Durvalumab and Tremelimumab ± Platinum-Based Chemotherapy in Patients With Metastatic Squamous or Non-Squamous NSCLC: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Durvalumab and Tremelimumab ± Platinum-Based Chemotherapy in Patients With Metastatic Squamous or Non-Squamous NSCLC (NCT03057106), a randomized phase 2 study. The primary outcome is Overall Survival.

2. Clinical Question

The trial addresses whether adding platinum-based chemotherapy to durvalumab and tremelimumab is associated with a different overall survival pattern in this population. The primary outcome is time from randomization to death, assessed over the stated 33 months. Random assignment supports a group-level comparison, while the open-label design means treatment assignment was not masked. The registry reports the overall survival results and a separate progression-free survival result. It lists objective response rate as another secondary outcome, but its detailed result is not reported in the registry record.

3. Trial Design

This was a randomized, open-label, parallel-group phase 2 trial sponsored by the Canadian Cancer Trials Group. The registry records enrollment of 301. The posted outcome groups contain 150 patients in the durvalumab-and-tremelimumab group and 151 in the platinum-based chemotherapy plus durvalumab-and-tremelimumab group. These outcome-group counts should be read as recorded, not as a substitute for information about every analysis population.

One group received durvalumab every 4 weeks until progression and tremelimumab every 4 weeks for 4 doses. The other received 4 cycles of platinum plus gemcitabine or pemetrexed with durvalumab and tremelimumab, given every 3 weeks for 4 cycles. The registry then describes maintenance durvalumab every 4 weeks until progression for squamous cell cancer, and pemetrexed plus durvalumab every 4 weeks until progression for non-squamous cell cancer.

The primary outcome was overall survival, defined as time from randomization to death. The primary outcome timeframe is recorded as 33 months. The registry identifies the primary population as intention-to-treat and notes that it included patients who refused the study treatments (4 patients). Progression-free survival using RECIST 1.1 and objective response rate using RECIST 1.1 and iRECIST are listed as secondary outcomes.

The trial status is active, not recruiting. The registry records a start date of 2017-03-28 and a primary completion date of 2020-02-24. It does not report the original sample-size justification, detailed allocation procedures, or whether interim analyses were planned.

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

Analysis population for the primary outcome, as defined in the registry: ITT population, included patients who refused the study treatments (4 patients)

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: ITT population, included 4 patients who refused the study treatments.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival, measured over 33 months.

Overall survival measures the time from randomization until death. It is a time-to-event outcome: each participant contributes information about how long they were followed and whether death was recorded during that period. The registry reports the median in months, with a 90% confidence interval, at the stated 33-month timeframe.

As a general statistical concept, the median survival time is the point at which the estimated survival experience reaches the halfway mark. It describes the timing of events in the group; it is not an average lifespan and does not mean that every participant experienced an event at that time. Time-to-event methods can use information from participants whose follow-up ends before an event is observed, but the registry record does not describe censoring rules or the handling of missing follow-up.

Overall survival is used in clinical trials because it is a direct measure of time until death and does not depend on defining radiographic progression. The between-group hazard ratio provides a separate summary of the relative event rates over follow-up, subject to the model and analysis choices. It complements the reported medians rather than replacing them. Neither summary by itself describes every individual outcome, and estimates remain uncertain, as reflected by their confidence intervals.

7. Results Posted in the Registry

• Durvalumab and Tremelimumab vs Platinum Based Chemotherapy + Durvalumab + Tremelimumab: hazard ratio 0.88 (95% CI 0.67–1.16), reported p = 0.460 (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 Using RECIST 1.1

• 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:
  • Durvalumab and Tremelimumab vs Platinum Based Chemotherapy + Durvalumab + Tremelimumab: hazard ratio 0.67 (95% CI 0.52–0.88), 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 Durvalumab and Tremelimumab vs Platinum Based Chemotherapy + Durvalumab + Tremelimumab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists progression-free survival using RECIST 1.1 and objective response rate using RECIST 1.1 and iRECIST as secondary outcomes. Detailed results are posted for progression-free survival, whose reported comparison favors the chemotherapy-containing group by the direction of the hazard-ratio estimate and has an interval that excludes 1. A detailed objective response rate estimate is not reported in the registry record. Thus, the secondary results do not form a complete set of reported outcome estimates.

The progression-free survival finding differs in statistical clarity from the primary overall-survival result, which has an interval including 1 and a p-value of 0.46. This contrast can be informative: the posted data show evidence of a group difference for progression-free survival, while the primary survival comparison remains uncertain. It does not establish that the secondary outcome predicts an overall-survival difference, and it does not turn the primary result into a positive finding.

Secondary outcomes generally carry less confirmatory weight than a prespecified primary outcome, particularly when several outcomes are tested. Testing multiple hypotheses raises the chance that at least one result will appear statistically notable by chance if no adjustment or hierarchy is used. The registry record does not report whether multiplicity was controlled, whether outcomes were tested in a prespecified sequence, or how missing secondary results should be interpreted. The progression-free survival result should therefore be read as a secondary signal in the context of the full outcome set, rather than as stand-alone confirmation. It is not possible to assess the complete pattern across all listed secondary endpoints because the objective response rate result is not reported in the registry record.

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

• Progression-free Survival Using RECIST 1.1
• Objective Response Rate Using RECIST 1.1 and iRECIST

9. Statistical Interpretation

10. Statistical Methods Explained

What does the median summarize?

For an outcome measured as time until an event, the median is a summary of the estimated time-to-event experience in a group. Here, the event for the primary outcome is death, and the registry defines the time origin as randomization. A median is not the arithmetic average of participants' observed follow-up times. It also does not imply that all participants experienced the event at that point. Follow-up can end without an event being recorded, and time-to-event methods ordinarily account for that incomplete observation. The registry record does not describe its censoring rules or the specific estimation method used for the medians.

What does the hazard ratio compare?

The registry reports a hazard ratio for durvalumab and tremelimumab versus platinum-based chemotherapy plus durvalumab and tremelimumab. In general, a hazard ratio compares event rates over follow-up between groups, rather than comparing their median times directly. A value of 1 represents equal relative event rates in the summary. A value below 1 indicates a lower estimated event rate for the first group in the stated comparison; a value above 1 indicates a higher estimated rate. These descriptions concern the estimate, not a guaranteed pattern for each participant. The registry reports the sponsor's hazard ratios and log-rank analyses; the specific assumptions underlying the hazard-ratio model are not reported in the registry record.

What does the log-rank test do?

As a general statistical method, the log-rank test compares time-to-event patterns between groups while accounting for the timing of events and the available follow-up. It tests a null hypothesis of no group difference under the test's framework. The p-value indicates how unusual a test result at least as extreme as the observed one would be if that null hypothesis held, subject to the analysis assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of an effect. The registry identifies the log-rank method for the reported comparisons.

How should confidence intervals be read?

A confidence interval expresses statistical uncertainty around an estimate. In general, a wider interval indicates less precision, while a narrower interval indicates greater precision, all else equal. It is not a range that contains a fixed proportion of individual participants' outcomes. The overall-survival median estimates are accompanied by 90% confidence intervals, and the sponsor-reported overall-survival hazard ratio also has a 90% confidence interval. For progression-free survival, the median estimates have 90% confidence intervals, while the sponsor-reported hazard-ratio analysis gives a 95% confidence interval. Confidence levels should therefore be considered when reading each estimate and should not be treated as interchangeable.

What is the role of randomization and masking?

Randomization assigns participants to groups by chance and is intended to support a fair comparison by reducing systematic differences at allocation. It does not guarantee that every characteristic is balanced, nor does it prevent differences in care or outcome assessment after allocation. This trial was open-label, meaning masking was not used. Knowledge of assigned treatment can affect participant experience, clinical decisions, follow-up, or assessments, particularly for outcomes involving judgment. Death is a comparatively direct endpoint, but the registry's open-label design remains relevant context for interpreting the trial as a whole. Details of the randomization sequence and allocation concealment are not reported in the registry record.

What is the analysis population?

The primary outcome is reported for the intention-to-treat population, including patients who refused the study treatments (4 patients). In general, an intention-to-treat analysis compares participants according to their assigned groups, preserving the randomized comparison even when treatment is not taken as planned. The registry does not provide detailed information about participant-level exclusions, missing outcome data, censoring, or sensitivity analyses. The outcome group counts are recorded separately from this population description, and should not be assumed to answer every question about analysis inclusion.

How do the secondary outcome methods fit?

Progression-free survival is a time-to-event outcome reported using RECIST 1.1. The registry also lists objective response rate using RECIST 1.1 and iRECIST. These criteria provide frameworks for classifying tumor response or progression; the detailed assessment schedule and adjudication procedures are not reported in the registry record. The registry posts a progression-free survival hazard ratio and log-rank result, but detailed objective response rate results are not reported. Secondary outcomes can add context about different aspects of disease course, but their interpretation depends on the analysis plan and on how multiple outcomes were considered.

The registry does not report whether proportional-hazards assumptions were evaluated, the precise handling of missing data, the detailed statistical analysis plan, or planned interim analyses. These details can matter for interpreting time-to-event estimates and the amount of confirmatory weight to place on any single result.

11. Confidence Intervals and Precision

The overall-survival median estimate was 14.1 months in the durvalumab-and-tremelimumab group, with a 90% confidence interval from 10.6 to 18.3 months. In the chemotherapy-containing group, the median was 16.6 months, with a 90% confidence interval from 12.6 to 19.1 months. These intervals show uncertainty around the group estimates; they do not describe the range of outcomes for individual participants.

For the overall-survival comparison, the hazard ratio was 0.88 and the 90% confidence interval was 0.67 to 1.16. Because this interval includes 1, it is compatible with no relative event-rate difference as well as estimates on either side of that value. The p-value was 0.46. The point estimate alone therefore gives an incomplete account of the evidence: the interval's width and inclusion of 1 are central to judging precision and the strength of the comparison.

For progression-free survival, the median estimates also have 90% confidence intervals. The sponsor-reported hazard ratio has a 95% confidence interval from 0.52 to 0.88, which does not include 1. Confidence intervals at different confidence levels are not directly interchangeable, and a confidence interval is not a probability statement about the true value. Overall, the estimates indicate uncertainty of differing degrees across outcomes; the registry does not provide a formal assessment of precision or a target width for the intervals.

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 display gives event totals for the time-to-event outcomes, not participants per arm. Its rows consider planning for 100%, 75%, or 50% of the observed effect; its columns give the total events associated with 80% or 90% power under the listed assumptions. For overall survival, planning for 100% of the observed effect corresponds to 1,922 total events for 80% power and 2,572 total events for 90% power. For 75% of the observed effect, the corresponding totals are 3,416 and 4,573; for 50%, they are 7,685 and 10,288. These are planning illustrations, not a count of events observed in this trial.

The stated 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 registry's own original design assumptions and sample-size justification are not reported in the registry record. These calculations are based on the observed effect and are not observed power.

In plain language, the 80% power column represents the event total that the planning calculation associates with an 80% chance of detecting the specified effect under its assumptions; the 90% column represents the corresponding higher chance. Smaller effects require more participants or, for a time-to-event analysis, more events to detect under otherwise similar assumptions. The figures should not be read as proof that a future study will achieve its target, since actual power depends on the true effect, event accrual, follow-up, and analysis choices.

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 = 0.880, so ln(HR) = -0.1278.
2. d = 4 × (1.96 + 0.842)² / [ln(0.880)]² = 1,921.2, rounded up to 1,922 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 <- 0.88                              # 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(3057106)
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 = 1,922 (80% power), 2,572 (90% power)
 75% of observed effect: events = 3,416 (80% power), 4,573 (90% power)
 50% of observed effect: events = 7,685 (80% power), 10,288 (90% power)
Simulated power with 1,922 events: 0.805 (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 3,416 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 301 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 <- 14.1          # control-arm median, months (converted from the posted unit)
hr          <- 0.88          # 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: 2724         p (1-sided)   0.0001
  Events: 867        ~HR at bound   0.7772
    Month: 19    P(Cross) if HR=1   0.0001
              P(Cross) if HR=0.88   0.0338
    IA 2: 67%                   Z   2.5114
      N: 3416         p (1-sided)   0.0060
 Events: 1734        ~HR at bound   0.8864
    Month: 30    P(Cross) if HR=1   0.0060
              P(Cross) if HR=0.88   0.5603
        Final                   Z   1.9930
      N: 3416         p (1-sided)   0.0231
 Events: 2601        ~HR at bound   0.9248
    Month: 60    P(Cross) if HR=1   0.0250
              P(Cross) if HR=0.88   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's sponsor-reported overall-survival analysis uses a log-rank method and reports a hazard ratio of 0.88, a 90% confidence interval from 0.67 to 1.16, and a p-value of 0.46. The posted result is the primary reference for the trial's reported comparison. Its interval includes 1, and the p-value is not strong evidence against the null comparison.

The recalculations described alongside the posted results are simple, unadjusted calculations from summary statistics. They cannot reproduce every feature of a participant-level sponsor analysis, such as its handling of follow-up and analysis details. The registry record does not flag a discrepancy between the sponsor result and the recalculated result. Where a discrepancy is flagged, the sponsor's own analysis takes precedence; no such discrepancy is flagged for this trial. The sponsor-reported analysis should therefore be used when describing the trial's formal comparison, while any summary-based recalculation is best treated as a limited cross-check rather than a replacement.

The same distinction applies to the secondary progression-free survival analysis: the registry reports the sponsor's log-rank hazard ratio and associated interval, while a simple recalculation from summaries would not be an equivalent participant-level analysis. The detailed assumptions and diagnostics of the sponsor's analysis are not reported in the registry record.

16. Multiplicity, Missing Data and Analysis Populations

The registry lists more than one secondary outcome in addition to the primary overall-survival endpoint. When a study evaluates multiple outcomes or makes multiple statistical comparisons, the chance of observing at least one apparently strong result by chance can increase. Statistical approaches such as adjusting significance thresholds or using a prespecified testing hierarchy can address that issue, but the registry record does not report whether either approach was used. It also does not report a detailed multiplicity plan. The secondary progression-free survival finding should therefore be interpreted with that uncertainty in mind, rather than assumed to have the same confirmatory status as the primary endpoint.

The primary outcome population is described as intention-to-treat and includes patients who refused the study treatments (4 patients). This identifies an important feature of the analysis population, but does not explain how every participant was handled. The registry record does not report the number or pattern of missing outcome observations, censoring rules, methods for addressing missing data, or sensitivity analyses. It also does not report whether analyses used additional populations or how deviations from assigned treatment were handled.

These omissions limit assessment of robustness. For time-to-event outcomes, follow-up and censoring can affect the information available for estimating survival patterns. For response outcomes, the definition and availability of assessments can affect who contributes to an estimate. Without the analysis plan and missing-data details, readers cannot determine from the registry record whether alternative reasonable analytic choices would materially change the conclusions. This uncertainty does not invalidate the posted results, but it narrows what can be concluded about their sensitivity to analytic decisions.

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: other neoplasms benign, malignant and unspecified (12.2%), diarrhea (10.7%), colitis (9.4%), anemia (8.1%), lung infection (8.1%).

The registry reports serious adverse event figures for 149 patients at risk in the durvalumab-and-tremelimumab group, of whom 84 were affected, and 148 at risk in the chemotherapy-containing group, of whom 102 were affected. It also reports other adverse event figures in which 149 of 149 and 148 of 148, respectively, were affected. These are the recorded at-risk and affected counts; they do not by themselves establish that an event was caused by treatment or indicate its duration, severity beyond the serious-event classification, or clinical course.

The most frequently listed serious-event terms include other neoplasms benign, malignant and unspecified (12.2%), diarrhea (10.7%), colitis (9.4%), anemia (8.1%), lung infection (8.1%), and fever (7.4%). These percentages are presented as recorded in the registry. The record also reports deaths of 80 and 77 in the respective groups, but does not provide further detail in these figures about cause or attribution. Adverse-event summaries are descriptive and should not be treated as a randomized comparison of each specific event unless the relevant analysis is reported.

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

The posted primary results describe median overall survival of 14.1 months in the durvalumab-and-tremelimumab group and 16.6 months in the group that also received platinum-based chemotherapy. The sponsor-reported hazard ratio is 0.88, with a 90% confidence interval from 0.67 to 1.16 and a p-value of 0.46. The interval includes 1, so the trial record does not provide clear statistical evidence of an overall-survival difference between the groups. The result does not establish that the treatments are equivalent or that any possible difference is absent.

The secondary progression-free survival analysis presents a more statistically distinct comparison, but it remains a secondary result, and the objective response rate result is not reported in the registry record. The registry does not report whether multiplicity was addressed, nor does it provide all details needed to assess missing data, follow-up, or analysis assumptions. Those gaps matter when weighing the strength and completeness of the evidence.

Trial-level summaries do not predict an individual participant's outcome. A median is not a personal prognosis, and a group hazard ratio does not describe the exact experience of every participant. The posted statistics also do not establish how results might apply outside the registered population, treatment regimens, and study setting. They describe what was recorded for this randomized comparison, with uncertainty reflected in the confidence intervals and limits arising from the reporting details.

No conclusion about an individual's care follows from these statistics alone. The results are a statistical account of a trial comparison, not medical advice or a guarantee about treatment response.

19. Why This Trial Matters Statistically

This trial illustrates why a primary endpoint and secondary endpoints need to be considered separately. The overall-survival comparison is the main question, and its hazard-ratio interval includes 1. The progression-free survival analysis has a different statistical pattern, but its secondary status and the unreported multiplicity approach affect how strongly it can support a broader conclusion. One outcome should not be used as a substitute for another simply because its result is more statistically distinct.

It also shows why estimates need to be read together with their uncertainty. A point estimate gives a compact summary, but the confidence interval indicates which values remain compatible with the data under the analysis. Here, the primary result is not precise enough to rule out relative event-rate estimates on either side of 1. The p-value alone cannot convey the size, direction, or clinical importance of a possible difference.

Finally, the registry record makes visible the boundary between reported results and unavailable design or analysis details. It identifies the trial as randomized and open-label, gives the primary population, and posts sponsor-reported comparisons, while leaving the original sample-size justification, missing-data procedures, and multiplicity plan unreported. Distinguishing what the data show from what the record does not establish is essential for sound interpretation.

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 does not report the original sample-size justification or detailed design assumptions. The posted sample-size planning calculations are illustrations based on the observed effect, not evidence of the trial's original planning or observed power. The record also does not report the event count used in the primary comparison, so the event-based planning figures cannot be directly related to the trial's accrued information from the record alone.

The trial was open-label, and the registry does not provide details about allocation concealment, treatment adherence, crossover, or subsequent therapies. These details could be relevant to understanding how the assigned groups were experienced and followed, but their absence does not establish that any particular issue occurred. The record also does not report the proportional-hazards assessment, censoring rules, missing-data procedures, or sensitivity analyses for the time-to-event outcomes.

The primary outcome is reported at a 33-month timeframe, but the registry record does not give detailed follow-up distributions or a fuller account of outcome maturity. Objective response rate is listed as a secondary outcome, yet its detailed results are not reported. The record does not describe whether multiplicity was controlled or provide the full statistical analysis plan. These gaps limit independent assessment of the completeness and robustness of the posted estimates.

Finally, the results pertain to the registered metastatic squamous or non-squamous non-small cell lung cancer population and the described treatment regimens. The registry record alone does not establish generalizability to different populations, schedules, or clinical settings.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Durvalumab and Tremelimumab ± Platinum-Based Chemotherapy in Patients With Metastatic Squamous or Non-Squamous NSCLC — NCT03057106.
• Registry record last updated 2026-09-01; status last verified 2026-02.
• Results first posted 2022-02-09.
• 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

NCT03057106 is a randomized, open-label, parallel-group phase 2 trial sponsored by the Canadian Cancer Trials Group. It enrolled patients with metastatic stage IV squamous or non-squamous non-small cell lung cancer and compared durvalumab plus tremelimumab with platinum-based chemotherapy plus the same immunotherapy drugs.

The primary outcome was overall survival, measured from randomization to death. The sponsor-reported comparison gives a hazard ratio of 0.88 with a 90% confidence interval from 0.67 to 1.16 and a p-value of 0.46; the interval includes 1, so the result does not establish a clear overall-survival difference. Progression-free survival is reported as a secondary outcome with a different statistical pattern. Objective response rate is listed, but its detailed result is not reported in the registry record.

The record is active, not recruiting. Details of the original sample-size justification, multiplicity plan, and missing-data procedures are not reported in the registry record.