LUX-Lung 7: A Phase IIb Trial of Afatinib(BIBW2992) Versus Gefitinib for the Treatment of 1st Line EGFR Mutation Positive Adenocarcinoma of the Lung: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for LUX-Lung 7: A Phase IIb Trial of Afatinib(BIBW2992) Versus Gefitinib for the Treatment of 1st Line EGFR Mutation Positive Adenocarcinoma of the Lung (NCT01466660), a randomized phase 2 study. The primary outcome is Progression-free Survival.

2. Clinical Question

The clinical question concerns a comparison of two first-line drug treatments in a defined group of patients with advanced lung adenocarcinoma and an EGFR mutation. The trial was randomised and open-label, so treatment assignment was random but participants and clinicians were not masked to the assigned treatment.

The main comparison is a time-to-event analysis rather than a simple comparison of the proportion of participants with an event at a fixed time. It uses the timing of progression or death and accounts for participants whose follow-up ended without a recorded event by censoring them. For the primary analysis, disease progression was assessed by independent central imaging review using RECIST version 1.1. At the final analysis, progression status and date were determined by investigators.

The registry record reports median progression-free survival of 12.78 months for afatinib and 11.17 months for gefitinib. The reported hazard ratio for afatinib versus gefitinib was 0.822, with a 95% confidence interval from 0.655 to 1.032 and p = 0.089. Those estimates describe the posted result and its uncertainty; they do not by themselves establish that either treatment is preferable for an individual patient.

3. Trial Design

LUX-Lung 7 was a randomised, open-label, parallel-group Phase IIb trial comparing afatinib with gefitinib as first-line treatment for patients with EGFR mutation-positive advanced adenocarcinoma of the lung. The registry lists an enrollment of 319. The treatment groups were afatinib, with 160 participants in the primary analysis, and gefitinib, with 159.

Participants were assigned to one of the two treatment groups, and each group received its assigned drug once daily. The trial had no masking. The parallel design means the registry describes separate groups receiving the two treatments, rather than a crossover in which participants would receive both in sequence. The record identifies Boehringer Ingelheim as the sponsor and lists the study as completed. Its recorded start date is 2011-12-13 and its primary completion date is 2016-04-08.

The primary endpoint was progression-free survival. The primary analysis population was the randomised set and included all patients randomised to receive treatment, whether treated or not. The primary endpoint description states that progression was primarily evaluated for the primary analysis through independent central imaging review according to RECIST version 1.1. Participants without a progression or death event were censored.

The registry also lists secondary outcomes including objective response, time to and duration of objective response, disease control and duration of disease control, tumour shrinkage, and health-related quality of life. The listed secondary outcome time frames include follow-up after drug administration. The record gives a final progression-free survival analysis cut-off date of 12 April 2019 and separately identifies cut-off dates for tumour shrinkage and health-related quality of life.

The registry record does not report the trial’s original sample-size justification or the complete statistical analysis plan. It also does not report whether interim analyses were planned or conducted. Those design details cannot be inferred from the posted results or from planning illustrations based on the observed estimates.

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: All subjects were screened for eligibility prior to participation in the trial. Subjects attended a specialist site which ensured that they (the subjects) strictly met all inclusion and none of the exclusion criteria. Subjects were not to be allocated to a treatment group if any of the entry criteria were violated.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Randomised set which included all patients randomised to receive treatment, whether treated or not.

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: Randomised set which included all patients randomised to receive treatment, whether treated or not.

6. Primary Endpoint

The registry lists the primary outcome as Progression-free Survival, measured over From first drug administration until 28 days after last drug administration + Follow-Up period for collecting information on disease progression or death, up to 2465 days..

Progression-free survival (PFS) measures the time from randomisation to disease progression or death, whichever occurred first. Participants with neither event were censored. This makes PFS a time-to-event endpoint: the analysis takes account of the timing of events and of follow-up that ends before an event is recorded, rather than treating all participants as if they had the same observation period.

The endpoint is useful in a trial of advanced cancer because it captures either documented progression or death as an event. The registry specifies that progression for the primary analysis was primarily evaluated by independent central imaging review according to RECIST version 1.1, using CT or MRI assessment. The description defines progression using changes in target lesions or the appearance of new lesions. At the final analysis, progression status and date were determined by investigator assessment.

The reported medians are 12.78 months for afatinib and 11.17 months for gefitinib. A median is a summary of the event-time distribution; it does not describe every participant’s experience or the full shape of that distribution. The hazard ratio adds a relative time-to-event comparison across follow-up, while its confidence interval conveys uncertainty. Neither summary alone gives a complete account of how individual event times or censoring were distributed. The registry record does not report further details about the pattern of censoring.

7. Results Posted in the Registry

• Afatinib vs Gefitinib: hazard ratio 0.82 (95% CI 0.66–1.03), reported p = 0.089 (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.

Objective Response Rate

• 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:
  • Gefitinib: 74.8% (119/159) versus 79.4% (127/160) on control; risk difference -4.5 percentage points (95% CI -13.7 to 4.7), risk ratio 0.94 (0.84–1.06); unadjusted two-proportion z-test p = 0.335.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Gefitinib and the control. The interval gives the range of effects that remain compatible with the data.

Disease Control

• 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:
  • Gefitinib: 93.7% (149/159) versus 94.4% (151/160) on control; risk difference -0.7 percentage points (95% CI -5.9 to 4.5), risk ratio 0.99 (0.94–1.05); unadjusted two-proportion z-test p = 0.802.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Gefitinib and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists several secondary outcomes, including objective response and disease control, as well as time to and duration of response or disease control, tumour shrinkage, and health-related quality of life. The posted numerical results include objective response and disease control. Across these outcomes, the reported group percentages are close, and the uncertainty intervals for the unadjusted contrasts include no difference. The sponsor’s reported logistic-regression analyses also have p-values that do not indicate strong evidence of a group difference under conventional testing.

These secondary outcomes complement, but do not replace, the primary progression-free survival analysis. Objective response and disease control describe response status, while progression-free survival incorporates the timing of progression or death. A similar pattern in secondary results may provide context, but it does not resolve uncertainty in the primary time-to-event comparison. Nor should these outcomes be combined into a single conclusion without a prespecified method.

The sponsor’s odds ratios and the recalculated risk differences are not the same estimand. The former come from logistic regression; the latter are unadjusted comparisons based on posted summary counts. Differences between their numerical values do not necessarily signal an error. Each should be interpreted on its own scale and in light of its method.

Multiple outcomes create multiple opportunities to observe a result that appears notable by chance. The registry record does not report whether a multiplicity procedure was planned or applied. Without that information, unadjusted secondary p-values should carry less weight than the prespecified primary endpoint and should be treated as supportive or exploratory rather than definitive evidence. Results for secondary outcomes not numerically detailed in the posted data cannot be inferred from the outcome list alone.

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

• Objective Response Rate
• Time to Objective Response
• Duration of Objective Response
• Disease Control
• Duration of Disease Control
• Tumour Shrinkage (Main Overall Survival Analysis Cut-off Date, 08 April 2016)
• Health-related Quality of Life (Primary Analysis Cut-off Date, 21 August 2015)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary analysis is a time-to-event comparison. Instead of recording only whether a participant experienced progression or death, it uses the time from randomisation to the first of those events. A participant who has no recorded event before follow-up ends is censored. Censoring means the exact event time is not observed during the available follow-up; it does not mean that the participant experienced no event thereafter.

What does the hazard ratio describe?

The registry reports a hazard ratio of 0.822 for afatinib versus gefitinib, calculated using a log-rank method. In general, a hazard ratio compares event hazards between groups over time. A hazard is an event rate at a particular point among participants who have not yet had the event. It is not the same as a ratio of medians, a risk difference at a fixed time, or the probability that a particular participant will experience an event.

The direction depends on the order of the treatment groups in the comparison. For afatinib versus gefitinib, an estimate below 1 points toward a lower event hazard for afatinib. The confidence interval and p-value are important alongside that point estimate: they show how uncertain the estimated comparison is and how compatible the results are with a null comparison under the method’s assumptions. The registry record does not report a separate test of the proportional-hazards assumption, so it should not be assumed that such a test was performed or passed.

What does the median add?

The median progression-free survival is the time by which half of the relevant event-time distribution has experienced progression or death, with estimation accounting for censoring. It offers a time-based summary that is relatively easy to describe, but it does not capture the full event-time pattern. Two groups can have medians that do not fully describe differences in the timing of events throughout follow-up. The hazard ratio and the medians therefore provide related but distinct summaries.

The reported medians are 12.78 months for afatinib and 11.17 months for gefitinib. Their individual 95% confidence intervals are 10.91 to 14.72 months and 9.49 to 12.81 months, respectively. Those intervals quantify uncertainty around each median. They are not a direct interval estimate for the difference between the medians.

Why does censoring matter?

Time-to-event methods can use information from participants who have not had an event by the end of their observed follow-up without treating them as if an event occurred at that point. The analysis still depends on how follow-up and censoring relate to the event process. The registry describes censoring participants with no progression or death but does not report a detailed account of censoring patterns or the reasons follow-up ended for each participant. Those details are not reported in the registry record.

The record describes the primary analysis population as the randomised set, including participants whether treated or not. This preserves the stated assignment-based population for the primary comparison. The record does not give further detail about how specific missing assessments were handled, or about the analysis populations used for every secondary outcome.

How do the binary secondary analyses differ?

Objective response and disease control are reported as percentages of participants, with confidence intervals. These outcomes reduce participant experience to whether a defined response or control status was recorded, rather than representing time to progression or death. The sponsor’s reported analyses use logistic regression and odds ratios. The recalculated comparisons use unadjusted risk differences based on posted summary counts. An odds ratio and a risk difference express effects on different scales, and an unadjusted calculation is not a reproduction of a regression analysis.

For these binary outcomes, a confidence interval gives a range of values compatible with the data and method. An interval for a risk difference is interpreted relative to no difference at zero; an interval for an odds ratio is interpreted relative to no association at 1. Such intervals should be considered together with the point estimate and the analysis method, rather than used as a stand-alone decision rule.

What role does the log-rank method play?

The log-rank method is a general approach for comparing time-to-event experience between groups while incorporating event timing and censoring. The registry identifies it as the method for the reported hazard-ratio comparison. The trial record does not provide enough detail to reconstruct all steps of the analysis, including any additional modelling choices, handling of protocol deviations, or sensitivity analyses. Such details should not be inferred from the method label alone.

The primary imaging assessment and the endpoint definition also matter to interpretation. Independent central review was specified as the primary-analysis approach to progression assessment, while the final analysis used investigator assessment for progression status and date. The outcome remains a composite of progression or death as defined in the record; the reported primary estimate does not separately quantify each component.

11. Confidence Intervals and Precision

The primary hazard-ratio estimate is 0.822, with a 95% confidence interval from 0.655 to 1.032. The interval indicates meaningful uncertainty around the point estimate: it includes values below 1 as well as values slightly above 1. Thus, the direction suggested by the estimate is not equally precise across all plausible values. The interval includes the null value of 1, consistent with the reported p-value of 0.089 not providing strong evidence against a no-difference comparison under conventional testing.

A confidence interval is not a guarantee that the true effect lies within its endpoints. In general, it describes the range of estimates that are reasonably compatible with the observed data and the statistical method, under its assumptions. It is useful to focus on both the interval’s width and the values it includes, rather than treating the point estimate as exact.

The median progression-free survival estimates also have 95% confidence intervals. For afatinib, the interval is 10.91 to 14.72 months; for gefitinib, it is 9.49 to 12.81 months. These intervals quantify uncertainty for the medians separately. They should not be read as an interval for the between-group contrast, which is represented by the reported hazard ratio and its confidence interval.

The registry also posts confidence intervals for the secondary outcome estimates. Their interpretation depends on the measure: a risk-difference interval is assessed against zero, while an odds-ratio interval is assessed against 1. For secondary results, estimates and intervals should also be read in light of the analysis method and the fact that multiple outcomes are reported. The registry record does not report a multiplicity adjustment.

12. Planning a Similar Trial: Sample Size

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

The planning figures are illustrations based on the observed effect, not the original trial design calculation. For the time-to-event comparison, the figures are total events required; they are not participants per arm. For the binary outcomes, the figures are participants per arm. The illustrations assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. They use the stated fraction of the observed effect as the effect to detect.

The 80% power column describes a planning scenario in which a study of the stated size would have an 80% chance of detecting the chosen effect under the calculation’s assumptions, if that effect were the true effect. The 90% power column describes the corresponding planning scenario with a 90% chance. These are design probabilities, not guarantees about the result of a future study.

For the primary time-to-event comparison, planning for 100% of the observed effect gives 818 total events for 80% power and 1,094 total events for 90% power. Planning for 75% of the observed effect gives 1,453 total events for 80% power and 1,945 total events for 90% power. Planning for 50% of the observed effect gives 3,269 total events for 80% power and 4,376 total events for 90% power. The pattern illustrates the general principle that smaller effects require more information to detect reliably.

For binary outcomes, the corresponding planning quantities are participants per arm, not total participants. A smaller effect likewise requires more participants per arm under the stated assumptions. These illustrations are not observed power: they are planning calculations, not a retrospective measure of how likely the completed trial was to detect its true effect. The registry record does not report the trial’s original sample-size assumptions or justification.

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 = 0.822, so ln(HR) = -0.1960.
2. d = 4 × (1.96 + 0.842)² / [ln(0.822)]² = 817.1, rounded up to 818 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 <- 0.822                              # 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(1466660)
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 = 818 (80% power), 1,094 (90% power)
 75% of observed effect: events = 1,453 (80% power), 1,945 (90% power)
 50% of observed effect: events = 3,269 (80% power), 4,376 (90% power)
Simulated power with 818 events: 0.794 (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 1,367 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 319 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 <- 11.17          # control-arm median, months (converted from the posted unit)
hr          <- 0.822          # 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: 1036         p (1-sided)   0.0001
  Events: 369        ~HR at bound   0.6794
    Month: 18    P(Cross) if HR=1   0.0001
              P(Cross) if HR=0.82   0.0338
    IA 2: 67%                   Z   2.5114
      N: 1368         p (1-sided)   0.0060
  Events: 737        ~HR at bound   0.8311
    Month: 29    P(Cross) if HR=1   0.0060
              P(Cross) if HR=0.82   0.5603
        Final                   Z   1.9930
      N: 1368         p (1-sided)   0.0231
 Events: 1106        ~HR at bound   0.8870
    Month: 60    P(Cross) if HR=1   0.0250
              P(Cross) if HR=0.82   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 sponsor’s reported primary analysis gives a hazard ratio of 0.822, a 95% confidence interval from 0.655 to 1.032, and p = 0.089 using a log-rank method. These are the primary comparison estimates used to interpret progression-free survival. The simple recalculation is based on the posted summary statistics; for this time-to-event endpoint, the reported hazard ratio is the relevant comparison rather than a newly derived contrast of the two medians.

The registry record does not flag a discrepancy for the primary result. The sponsor’s reported hazard ratio and the posted median estimates answer related but different questions, so the medians should not be substituted for the hazard-ratio analysis. The interval around the hazard ratio includes 1, and the p-value is 0.089; the result therefore remains uncertain about the direction and size of the relative difference.

For the secondary binary outcomes, the sponsor reports logistic-regression odds ratios, while the simple recalculations use unadjusted risk differences from the posted group counts. These estimates need not match because they use different effect measures and analysis approaches. The sponsor’s reported analyses should be distinguished from those unadjusted recalculations rather than treated as competing calculations of exactly the same quantity.

16. Multiplicity, Missing Data and Analysis Populations

The registry identifies progression-free survival as the primary endpoint and lists multiple secondary outcomes, including response, disease control, tumour shrinkage, and health-related quality of life. When a study assesses multiple outcomes, the chance of obtaining at least one apparently unusual result can increase if each is tested separately without adjustment. A multiplicity plan can define how such testing is controlled and which outcomes are intended to support confirmatory conclusions.

The registry record does not report a multiplicity adjustment, a hierarchy for testing secondary outcomes, or whether the listed secondary analyses were designated as confirmatory or exploratory. Their p-values should therefore be read cautiously and should not be treated as if they were all part of a single error-controlled testing strategy.

The primary analysis population is described as the randomised set, including patients assigned to treatment whether treated or not. That description clarifies the population for the primary endpoint, but the registry record does not provide a complete account of missing-data procedures across all outcomes. It states that participants without progression or death were censored. It does not report detailed censoring patterns, reasons for loss to follow-up, or the handling of missing imaging or other assessments. These omissions limit assessment of how incomplete follow-up may have influenced the estimates.

The record also does not report the complete statistical analysis plan, the original sample-size justification, or whether interim analyses were planned or conducted. These are not details that can be recovered from the posted endpoint results. The absence of this information in the registry record does not establish that a procedure was absent; it means the procedure is not reported there.

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: diarrhoea (6.9%), pleural effusion (5.6%), pneumonia (4.4%), malignant neoplasm progression (4.4%), pulmonary embolism (3.8%).

The registry’s safety figures list serious adverse events affecting 75 of 160 participants in the afatinib group and 64 of 159 in the gefitinib group. It also lists other adverse events affecting 157 of 160 participants with afatinib and 159 of 159 with gefitinib. The registry separately records deaths of 126 in the afatinib group and 132 in the gefitinib group. These are descriptive counts from the posted safety record and should not be interpreted as evidence that treatment caused an event or death.

The most frequently listed serious adverse event terms include diarrhoea at 6.9%, pleural effusion at 5.6%, pneumonia at 4.4%, malignant neoplasm progression at 4.4%, pulmonary embolism at 3.8%, and dizziness at 3.1%. The registry record does not provide a comparative statistical analysis for these safety figures. The record also does not clarify in this summary whether the listed deaths were attributed to treatment, arose from disease, or occurred within a specific follow-up window. Counts alone do not establish causation or comparative risk.

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

The posted primary result is an estimated progression-free survival comparison in a randomised trial of afatinib and gefitinib. The median estimates are 12.78 months and 11.17 months, respectively, and the reported hazard ratio for afatinib versus gefitinib is 0.822. Its 95% confidence interval, 0.655 to 1.032, includes 1, and the reported p-value is 0.089. The data therefore suggest a direction in the estimate but leave uncertainty about the size and direction of the underlying relative difference.

This result does not prove that the treatments are equivalent. A confidence interval that includes the null value is not the same as evidence that the true effect is exactly null. It also does not establish a definitive advantage for afatinib: the interval includes the null and extends slightly beyond it. The registry record does not report a non-inferiority or equivalence framework, so the result should not be interpreted as demonstrating either equivalence or non-inferiority.

The finding concerns the trial population, endpoint definition, follow-up, and analyses described in the registry. It does not establish effects in populations not represented by that description, or on outcomes not reported. The primary endpoint combines progression and death, and the posted hazard ratio does not separate those components. The secondary outcomes add context but are subject to the limitations of multiple analyses and differing effect measures.

Open-label assignment is a design feature to keep in mind, while independent central imaging review was used for progression assessment in the primary analysis. The final analysis used investigator assessment for progression status and date. Neither fact alone determines whether the result is biased; they describe aspects of how the outcome was assessed.

These statistical findings are not medical advice and do not determine which treatment is appropriate for an individual. They describe the registry’s reported estimates and the uncertainty around them.

19. Why This Trial Matters Statistically

This trial is statistically informative because it compares two active first-line treatments through random allocation and a defined time-to-event primary endpoint. Randomisation supports a comparison of assigned groups, while the progression-free survival definition combines progression and death and accounts for their timing. The analysis population is explicitly described as the randomised set, including participants whether treated or not.

The result also illustrates why an effect estimate should not be read without its confidence interval. The hazard-ratio point estimate is 0.822, but its 95% confidence interval extends from 0.655 to 1.032 and includes 1. The point estimate alone might suggest a direction, whereas the interval makes clear that the estimate is imprecise enough to include a null comparison. The p-value of 0.089 is consistent with that uncertainty and should not replace consideration of the estimate and interval.

The trial’s assessment procedures are relevant to interpretation. It was open-label, but independent central imaging review was specified for primary-analysis progression assessment. The final analysis used investigator assessment. Recognising these details helps readers consider how outcome measurement may affect a time-to-event result without assuming that bias necessarily occurred.

The record also demonstrates the difference between a primary time-to-event outcome and secondary binary outcomes. Response and disease-control percentages answer distinct questions, and the sponsor’s logistic-regression odds ratios are not the same measures as unadjusted risk differences. Keeping these methods and estimands distinct prevents unlike statistics from being compared as if they were interchangeable.

Finally, planning illustrations based on the observed effect are not the trial’s original design justification and are not observed power. The registry record does not report the original sample-size assumptions or multiplicity plan. Those absences limit what can be said about the study’s intended precision and the confirmatory status of secondary 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 does not report the complete statistical analysis plan, including detailed modelling choices, sensitivity analyses, or a multiplicity strategy. It also does not report the original sample-size justification or whether interim analyses were planned or conducted. These omissions make it difficult to assess all analytic decisions from the registry entry alone.

The trial was open-label. Independent central imaging review was used for the primary analysis, but the final progression assessment was made by investigators. The registry description does not provide enough information to evaluate how assessment differences across analyses affected the estimates. It also does not report detailed censoring patterns or the reasons that participants’ follow-up ended.

The primary endpoint combines disease progression and death, so the primary hazard ratio does not show the separate contribution of each component. The registry describes progression assessment and censoring but does not provide further detail about the distribution of event times or missing assessments in this summary. The medians and hazard ratio each compress the time-to-event data into a limited summary.

The secondary outcomes involve multiple comparisons, and the record does not state whether adjustment was planned. The recalculated binary contrasts are unadjusted summaries, whereas the sponsor reports logistic-regression odds ratios. These measures are not directly interchangeable. The posted safety counts are descriptive and do not, in this summary, establish attribution or provide comparative statistical estimates.

These limitations concern what can be assessed from the registry record. They do not establish that unreported procedures were not performed, nor do they show that the reported analyses are invalid.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: LUX-Lung 7: A Phase IIb Trial of Afatinib(BIBW2992) Versus Gefitinib for the Treatment of 1st Line EGFR Mutation Positive Adenocarcinoma of the Lung — NCT01466660.
• Registry record last updated 2020-04-07; status last verified 2020-03.
• Results first posted 2017-06-19.
• 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

LUX-Lung 7 was a completed, randomised, open-label Phase IIb parallel-group trial sponsored by Boehringer Ingelheim. It compared once-daily afatinib with once-daily gefitinib as first-line treatment for patients with EGFR mutation-positive advanced adenocarcinoma of the lung. The registry lists enrollment of 319.

The primary endpoint was progression-free survival, defined from randomisation to progression or death, with participants without either event censored. The reported medians were 12.78 months for afatinib and 11.17 months for gefitinib. The reported hazard ratio was 0.822, with a 95% confidence interval from 0.655 to 1.032 and p = 0.089. The interval includes 1, so the estimate remains uncertain about the relative difference.

The record includes secondary response and disease-control outcomes and descriptive safety counts. It does not report the original sample-size justification, a multiplicity plan, or detailed procedures for all missing data. The findings describe the posted analyses and should be interpreted within those limits.