they do not replace the primary endpoint. Some listed secondary outcomes have no results reported in the registry record, and the record does not state whether multiplicity was addressed. Safety figures are also limited to the reported counts and terms.

These estimates summarize the trial population and registry analysis, not the likely outcome for a particular person. They do not offer medical advice or establish which strategy should be chosen. Interpretation depends on the study population, treatment definitions, follow-up, censoring, and the uncertainty reflected in the confidence intervals.

19. Why This Trial Matters Statistically

This trial illustrates why a clinical comparison should be read through both its point estimate and its uncertainty. The overall-survival hazard ratio is close to the no-difference reference, but the confidence interval extends on both sides of that reference. Reporting only the point estimate could make the comparison appear more conclusive than the data support; reporting only the p-value would also conceal the range of effects compatible with the analysis.

It also shows the distinction between a descriptive arm summary and a formal comparative analysis. Median survival in each arm is informative, but the reported hazard ratio and Log Rank test address the time-to-event comparison. Because follow-up is not complete for every participant at the data cut-off, censoring is part of the analysis rather than a reason to discard those observations. Understanding how censoring was defined helps readers interpret what the survival estimates represent.

The design compares complex treatment strategies, each with concurrent chemoradiotherapy and consolidation. This has practical relevance for the trial question but limits attribution: the analysis cannot identify the contribution of an individual agent or phase. Randomization supports comparison of assigned strategies, while the unmasked design and incomplete reporting of treatment adherence and assessment procedures remain important context.

Finally, secondary outcomes and safety summaries broaden the picture but require careful handling. Multiple endpoints can produce chance findings, and the registry does not report a multiplicity plan. Sample-size illustrations based on observed effects can help readers understand how much information smaller effects require, but they are not the trial’s original design rationale and are not observed power. Together, these distinctions support a more calibrated reading of registry results.

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.

The registry record leaves several trial-specific details unavailable. It does not report the original sample-size justification, the complete statistical analysis plan, interim-analysis procedures, or whether multiplicity adjustment was prespecified. Without those details, it is not possible to assess the original power target, the role of any interim decisions, or the intended hierarchy for interpreting secondary findings.

The primary endpoint population is described as all randomized participants, and censoring is described for participants not known to have died by the data cut-off. However, the record does not provide a fuller account of follow-up duration by arm, reasons for loss to follow-up, or sensitivity analyses for censoring assumptions. The reported censoring counts alone do not show whether censoring was related to prognosis.

For secondary outcomes, analysis denominators differ from those listed for overall survival, and the record does not explain all missing-data handling. Results are not reported for every secondary outcome named in the registry. The availability of results for a subset of listed outcomes makes it difficult to assess the full pattern of planned secondary analyses or selective reporting.

The treatment arms are multi-component strategies, so differences cannot be attributed to one drug, the radiation component, or consolidation alone. The registry record also gives only limited safety detail: serious-event counts and selected terms are posted, but arm-specific term estimates, event timing, attribution, and other safety details are not reported in the summary. No deaths are listed in the safety data.

These limitations do not invalidate the reported randomized comparison. They define what can be concluded from the registry record and where a more detailed protocol, analysis plan, or full report would be needed for further assessment.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Chemotherapy and Radiation in Treating Participants With Stage 3 Non-Small Cell Lung Cancer — NCT00686959.
• Registry record last updated 2016-06-28; status last verified 2016-05.
• Results first posted 2015-11-26.
• 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

NCT00686959 is a completed randomized, parallel-group Phase 3 trial with no masking, sponsored by Eli Lilly and Company. The registry lists enrollment of 598 and a primary completion date of 2014-10. It compares pemetrexed, cisplatin, and thoracic radiation followed by consolidation pemetrexed with etoposide, cisplatin, and thoracic radiation followed by consolidation cytotoxic chemotherapy of choice in participants with unresectable, locally advanced Stage III non-small cell lung cancer other than predominantly squamous cell histology.

The primary endpoint is overall survival from baseline to death from any cause. The reported hazard ratio is 0.98, with a 95% confidence interval of 0.79 to 1.20 and p = 0.831. The result does not show a statistically clear difference between the strategies and does not establish equivalence. Secondary outcomes and serious adverse-event summaries add context, but some listed secondary results and important analysis details are not reported in the registry record.