An Investigational Drug, PF-02341066 Is Being Studied Versus Standard Of Care In Patients With Advanced Non-Small Cell Lung Cancer With A Specific Gene Profile Involving The Anaplastic Lymphoma Kinase (ALK) Gene: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for An Investigational Drug, PF-02341066 Is Being Studied Versus Standard Of Care In Patients With Advanced Non-Small Cell Lung Cancer With A Specific Gene Profile Involving The Anaplastic Lymphoma Kinase (ALK) Gene (NCT00932893), a randomized phase 3 study. The primary outcome is Progression-Free Survival (PFS).

2. Clinical Question

3. Trial Design

This was a phase 3 randomized, open-label, parallel-group trial of PF-02341066, reported as crizotinib, versus standard-of-care chemotherapy. The chemotherapy arm used pemetrexed or docetaxel, selected by the investigator. The registry record lists 347 participants enrolled and identifies the full analysis set as all participants randomized to treatment. The primary endpoint results include 173 participants in the crizotinib group and 174 in the chemotherapy group.

The enrolled population had advanced non-small cell lung cancer with a translocation or inversion event involving the ALK gene locus. The brief study description also says participants had experienced failure of one previous chemotherapy regimen that included one platinum drug. The condition listed in the registry is carcinoma, non-small-cell lung. The study was sponsored by Pfizer and is recorded as completed.

There was no masking. Open-label treatment means that participants and study personnel were not kept unaware of assigned treatment, although the primary endpoint description specifies independent radiology review for disease progression. That separation can be relevant when interpreting a radiographic outcome: independent review is part of the stated endpoint definition, but the registry record does not describe the review procedures in further detail.

The primary endpoint was progression-free survival, assessed every 6 weeks, with the stated timeframe extending up to 112 weeks. The secondary outcomes listed include overall survival, objective response, disease control at specified weeks, duration of response, time to tumor response, and plasma concentration of crizotinib. The posted overall-survival timeframe extends up to 4.5 years. Details of randomization procedures, stratification factors, the original sample-size justification, and the trial's full statistical analysis plan are not reported in the registry record.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

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

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Full analysis set (FAS) included all participants who were randomized to study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS), measured over Randomization until progressive disease (PD) or initiation of antitumor therapy in the absence of PD or death, assessed every 6 weeks (up to 112 weeks).

Progression-free survival measures the time from randomization until the first qualifying event: objective disease progression or death from any cause. The registry definition says progression was determined by independent radiology review and applied RECIST v1.1 criteria. Those criteria include a specified increase in the summed diameters of target lesions, unequivocal progression of existing non-target lesions, or the appearance of new lesions. The registry also describes the endpoint timeframe as progression, initiation of antitumor therapy in the absence of progression, or death, assessed every 6 weeks; its endpoint definition specifically describes progression or death as the events used to calculate progression-free survival.

The posted outcome is the median time to an event in each group, with a 95% confidence interval, alongside a hazard ratio comparison. A median provides a time-based summary of the observed event experience; it does not describe every participant's course or the probability of an event at a particular time. A time-to-event analysis can use information about when events occur and account for participants whose event time is not observed during follow-up, subject to the analysis methods and assumptions. The registry record does not report further details about censoring rules.

Progression-free survival is used in oncology trials because it captures disease progression as well as death, rather than requiring death to occur before an endpoint event is counted. It is distinct from overall survival, which measures time to death. The trial's primary result therefore addresses the time-to-progression-or-death outcome defined in the record, not every aspect of disease control, survival, safety, or quality of life.

7. Results Posted in the Registry

• Crizotinib vs Chemotherapy: hazard ratio 0.49 (95% CI 0.37–0.64), reported p < 0.001 (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.

Overall Survival (OS)

• 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:
  • Crizotinib vs Chemotherapy: hazard ratio 0.85 (95% CI 0.66–1.10), reported p = 0.115 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Crizotinib vs Chemotherapy and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Objective Response (OR)

• 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:
  • Chemotherapy: 19.5% (34/174) versus 65.3% (113/173) on control; risk difference -45.8 percentage points (95% CI -55.0 to -36.6), risk ratio 0.30 (0.22–0.41); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Chemotherapy than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants With Disease Control at Week 6

• 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:
  • Chemotherapy: 55.2% (96/174) versus 81.5% (141/173) on control; risk difference -26.3 percentage points (95% CI -35.7 to -16.9), risk ratio 0.68 (0.58–0.79); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Chemotherapy than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover different dimensions of disease course. Overall survival measures time to death, while objective response and disease control describe tumor status at assessments. The posted results show a distinct pattern across these types of outcomes: the reported response and Week 6 disease-control comparisons favor crizotinib by their sponsor-reported risk ratios, whereas the overall-survival comparison has an interval that includes 1 and a p-value of 0.115.

This pattern is not inherently contradictory. Progression-free survival, response, disease control, and overall survival are related but non-identical outcomes. They use different event definitions, observation periods, and statistical comparisons. A difference in progression or response need not translate into an observed overall-survival difference within the follow-up and analysis reported. The registry record does not provide enough information to attribute the overall-survival result to subsequent treatments, crossover, or any other explanation.

Secondary endpoints provide context for the primary result, but they generally carry less confirmatory weight when many outcomes are examined. Each additional comparison creates more opportunities for a seemingly notable result to arise by chance. The registry record lists multiple secondary outcomes and 15 outcome measures, but it does not report a multiplicity-control strategy for the secondary analyses. Without such details, p-values for secondary outcomes should be read as nominal evidence rather than as a fully adjusted family of confirmatory tests.

Not all listed secondary outcomes have numerical results included in the posted results. The record lists duration of response, time to tumor response, disease control at Week 12, survival probabilities at specified months, and plasma concentration of crizotinib among the secondary outcomes, but detailed numerical results for those outcomes are not reported in the registry record. Their presence in the outcome list should not be mistaken for evidence of a particular result.

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

• Overall Survival (OS)
• Overall Survival Probability at Months 6 and 12
• Percentage of Participants With Objective Response (OR)
• Percentage of Participants With Disease Control at Week 6
• Percentage of Participants With Disease Control at Week 12
• Duration of Response (DR)
• Time to Tumor Response (TTR)
• Plasma Concentration of Crizotinib

9. Statistical Interpretation

10. Statistical Methods Explained

What does a median time-to-event result mean?

For a time-to-event outcome, each participant contributes information about the time from a defined starting point until an event, or until the last time the event status is known. The registry defines the progression-free survival starting point as randomization and the event as objective progression or death, whichever occurred first. The median is a central summary of the event-time distribution: it is the time by which half of the analyzed group has had an event, under the estimation method. It is not an average, and it does not mean that every participant experienced an event at that time.

Time-to-event methods can incorporate participants whose event has not been observed by the end of their recorded follow-up through censoring. Censoring is a statistical way to retain the information that a participant was event-free up to a known time, without treating that time as the actual event time. The registry record does not provide detailed censoring rules or the amount of follow-up for individual participants, so those implementation details cannot be assessed from the posted results.

What does the hazard ratio describe?

A hazard is the instantaneous event rate among participants who have not yet had the event, considered over time. A hazard ratio compares that rate between groups. The primary comparison is reported as crizotinib versus chemotherapy, with a hazard ratio of 0.487. In general, a hazard ratio below 1 indicates a lower estimated event hazard in the first-named group over the analyzed period, while a value above 1 indicates a higher estimated hazard. This is a relative comparison of event rates, not an absolute probability and not a direct comparison of median times.

The reported 95% confidence interval for the primary hazard ratio is 0.371 to 0.638. A confidence interval gives a range of values compatible with the data and statistical model at the stated confidence level. It conveys uncertainty in the estimated comparison; it does not guarantee that the true value lies in the interval. The reported p < 0.001 is associated with the Log Rank comparison. In general, a p-value describes how incompatible the observed data are with a specified null hypothesis under the test assumptions; it does not measure the size or clinical importance of an effect.

The Log Rank method compares time-to-event experience between groups across follow-up. It is commonly used for a group comparison when event timing and censoring matter. Interpretation of a single hazard ratio is simplest when the relative hazards are reasonably stable over time. The registry record does not report whether proportional-hazards assumptions were assessed, whether any adjusted analyses were conducted, or which additional modeling checks were performed. Those assumptions and procedures should not be inferred from the reported test name alone.

How do the binary secondary outcomes differ?

Objective response and disease control at a stated week are binary outcomes: a participant either meets the outcome definition or does not. The posted percentages and confidence intervals summarize the proportion meeting each outcome in each group. For objective response, the sponsor reports a Cochran-Mantel-Haenszel risk ratio of 3.394, with a 95% confidence interval from 2.463 to 4.676 and p < 0.001. For disease control at Week 6, the sponsor reports a risk ratio of 1.502, with a 95% confidence interval from 1.297 to 1.741 and p < 0.001.

A risk ratio compares proportions between groups; a value above 1 means the outcome proportion is higher in the group in the numerator under the reported orientation. The Cochran-Mantel-Haenszel method can combine comparisons across strata, but the registry record does not report which strata were used for these analyses. It is therefore appropriate to describe the sponsor's reported method and estimate without assuming details of the adjustment.

Separate simple, unadjusted recalculations can summarize the binary group proportions using a risk difference and an unadjusted p-value. Those calculations use posted summary counts and are not the same analysis as the sponsor's reported Cochran-Mantel-Haenszel risk ratio. They answer a different statistical contrast and may use different assumptions. Where the posted sponsor analysis is available, it is the primary reported inferential result; a simple recalculation is supplementary and should not be presented as replacing it.

Why distinguish reported analyses from planning illustrations?

The reported analyses estimate trial outcomes using the registry's stated methods. Sample-size calculations based on an observed effect are different: they illustrate what sample sizes might be needed under stated assumptions to detect an effect of a chosen size. They are not a reconstruction of the original trial design and do not tell whether the completed study had a particular probability of success. The registry record does not report the original sample-size justification or enough information to recreate the full original analysis plan.

For time-to-event planning, event counts matter because participants may have different follow-up times and event status. For binary planning, the number of participants per group is the relevant planning quantity. Both depend on the assumed effect and analysis conditions. A smaller target effect is harder to distinguish from random variation and generally requires more information, often through more participants or more events. These are general statistical principles; they do not establish what assumptions governed the trial's original design.

11. Confidence Intervals and Precision

The 95% confidence intervals communicate uncertainty around the posted estimates. For median progression-free survival, the interval is 6.0 to 8.8 months for crizotinib and 2.6 to 4.3 months for chemotherapy. For the hazard ratio, the interval is 0.371 to 0.638. These intervals show that the estimates are not exact; the range of plausible values depends on the data and the statistical method. They should be considered alongside the point estimates rather than treated as guarantees about the true values.

The primary hazard-ratio interval lies below 1, consistent with the reported p < 0.001. The interval describes uncertainty in the relative hazard estimate, not uncertainty in a directly measured percentage reduction in median progression-free survival. The separate median confidence intervals describe uncertainty around each group's median and are not themselves a formal test of the between-group hazard ratio.

For overall survival, the hazard-ratio interval is 0.661 to 1.104 and includes 1. This means the posted estimate is compatible with a range that includes no relative hazard difference as well as effects in either direction. The p-value of 0.115 likewise does not meet conventional evidence against the no-difference null, but it does not establish equivalence or prove that there is no effect.

Confidence intervals are influenced by the amount of information available and the variability of the data. Time-to-event precision depends importantly on observed events and follow-up, while binary-outcome precision depends on the number of participants and the distribution of outcomes. The registry record does not give all details needed to assess precision beyond the reported intervals and group counts. Statistical precision is also distinct from clinical importance: an interval can be relatively narrow or exclude a null value without answering whether an effect matters to patients.

12. Planning a Similar Trial: Sample Size

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

The sample-size illustrations are planning calculations based on observed trial effects, not the original design justification. For the time-to-event comparisons, the figures are total events, not participants per arm. For binary outcomes, the figures are participants per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two groups, and a normal approximation. Each row considers an effect equal to the stated fraction of the observed effect.

The 80% power column describes the planning sample size or event total that would give an 80% chance of detecting the specified effect under the calculation assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding planning quantity for a higher chance of detecting that effect. Power is a design calculation for a hypothesized effect, not a guarantee that a study will produce a statistically significant result.

The illustrations make the effect-size dependence visible: planning for 75% or 50% of the observed effect requires more participants or events than planning for 100% of the observed effect. Smaller effects are harder to distinguish from random variation, so more information is generally needed. For example, the primary time-to-event planning row gives total event counts of 61 for 80% power and 82 for 90% power at 100% of the observed effect; the corresponding row for 50% of the observed effect gives 243 and 325 total events. For the objective-response calculation, the per-arm figures range from 18 for 80% power at 100% of the observed effect to 74 for 80% power at 50% of the observed effect.

These are planning illustrations computed from the observed effect and the stated assumptions. They are not observed power, and they should not be used to infer the probability that the completed trial's results were true or reproducible. The original sample-size assumptions and justification are not reported in the registry record.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.487, so ln(HR) = -0.7195.
2. d = 4 × (1.96 + 0.842)² / [ln(0.487)]² = 60.6, rounded up to 61 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 <- 0.487                              # 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(932893)
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 = 61 (80% power), 82 (90% power)
 75% of observed effect: events = 108 (80% power), 145 (90% power)
 50% of observed effect: events = 243 (80% power), 325 (90% power)
Simulated power with 61 events: 0.760 (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 88 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 347 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 <- 3          # control-arm median, months (converted from the posted unit)
hr          <- 0.487          # 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: 50         p (1-sided)   0.0001
 Events: 28        ~HR at bound   0.2413
  Month: 13    P(Cross) if HR=1   0.0001
            P(Cross) if HR=0.49   0.0338
  IA 2: 67%                   Z   2.5114
      N: 80         p (1-sided)   0.0060
 Events: 55        ~HR at bound   0.5064
  Month: 22    P(Cross) if HR=1   0.0060
            P(Cross) if HR=0.49   0.5603
      Final                   Z   1.9930
      N: 88         p (1-sided)   0.0231
 Events: 82        ~HR at bound   0.6435
  Month: 60    P(Cross) if HR=1   0.0250
            P(Cross) if HR=0.49   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-reported primary comparison is a Log Rank hazard ratio of 0.487, with a 95% confidence interval from 0.371 to 0.638 and p < 0.001. The primary time-to-event result is presented using that reported hazard ratio; there is no conflicting recalculation identified for this comparison. The posted medians and their confidence intervals provide additional descriptive summaries of the progression-free survival experience in each group.

For the binary secondary outcomes, simple unadjusted recalculations based on posted group summaries are not interchangeable with the sponsor's analyses. The sponsor reports Cochran-Mantel-Haenszel risk ratios for objective response and disease control at Week 6. Those reported analyses are the appropriate figures to cite as the sponsor's results. A recalculated unadjusted risk difference or p-value describes a simpler contrast from the posted counts and may differ because the estimand and method differ.

No discrepancy is identified between the sponsor's reported primary result and the posted recalculation record. The registry record does not provide enough detail to reproduce all sponsor analyses independently, including any strata used for the Cochran-Mantel-Haenszel calculations. The sponsor's own reported analysis therefore remains the central inferential result where it is available, while unadjusted recalculations are supplementary descriptions rather than corrections.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a trial examines several outcomes, time points, or statistical comparisons. Even when each test uses a conventional significance threshold, conducting many tests increases the chance that at least one result will appear statistically notable by chance alone. Common approaches include designating a single primary endpoint, arranging tests in a prespecified hierarchy, or adjusting the interpretation or threshold for a family of comparisons. These are general statistical approaches; the registry record does not report which multiplicity strategy, if any, was used for the secondary endpoints.

The registry identifies progression-free survival as the primary outcome and lists multiple secondary outcomes. The sponsor-reported result for the primary comparison is therefore the main confirmatory result available in the record. The response and disease-control results are additional evidence, but the registry record does not state whether their p-values were adjusted for multiple testing or whether they were intended as confirmatory. Their reported p-values should consequently be interpreted with that uncertainty in mind.

Missing-data handling can also affect interpretation. In time-to-event analysis, a participant without a recorded event may be censored at the last time known to be event-free; the assumptions behind censoring and the rules for follow-up can influence results. For binary outcomes, missing assessments can affect the denominator and classification of response. The registry record does not report detailed rules for censoring, missing outcome assessments, treatment discontinuation, or sensitivity analyses addressing missing data.

The primary endpoint population is identified as the full analysis set, which included all participants randomized to study treatment. That definition is reported, but the record does not provide a fuller account of protocol deviations, exclusions from particular analyses, or the analysis populations used for every secondary outcome. It also does not report the original sample-size justification, interim analyses, or a complete statistical analysis plan. These omissions limit assessment of prespecification and robustness; they do not, by themselves, establish that an analysis was inappropriate.

17. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: disease progression (10.5%), febrile neutropenia (7.0%), pneumonia (4.7%), pulmonary embolism (4.1%), dyspnoea (3.5%).

The adverse-event record lists serious adverse events for 80 of 172 participants at risk in the crizotinib group and 42 of 171 participants at risk in the chemotherapy group. It also lists other adverse events for 172 of 172 participants at risk in the crizotinib group and 169 of 171 in the chemotherapy group. These are descriptive counts in the registry record. They do not by themselves establish that a treatment caused an event, and they should be interpreted in light of the event definitions, observation periods, and reporting procedures, details of which are not reported in the registry record.

The registry lists disease progression at 10.5%, febrile neutropenia at 7.0%, pneumonia at 4.7%, pulmonary embolism at 4.1%, dyspnoea at 3.5%, and pleural effusion at 1.8% among the serious adverse-event terms. The record does not identify arm-specific percentages for these terms in the posted figures, so those percentages should not be assigned to an individual group. Death counts in the adverse-event results are not reported in the registry record.

Safety comparisons require more than comparing raw counts: the number at risk, exposure time, event definitions, and whether events were counted once per participant or as repeated events can matter. The registry figures give the participant counts at risk and affected for the broad categories, but do not report all of these details. The serious-event counts are therefore a descriptive account, not a complete comparative safety analysis.

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

The primary result indicates that the randomized groups had different progression-free survival experience in the posted analysis. The crizotinib group had a median progression-free survival of 7.7 months, compared with 3.0 months in the chemotherapy group, and the reported hazard ratio was 0.487 with a 95% confidence interval from 0.371 to 0.638. The reported Log Rank p-value was < 0.001. These statistics describe the trial groups and the defined time-to-event outcome.

They do not show that every participant in one group had a longer progression-free time than every participant in the other group. They do not give an individual's chance of benefit, and the hazard ratio is not a direct measure of an absolute probability or median-time difference. The confidence interval expresses uncertainty about the estimated relative hazard, while the median confidence intervals describe uncertainty in each group median.

The overall-survival analysis is separate. Its reported hazard ratio was 0.854, with a 95% confidence interval from 0.661 to 1.104 and p = 0.115. This result does not establish a difference in overall-survival hazards, but it also does not prove the groups are equivalent. The record does not report enough information to explain why the overall-survival comparison differs from the progression-free survival result.

The results are specific to participants with advanced non-small cell lung cancer involving the ALK gene locus who had experienced failure of one previous platinum-containing chemotherapy regimen, and to the trial's comparison of crizotinib with investigator-selected pemetrexed or docetaxel. The open-label design and incomplete registry reporting of analysis details constrain interpretation. The posted outcomes do not resolve questions about unreported endpoints, individual treatment decisions, or outcomes outside the studied population. They should be read as statistical evidence about this trial comparison, not as medical advice.

19. Why This Trial Matters Statistically

This trial illustrates why the endpoint definition matters. Progression-free survival counts either objective disease progression or death, whereas overall survival counts death. The two endpoints can therefore tell different statistical stories without being inconsistent. Reading both requires attention to their event definitions and follow-up periods rather than treating them as interchangeable measures.

It also illustrates the complementary roles of a median and a hazard ratio. The medians give a time-based summary within each group, while the hazard ratio compares event rates over follow-up. Their confidence intervals provide distinct information about precision. A reader who relies on only one number may miss how the analysis summarizes the data and what remains uncertain.

The trial also makes the distinction between sponsor analyses and simple recalculations important. A reported time-to-event hazard ratio or a stratified binary-outcome analysis is not necessarily reproduced by an unadjusted comparison of summary counts. Keeping the estimand and method attached to each result prevents unlike statistics from being treated as competing answers to the same question.

Finally, secondary endpoints and planning calculations require careful interpretation. Multiple secondary comparisons can weaken the force of isolated p-values when multiplicity handling is unknown, and sample-size illustrations based on observed effects do not reveal the original design or the probability that a completed result is true. The registry record gives meaningful outcome estimates, but its omissions about analysis plans, missing data, and multiplicity show why statistical evidence must be read together with the limits of what the record documents.

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, the complete statistical analysis plan, interim analyses, randomization procedures, or stratification factors. It also does not provide detailed proportional-hazards assessments or the censoring rules used for the time-to-event analyses. These details would help readers evaluate prespecification, assumptions, and the reproducibility of the analyses.

The study was open-label, so treatment assignment was not masked. The primary endpoint description specifies independent radiology review, which is relevant to progression assessment, but the record does not provide further details about review procedures or whether all relevant assessments were handled without knowledge of treatment. The registry does not report detailed missing-data procedures or sensitivity analyses.

The comparator was investigator selection of pemetrexed or docetaxel, rather than one single chemotherapy drug for all participants. The posted comparison is therefore against the chemotherapy group as a whole; the record does not provide enough information to assess outcomes separately by selected chemotherapy drug. The results also concern a defined population with advanced disease and the specified ALK gene event after prior platinum-containing chemotherapy, limiting how broadly they can be generalized.

Several secondary outcomes are listed, but numerical results for all listed outcomes are not reported in the registry record. The record also does not report a multiplicity strategy for secondary comparisons. Safety figures are descriptive and lack some details needed for a complete comparison, including arm-specific percentages for the listed serious-event terms and reported death counts. These limits narrow what can be concluded from the posted results; they do not negate the reported primary comparison.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: An Investigational Drug, PF-02341066 Is Being Studied Versus Standard Of Care In Patients With Advanced Non-Small Cell Lung Cancer With A Specific Gene Profile Involving The Anaplastic Lymphoma Kinase (ALK) Gene — NCT00932893.
• Registry record last updated 2017-01-02; status last verified 2016-10.
• Results first posted 2013-06-06.
• 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

The registry describes a completed, randomized, open-label phase 3 trial sponsored by Pfizer, comparing PF-02341066, reported as crizotinib, with investigator-selected pemetrexed or docetaxel in advanced non-small cell lung cancer involving the ALK gene locus. The study enrolled 347 participants, and the primary analysis used the full analysis set of randomized participants.

The primary endpoint was progression-free survival. Median progression-free survival was 7.7 months with crizotinib and 3.0 months with chemotherapy; the reported hazard ratio was 0.487, with a 95% confidence interval from 0.371 to 0.638 and p < 0.001. The overall-survival hazard ratio was 0.854, with a 95% confidence interval from 0.661 to 1.104 and p = 0.115. Secondary response and disease-control outcomes add context but require cautious interpretation because the record does not report a multiplicity strategy. Important details of the original design and analysis plan are not reported in the registry record.