OSE2101 Versus Chemotherapy in HLA-A2 Positive Patients With Advanced NSCLC After Immune Checkpoint Inhibitor Failure: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for OSE2101 Versus Chemotherapy in HLA-A2 Positive Patients With Advanced NSCLC After Immune Checkpoint Inhibitor Failure (NCT02654587), a randomized phase 3 study. The primary outcome is Overall Survival (OS).

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group phase III trial comparing OSE2101 with docetaxel or pemetrexed. The registry lists enrollment of 219 participants, no masking, and a terminated status. It records a start date of 2016-02-12 and primary completion on 2021-01-15. The reason for termination and the timing or circumstances of that decision are not reported in the registry record.

The intervention was OSE2101; the comparator arm could receive docetaxel or pemetrexed. The brief summary describes the study population as patients with metastatic NSCLC who progressed after sequential or concurrent chemotherapy and an immune checkpoint inhibitor in first- or second-line treatment. The primary endpoint description is narrower: it specifies patients with ICI secondary resistance, defined as progression after ICI second line > or equal to 12 weeks. A separate analysis is identified for the ITT population with primary and secondary ICI resistance.

Overall survival was the primary outcome, with the timeframe recorded as approximately 24 months. Secondary outcomes included post-progression survival, time to worsening ECOG performance status, quality-of-life changes, disease control rate at 6 months, and progression-free survival in patients with ICI secondary resistance. The registry lists 12 outcome measures and 41 analyses. The precise allocation ratio, randomization procedure, concealment method, stratification factors, and full protocol are not reported in the registry record. The registry also does not give the reason for termination or a full account of how the listed analyses were selected.

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: Only patient with HLA-A2 positive phenotype were eligible

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: OS analysis in patients with ICI secondary resistance defined as disease progression after ICI second line \> or equal to 12 weeks (n=118) OS analysis in the ITT population with ICI resistance (primary and secondary resistance) is described in statistical analysis 2 (n=219)

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: 219 HLA-A2 positive patients with advanced NSCLC who progressed after sequential or concurrent chemotherapy and Immune Checkpoint Inhibitor (either primary or secondary resistance) were randomized. Safety analysis was done in overall population. Primary efficacy analysis was carried out from stratification factor in subgroup of patients with ICI secondary resistance (who had a disease progression after ICI second line \> or equal to 12 weeks)

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS), measured over Approx. 24 months.

Overall survival is the time from randomization to death from any cause. It is a time-to-event endpoint: each participant contributes information about how long they were followed and whether death was observed during that follow-up. Participants who have not had the event by the end of their observed follow-up can still contribute information without being treated as if their eventual outcome were known.

Overall survival is used because it captures survival time directly and does not depend on a judgment about whether a tumor has progressed. The all-cause definition also avoids requiring the cause of death to be assigned to the cancer or treatment. In this trial, the registry gives an approximate 24-month timeframe and identifies an overall-survival analysis in patients with ICI secondary resistance. It also describes an ITT analysis including primary and secondary ICI resistance.

The reported group summaries are medians with 95% confidence intervals, while the comparison is expressed as a hazard ratio with its own 95% confidence interval. These quantities answer different questions: a median describes the time by which half of the relevant group has experienced the event, whereas a hazard ratio compares event rates over follow-up under a time-to-event model. Neither alone provides each participant’s survival time or guarantees that the relative comparison is constant throughout follow-up. The registry record does not report the assessment of that modeling assumption.

7. Results Posted in the Registry

• OSE2101 vs Docetaxel or Pemetrexed: hazard ratio 0.59 (95% CI 0.38–0.91), reported p < 0.050 (as reported in the registry).
• OSE2101 vs Docetaxel or Pemetrexed: hazard ratio 0.86 (95% CI 0.62–1.19), reported p = 0.360 (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.

Post-Progression Survival

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (t-test, 2 sided); it was not recalculated.
• Result:
  • OSE2101 vs Docetaxel or Pemetrexed: hazard ratio 0.46 (95% CI 0.27–0.79), reported p = 0.004 (as reported in the registry).
  • OSE2101 vs Docetaxel or Pemetrexed: hazard ratio 0.57 (95% CI 0.39–0.83), reported p = 0.004 (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 OSE2101 vs Docetaxel or Pemetrexed than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for OSE2101 vs Docetaxel or Pemetrexed than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Time to Worsening ECOG PS

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (t-test, 2 sided); it was not recalculated.
• Result:
  • OSE2101 vs Docetaxel or Pemetrexed: hazard ratio 0.43 (95% CI 0.23–0.80), reported p = 0.006 (as reported in the registry).
  • OSE2101 vs Docetaxel or Pemetrexed: hazard ratio 0.57 (95% CI 0.35–0.92), reported p = 0.020 (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 OSE2101 vs Docetaxel or Pemetrexed than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for OSE2101 vs Docetaxel or Pemetrexed than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Disease Control Rate (DCR) at 6 Months

• 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:
  • Docetaxel or Pemetrexed: 23.7% (9/38) versus 24.4% (19/78) on control; risk difference -0.7 percentage points (95% CI -17.2 to 15.9), risk ratio 0.97 (0.49–1.94); unadjusted two-proportion z-test p = 0.936.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Docetaxel or Pemetrexed and the control. The interval gives the range of effects that remain compatible with the data.

The available secondary results show a generally favorable direction for OSE2101 on the reported time-to-event comparisons: the post-progression survival and time-to-worsening ECOG performance status hazard-ratio estimates are below 1 in both reported populations. Their confidence intervals and p-values are listed with those outcomes. These results are supportive in direction, but they are secondary to overall survival and do not independently establish the primary endpoint finding.

Disease control at 6 months presents a less distinct pattern. The group-level estimates are close, and the recalculated unadjusted comparison is imprecise. The sponsor’s odds-ratio analyses also have p-values that do not indicate a clear difference. These distinct analyses should not be collapsed into a single effect measure.

The registry lists quality-of-life changes and progression-free survival among the secondary outcomes, but numerical results for those measures are not reported in the registry record. The record lists 12 outcome measures and 41 analyses, while the full analysis plan and multiplicity approach are not reported. When many outcomes or analyses are considered, some apparently notable findings may arise by chance even if no corresponding effect exists. Unless an analysis hierarchy or adjustment strategy is documented, individual secondary p-values carry less weight than a prespecified primary analysis. The secondary findings are best viewed as context for the primary result, not as independent confirmation of it.

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

• Post-Progression Survival
• Time to Worsening ECOG PS
• Mean Changes in Functional Subscales of European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Core Questionnaire (QLQ-C30) From Baseline to Treatment Discontinuation in Patients With ICI Secondary Resistance
• Mean Changes in Symptoms Subscales of the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Core Questionnaire (QLQ-C30) From Baseline to Treatment Discontinuation in Patients With ICI Secondary Resistance
• Mean Changes in Symptoms Subscales of European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Core Lung Cancer (QLQ-LC13) From Baseline to Treatment Discontinuation in Patients With ICI Secondary Resistance
• Disease Control Rate (DCR) at 6 Months
• Progression Free Survival (PFS) in Patients With ICI Secondary Resistance

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is time to death from any cause after randomization. The registry reports median survival for each arm and a hazard ratio for the comparison. A median time is the point at which the estimated survival experience reaches the halfway mark. A confidence interval around a median communicates uncertainty in that estimate. The registry reports these medians and intervals for the secondary-resistance population, but does not report the method used to estimate them.

What does a hazard ratio describe?

As a general statistical explanation, a hazard is the instantaneous rate of an event among participants who have not yet experienced it. A hazard ratio compares that rate between groups over follow-up. A value below 1 points toward a lower estimated event hazard in the first-named group relative to the comparator; it does not directly state how much longer an individual will live. Interpretation as a single constant relative comparison relies on modeling assumptions. Whether those assumptions were assessed is not reported in the registry record.

The sponsor-reported hazard ratios are 0.59 for the analysis in patients with ICI secondary resistance and 0.86 for the ITT population with primary and secondary resistance. Their 95% confidence intervals are 0.38 to 0.91 and 0.62 to 1.19, respectively. The registry lists “t-test, 2 sided” as the method for these analyses. That label is reproduced as reported; the registry does not explain the underlying time-to-event analysis procedure, how censoring was handled, or how the listed method relates to the hazard-ratio estimate.

How should the p-values be read?

A p-value describes how compatible the observed data are with a specified null hypothesis under the analysis model. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The primary secondary-resistance analysis is reported with p <0.05; the broader ITT analysis is reported with p=0.36. These values should be read alongside the effect estimates and confidence intervals, not as stand-alone conclusions. The significance threshold used in the trial’s original analysis plan is not reported in the registry record.

The reported 95% confidence interval gives a range of values compatible with the observed data under the interval procedure and its assumptions. For the first overall-survival analysis, the interval excludes 1, while the interval for the broader ITT analysis includes 1. A confidence interval does not assign a 95% probability to the specific realized interval containing the true value. It indicates the uncertainty and range of estimates supported by the data under the procedure.

What do the secondary analyses add?

The registry lists post-progression survival, time to worsening ECOG performance status, disease control rate, quality-of-life measures, and progression-free survival among secondary outcomes. For the available time-to-event outcomes, hazard ratios are reported for both the secondary-resistance and broader resistance populations. Disease control is a binary outcome, expressed as the proportion meeting the disease-control definition at the specified assessment time. Its comparison can be summarized using a risk difference or odds ratio, which are distinct measures and need not have the same numerical scale.

Some recalculated binary comparisons are simple unadjusted calculations from the posted summary statistics; they are not adjusted estimates from the sponsor’s analysis. In particular, the recalculated disease-control risk difference is -0.007, with a 95% confidence interval from -0.172 to 0.159 and an unadjusted p-value of 0.936. The sponsor reports odds-ratio analyses with different point estimates and p-values. These are different measures and may reflect different analysis populations or methods; the registry does not fully explain their relationship. The sponsor’s reported results remain the appropriate account of its own analysis.

For time-to-event outcomes, planning calculations are expressed as total events; for the binary outcome, they are expressed as participants per arm. Such calculations describe hypothetical planning under stated assumptions, not the statistical power achieved by the completed trial. The trial’s original sample-size justification and planning assumptions are not reported in the registry record.

The registry identifies an overall count of 41 analyses, but the complete analysis plan, endpoint hierarchy, multiplicity adjustment strategy, interim analysis plan, missing-data procedures, and detailed analysis populations are not reported in the registry record. Those omissions constrain assessment of whether each reported test was prespecified and how multiple comparisons were handled.

11. Confidence Intervals and Precision

Precision is reflected in the width of a confidence interval: narrower intervals indicate less uncertainty about the estimate, while wider intervals permit a broader range of plausible values. The primary secondary-resistance estimate has a hazard ratio of 0.59 and a 95% confidence interval from 0.38 to 0.91. The interval does not include 1, but its range still allows meaningfully different relative hazard estimates. The reported p-value is p <0.05.

The broader ITT analysis has a hazard ratio of 0.86, with a 95% confidence interval from 0.62 to 1.19. This interval includes 1 and extends to estimates on either side of it. Its p-value is 0.36. The distinction between the analyses is not simply a difference in point estimates: the populations differ, and the interval for the broader population does not rule out no relative difference under the reported model.

The median estimates also have uncertainty. For the secondary-resistance analysis, the OSE2101 median is 11.1 months, with an interval from 8.6 to 13.5 months; the chemotherapy median is 7.5 months, with an interval from 4.7 to 10.3 months. Confidence intervals for the separate medians are not themselves a formal test of the difference between groups. The hazard-ratio analysis is the registry’s reported comparative measure.

Confidence intervals quantify statistical uncertainty under an analysis and its assumptions; they do not capture every limitation, including possible bias, incomplete outcome ascertainment, or choices in analysis. The registry does not report enough detail to evaluate all such sources. The intervals therefore support a calibrated reading of the estimates, not certainty about the true effect or its applicability beyond the studied populations.

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 for time-to-event outcomes are total event counts, not participants per arm. The disease-control planning figures are participants per arm. The calculations use a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. They use the observed effect as a starting point and consider detecting the stated fraction of that effect. The original trial’s design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the planned number of events or participants under the calculation assumptions for an 80% chance of rejecting the null hypothesis if the assumed effect were true. The 90% column gives the corresponding planning target for a 90% chance. These are conditional planning probabilities, not guarantees. For example, the planning calculation for 100% of the observed primary hazard-ratio effect gives 113 total events for 80% power and 151 for 90% power. For 75% of that observed effect, the corresponding event targets are 201 and 269. For 50%, they are 452 and 604.

Smaller effects are harder to distinguish from random variation, so the planning calculations require more information: more total events for time-to-event endpoints or more participants per arm for binary endpoints. The values are planning illustrations based on the observed effect, not observed power. They do not show that the completed study had those event totals or that its original design used these assumptions.

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 = 0.590, so ln(HR) = -0.5276.
2. d = 4 × (1.96 + 0.842)² / [ln(0.590)]² = 112.8, rounded up to 113 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 <- 0.59                              # 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(2654587)
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 = 113 (80% power), 151 (90% power)
 75% of observed effect: events = 201 (80% power), 269 (90% power)
 50% of observed effect: events = 452 (80% power), 604 (90% power)
Simulated power with 113 events: 0.807 (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 179 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 219 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 <- 7.5          # control-arm median, months (converted from the posted unit)
hr          <- 0.59          # 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: 126         p (1-sided)   0.0001
  Events: 51        ~HR at bound   0.3520
   Month: 17    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.59   0.0338
   IA 2: 67%                   Z   2.5114
      N: 180         p (1-sided)   0.0060
 Events: 102        ~HR at bound   0.6067
   Month: 27    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.59   0.5603
       Final                   Z   1.9930
      N: 180         p (1-sided)   0.0231
 Events: 152        ~HR at bound   0.7234
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.59   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 sponsor analyses for overall survival in two populations. In patients with ICI secondary resistance, the sponsor reports a hazard ratio of 0.59, a 95% confidence interval from 0.38 to 0.91, and p <0.05. In the ITT population with primary and secondary ICI resistance, the sponsor reports a hazard ratio of 0.86, a 95% confidence interval from 0.62 to 1.19, and p=0.36. These are the sponsor’s stated results and should be distinguished from simple calculations based on summary statistics.

The posted disease-control summary also permits an unadjusted recalculation of a risk difference. That recalculation is -0.007, with a 95% confidence interval from -0.172 to 0.159 and an unadjusted p-value of 0.936. The sponsor’s disease-control analyses are reported as odds ratios, including 1.09 with p=0.87 and 0.73 with p=0.37. An odds ratio and a risk difference are different statistical measures, and the registry does not fully explain why the sponsor lists two analyses. The discrepancy flag is false; where a recalculation differs in measure or method, it should not be presented as replacing the sponsor’s analysis.

The registry labels the reported comparison method “t-test, 2 sided” for the hazard-ratio analyses. It does not provide sufficient detail to reconstruct the sponsor’s analysis procedure from that label alone.

16. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple secondary outcomes and reports 41 analyses. Testing several outcomes or populations increases the chance that at least one result will appear statistically notable by chance. Statistical methods can address this through a prespecified hierarchy, adjustment of significance thresholds, or other procedures. Which approach, if any, was planned or applied is not reported in the registry record.

The registry identifies overall survival as the primary endpoint, but it includes two reported overall-survival analyses: one in patients with ICI secondary resistance and one in the broader ITT population with primary and secondary resistance. The primary endpoint description specifies the secondary-resistance population, while the broader analysis is described separately. The complete protocol, analysis plan, endpoint hierarchy, and rationale for the two populations are not reported in the registry record. Without those details, it is difficult to determine how the analyses were prioritized or whether the reported populations were prespecified.

The registry does not report an interim analysis plan, stopping boundaries, or any interim efficacy assessment. Although the trial status is terminated, the reason for termination and whether it was related to efficacy, safety, recruitment, or another consideration are not reported in the registry record. Termination alone does not establish why the trial ended or whether it affected the analyses.

Missing-data methods, censoring rules, handling of protocol deviations, and detailed analysis-population definitions are also not reported in the registry record. These details matter because survival analyses depend on follow-up and censoring, while quality-of-life outcomes can be affected by missing assessments. The available record supports interpretation of the posted estimates, but not a full audit of analysis choices or multiplicity control.

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: general physical health deterioration (6.2%), cytokine release syndrome (4.3%), febrile neutropenia (3.8%), malignant neoplasm progression (3.8%), diarrhea (2.5%).

The registry reports serious adverse event counts among 139 participants at risk in the OSE2101 group and 80 at risk in the docetaxel or pemetrexed group. Serious adverse events affected 47 and 33 participants, respectively. Other adverse events affected 125 of 139 in the OSE2101 group and 67 of 80 in the chemotherapy group. Death counts are reported as 116 and 65, respectively.

The registry lists general physical health deterioration at 6.2%, cytokine release syndrome at 4.3%, febrile neutropenia at 3.8%, malignant neoplasm progression at 3.8%, diarrhea at 2.5%, and gastrooesophageal reflux disease at 2.5% among the top serious adverse event terms. These entries describe recorded events; the registry information does not establish that an event was caused by study treatment. The record does not fully explain the event coding, exposure duration, overlap between participants with different events, or the basis of the percentages.

Counts from arms with different numbers at risk should not be compared as if they were rates without accounting for the denominators and follow-up. The death counts are not identified as treatment-related deaths in the posted figures. A full safety comparison would require event definitions, timing, severity, attribution, and follow-up details, several of which are not reported in the registry record.

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

The registry reports an overall-survival hazard ratio below 1 for the ICI secondary-resistance analysis, with a confidence interval that excludes 1 and p <0.05. The reported medians are also longer in the OSE2101 group than in the chemotherapy group for this analysis. In the broader ITT population with primary and secondary resistance, the hazard-ratio estimate is closer to 1, its confidence interval includes 1, and the reported p-value is 0.36. The two population-specific analyses therefore do not give identical levels of evidence.

These findings describe group-level estimates in the populations and analyses recorded in the registry. They do not establish what would happen to a particular patient, do not guarantee an outcome, and do not show that every subgroup would have the same result. A hazard ratio is not an absolute survival probability, and the reported medians are not individual predictions.

The record also cannot establish why the trial was terminated, whether all planned analyses were completed, or how missing data and multiple comparisons were handled. The available secondary outcome results provide context, but secondary analyses are less definitive when the analysis hierarchy and multiplicity strategy are not reported. Safety counts describe events recorded in the registry and do not by themselves establish causation or permit a complete comparison of risk.

Accordingly, the statistical reading is limited to the estimates, confidence intervals, and methods recorded. It is not medical advice and does not establish a treatment recommendation. Interpretation beyond these trial populations requires evidence and clinical judgment not contained in the registry record.

19. Why This Trial Matters Statistically

This trial illustrates why the population attached to an endpoint matters. The registry describes an overall-survival analysis in patients with ICI secondary resistance and a separate analysis in a broader ITT population with primary and secondary resistance. The estimated hazard ratios and their precision differ. Treating them as if they described the same population would obscure an important part of the result.

It also shows the value of reading an estimate together with its confidence interval and p-value. The secondary-resistance analysis has a hazard ratio below 1 and an interval that excludes 1, while the broader ITT interval includes 1. A point estimate alone would hide that difference in uncertainty. Conversely, a p-value alone would not convey the range of effect estimates compatible with the data.

The secondary endpoints provide another statistical lesson: multiple outcomes can help describe a study, but they create interpretive challenges when the prespecified hierarchy and multiplicity plan are unavailable. Results across time-to-event, binary, and quality-of-life measures cannot be treated as interchangeable. Their scales, populations, and inferential roles differ.

Finally, the registry’s planning illustrations are not observed power and do not recover the original trial design. They show how planning targets depend on the assumed effect and whether the endpoint is based on events or participants. The missing original design justification, termination reason, and analysis details limit how completely the trial can be reconstructed from the registry record.

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 identifies the trial as terminated but does not report why it ended. Without that information, it is not possible to assess whether termination was related to recruitment, safety, efficacy, resources, or another reason. The record does not describe the timing or decision process behind termination.

The overall-survival result is reported for a defined secondary-resistance population and for a broader ITT population. The registry gives the population descriptions but does not provide the full protocol or analysis plan explaining the relationship between these analyses. The reported arm sample sizes for the primary analysis are 80 and 38, while the broader ITT population is described as n=219; detailed participant flow and the reasons participants may be excluded from a particular analysis are not reported in the registry record.

Important design and analysis details are absent, including the allocation ratio and procedure, concealment, stratification, censoring rules, proportional-hazards assessment, missing-data methods, and original sample-size justification. The listed method for the hazard-ratio comparisons is “t-test, 2 sided,” without enough explanation to reproduce the analysis. The registry also does not report the multiplicity strategy despite listing 41 analyses.

Several secondary outcomes are named, but numerical results for quality-of-life changes and progression-free survival are not reported in the registry record. Safety summaries lack detail about event attribution, duration of exposure, and overlap among event categories. These omissions make the posted estimates interpretable as reported, but restrict independent evaluation of the full trial conduct and analysis.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: OSE2101 Versus Chemotherapy in HLA-A2 Positive Patients With Advanced NSCLC After Immune Checkpoint Inhibitor Failure — NCT02654587.
• Registry record last updated 2024-02-02; status last verified 2024-01.
• Results first posted 2024-02-02.
• 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

NCT02654587 is a randomized, open-label, parallel-group phase III trial comparing OSE2101 with docetaxel or pemetrexed in HLA-A2 positive patients with advanced NSCLC after immune checkpoint inhibitor failure. The registry records enrollment of 219 and a terminated status. Overall survival was the primary endpoint, with an analysis in patients with ICI secondary resistance and a separate analysis in the broader ITT population with ICI resistance.

The sponsor reports a hazard ratio of 0.59 for the secondary-resistance analysis and 0.86 for the broader ITT analysis. The associated confidence intervals and p-values indicate different levels of precision and evidence across these populations. Secondary outcome findings are mixed in strength, and the registry does not report numerical results for every listed outcome. Reasons for termination, the complete analysis plan, multiplicity handling, and several other design and analysis details are not reported in the registry record.