Phase 2 Study of Alisertib (MLN8237) in Combination With Paclitaxel Versus Placebo in Combination With Paclitaxel as Second Line Therapy for Small Cell Lung Cancer (SCLC): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase 2 Study of Alisertib (MLN8237) in Combination With Paclitaxel Versus Placebo in Combination With Paclitaxel as Second Line Therapy for Small Cell Lung Cancer (SCLC) (NCT02038647), a randomized phase 2 study. The primary outcome is Progression-Free Survival (PFS) as Determined by Investigator, Analyzed Using FDA Guidelines.

2. Clinical Question

The study was randomized, parallel-group, and placebo-controlled. Its title and summary describe it as a phase 2 study of second-line treatment for small cell lung cancer. The two regimens paired the study drug or matching placebo with paclitaxel, but the paclitaxel doses listed for the arms were not the same. The comparison therefore concerns the two assigned treatment regimens as described, rather than alisertib versus placebo with all other treatment details held identical.

The registry defines progression-free survival in the primary outcome description as time from the start of study treatment to the first investigator-documented objective tumor progression or death from any cause, whichever occurred first. The brief summary instead describes the time origin as randomization. This difference in the record is a limitation when interpreting the endpoint definition; the precise operational resolution is not reported in the registry record.

3. Trial Design

This was a randomized, parallel-group, placebo-controlled phase 2 trial. The registry records quadruple masking and describes the study as double-blind in its title and brief summary. It was multicenter. The registry lists 178 participants as the enrollment and identifies the study status as completed.

One group received alisertib plus paclitaxel; the comparator group received placebo plus paclitaxel. Alisertib was given orally at 40 mg twice a day, for 3 days on and 4 days off over 3 weeks of a 28-day cycle, on Days 1–3, 8–10, and 15–17. Paclitaxel in this group was 60 mg/m^2 intravenously once a week on Days 1, 8, and 15. The comparator used placebo-matching tablets on the same schedule and paclitaxel at 80 mg/m^2 intravenously on the same weekly schedule. Treatment was described as continuing until disease progression, with up to 17 cycles for the alisertib regimen and up to 22 cycles for the placebo regimen.

The primary outcome was investigator-determined progression-free survival. The registry reports assessment every cycle for the first 6 months and then every 2 months, until progression, death, or the data cut-off of 03 January 2016. It also lists overall survival, response outcomes, symptoms, and adverse events among secondary outcomes.

The registry gives the study start as 2014-05-12 and primary completion as 2016-01-03. The trial's original sample-size justification, detailed randomization procedure, exact allocation concealment process, and role-specific masking details are not reported in the registry record. The number of participants included in each individual endpoint analysis may differ from the randomized population; endpoint-specific analysis populations are not fully described in the posted results.

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: Participants with a diagnosis of Small Cell Lung Cancer (SCLC) were enrolled in 1 of 2 treatment groups: alisertib + paclitaxel or placebo + paclitaxel arm group.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The intent-to-treat (ITT) population was defined as all participants who were randomized to study treatment. For participants who have not progressed and is last known to be alive, PFS was censored at the last response assessment that is stable disease (SD) or better as determined by Investigator, and analyzed using FDA Guidelines.

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: The ITT population was defined as all participants who were randomized to study treatment, whether or not any treatment was administered.

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS) as Determined by Investigator, Analyzed Using FDA Guidelines, measured over Every cycle for first 6 months and then every 2 months until disease progression or death or up to data cut-off: 03 January 2016 (approximately 22 months).

Progression-free survival measures the time until the first documented disease progression or death, whichever occurs first. In this trial, the registry describes investigator assessment using RECIST v1.1 and defines progression using changes in tumor measurements or new lesions. The endpoint is a time-to-event measure: it uses both whether an event occurred and the time at which it occurred, while participants without a recorded event may contribute follow-up until censoring.

Progression-free survival is used in trials because it captures disease progression and death in one endpoint and can be assessed before every participant has died or reached a later outcome. It is not interchangeable with overall survival: a difference in progression-free survival does not by itself establish a difference in how long participants live. Nor does the endpoint directly summarize symptoms, quality of life, or treatment burden.

The registry reports median progression-free survival of 101.0 days for alisertib plus paclitaxel, with a 95% confidence interval of 80.0 to 113.0 days, and 66.0 days for placebo plus paclitaxel, with a 95% confidence interval of 53.0 to 83.0 days. It also reports a hazard ratio of 0.77, with a 95% confidence interval of 0.557 to 1.067 and p = 0.113. The median is a summary of the time-to-event distribution, while the hazard ratio compares event rates over follow-up under the model used. Those summaries describe different aspects of the observed outcome.

The record gives two different time origins in its endpoint descriptions: the detailed primary outcome uses start of study treatment, while the brief summary uses randomization. That discrepancy in wording is not resolved in the registry record.

7. Results Posted in the Registry

• Alisertib + Paclitaxel vs Placebo + Paclitaxel: hazard ratio 0.77 (95% CI 0.56–1.07), reported p = 0.113 (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:
  • Alisertib + Paclitaxel vs Placebo + Paclitaxel: hazard ratio 0.93 (95% CI 0.65–1.34), reported p = 0.714 (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 Alisertib + Paclitaxel vs Placebo + Paclitaxel and the control. The interval gives the range of effects that remain compatible with the data.

Overall Response Rate (ORR)

• 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:
  • Alisertib + Paclitaxel: 22.5% (20/89) versus 18.0% (16/89) on control; risk difference 4.5 percentage points (95% CI -7.3 to 16.3), risk ratio 1.25 (0.69–2.25); unadjusted two-proportion z-test p = 0.455.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alisertib + Paclitaxel and the control. The interval gives the range of effects that remain compatible with the data.

Complete Response Rate (CRR)

• 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:
  • Alisertib + Paclitaxel: 1.1% (1/89) versus 0.0% (0/89) on control; risk difference 1.1 percentage points (95% CI -1.1 to 3.3); unadjusted two-proportion z-test p = 0.316.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alisertib + Paclitaxel and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover several distinct domains: overall survival, tumor response, symptom scores, symptom relief, and treatment-emergent and serious adverse events. The posted results include numeric estimates for overall survival and response measures, while the registry lists additional symptom and safety outcomes. These endpoints help describe other aspects of the trial but do not change which outcome was designated primary.

The overall survival hazard-ratio estimate is 0.93, with a 95% confidence interval of 0.652 to 1.341 and p = 0.714. For overall response rate, the sponsor-reported odds ratio is 0.74, with a 95% confidence interval of 0.35 to 1.55 and p = 0.406. The complete response rate is reported as 1.0% in the alisertib group and 0.0% in the placebo group; its sponsor analysis gives an odds ratio of 0.01, with a 95% confidence interval of 0.01 to 9999.99 and p = 0.283. These results are not a consistent basis for claiming a clear difference across the secondary efficacy outcomes.

Secondary outcomes can provide useful context, but they are often less conclusive than a prespecified primary endpoint. Testing many outcomes creates multiple opportunities for a result to appear noteworthy by chance. Unless a multiplicity strategy is specified and applied, individual secondary p-values should be treated as exploratory rather than as independent confirmatory evidence. The registry does not report a multiplicity adjustment or a testing hierarchy.

The registry lists changes in cough, dyspnea, and pain scores and the percentage of participants experiencing symptom relief, but numeric results for those measures are not included in the registry record in the registry record. It also lists multiple secondary outcomes beyond the efficacy estimates. Without their posted results and analysis details, the overall secondary pattern cannot be fully characterized. These endpoints should be considered alongside, not substituted for, the primary progression-free survival analysis.

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

• Percentage of Participants Who Experience at Least One Treatment Emergent Adverse Events (TEAEs) and Serious Adverse Events (SAEs)
• Overall Survival (OS)
• Overall Response Rate (ORR)
• Complete Response Rate (CRR)
• Disease Control Rate (DCR)
• Duration of Response (DOR)
• Change From Baseline in Symptom (QLQ-LC13 Cough Scale, QLQ-C30 Dyspnea Scale, QLQ-C30 Pain Scale) Score at Cycle 5
• Percentage of Participants Experiencing Symptom Relief

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is a time-to-event outcome. For each participant, the event is the first investigator-documented tumor progression or death, whichever comes first, according to the detailed primary outcome description. Participants who have not experienced an event by the end of their available observation period are censored. The registry describes censoring at the last response assessment showing stable disease or better for participants who had not progressed and were last known to be alive, but the text of that description is incomplete. Other details of censoring and the handling of missing assessments are not reported in the registry record.

What does a median time represent?

A median time-to-event value marks the point in the estimated event-time distribution at which half of the distribution is on either side. It is not the average time for all participants, nor does it describe each participant's individual outcome. In this trial, the reported medians are 101.0 days for alisertib plus paclitaxel and 66.0 days for placebo plus paclitaxel. The associated confidence intervals express uncertainty around each median estimate.

How should the hazard ratio be read?

A hazard ratio compares the estimated event rate in one group with the other over the observed follow-up, using the analysis method specified by the sponsor. A value of 1 represents equal hazards under that comparison; values below or above 1 indicate different directions of the estimated relative hazard. The primary reported estimate is 0.77. It is a relative measure over time, not a ratio of the two medians, and it does not state how many days an individual participant gained or lost. The registry reports that the comparison used a log-rank method and gives a 95% confidence interval of 0.557 to 1.067 and p = 0.113.

General statistical context: time-to-event methods account for participants having different observed follow-up times and for censoring, rather than treating every participant as if their event status were known at one fixed time. The methods rely on assumptions about the observation and censoring process. Whether the trial assessed those assumptions, or used additional modeling adjustments, is not reported in the registry record.

What does a confidence interval add?

A confidence interval conveys the precision of an estimate by giving a range of values compatible with the data and method under repeated-sampling reasoning. A wider interval indicates less precision than a narrower one. It is not a range guaranteed to contain the true value, and it does not give the probability that the true value is inside the particular interval. The primary hazard-ratio interval includes 1, which is consistent with no relative hazard difference as one possibility among the values compatible with the reported analysis.

What does the p-value tell us?

The p-value describes how unusual the observed result, or a result more extreme, would be under the null hypothesis and the assumptions of the test. It is not the probability that the null hypothesis is true. The primary reported p-value is 0.113. It should be read alongside the effect estimate, confidence interval, endpoint definition, and study context rather than as a stand-alone measure of importance.

Why can over-dispersion matter?

General statistical context: over-dispersion occurs when observed variation is greater than a model assumes. If it is present and ignored, standard errors and confidence intervals can be too small, making estimates appear more precise than warranted. It is particularly relevant to some count outcomes. The primary outcome is time-to-event, and the registry does not report an over-dispersion assessment or a count-outcome model for the primary analysis.

The posted primary result is an investigator-assessed analysis under FDA guidelines, according to its title. The detailed statistical plan, the exact model assumptions, any adjustment variables, and the handling of intercurrent events are not reported in the registry record. The brief summary and detailed outcome description also use different time origins for progression-free survival, so that detail should not be silently assumed away.

11. Confidence Intervals and Precision

The primary progression-free survival estimate is reported as a hazard ratio of 0.77. Its 95% confidence interval runs from 0.557 to 1.067. Because that interval includes 1, the posted result does not pin down a single direction of relative hazard with high precision: values consistent with a lower hazard and values consistent with a higher hazard are both included. The p-value is 0.113, which likewise does not meet the conventional 0.05 threshold for statistical significance.

The medians are 101.0 days for alisertib plus paclitaxel and 66.0 days for placebo plus paclitaxel. Their respective 95% confidence intervals are 80.0 to 113.0 days and 53.0 to 83.0 days. These intervals describe uncertainty in each group's median estimate. They are not a confidence interval for a difference between medians, and the registry does not report a direct median-difference estimate.

General statistical context: confidence intervals show a range of estimates compatible with the data and analysis method; they do not guarantee that the true value lies within the observed interval. More precise estimates typically have narrower intervals, while less precise estimates have wider intervals. Precision is distinct from clinical importance: a precise estimate could describe a small effect, and an imprecise estimate can leave substantial uncertainty about the effect's direction or magnitude.

The overall survival estimate is also imprecise: the reported hazard ratio is 0.93, with a 95% confidence interval of 0.652 to 1.341. This interval spans values on either side of 1. The response outcomes have their own uncertainty, and the registry's response summaries should be interpreted with their corresponding confidence intervals and analysis methods. No inference about certainty should be based solely on comparing whether separate group-specific intervals overlap.

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 table and curve are planning illustrations based on the observed effect, not the trial's original design assumptions. The original sample-size justification is not reported in the registry record. The stated illustration assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set at a stated fraction of the observed effect.

For time-to-event outcomes, the entries are total events needed across the comparison, using the Schoenfeld method; they are not participants per arm. For other outcomes, entries are participants per arm. Thus, read each row according to its outcome type rather than treating all values as the same unit.

The 80% power column gives the planned event or participant count associated with an 80% chance of detecting the specified effect under the illustration's assumptions, if that effect and the assumptions were correct. The 90% power column describes the corresponding count for a 90% chance. These are planning targets, not observed outcomes or evidence that the trial achieved a particular probability of detecting an effect. The figures are not observed power.

Rows for 100% of the observed effect use the full observed estimate as the target; the 75% and 50% rows use progressively smaller target effects. Smaller effects generally require more information to detect, so those rows require more events or participants. For example, the primary time-to-event illustration lists 460 total events for 80% power and 616 for 90% power when using 100% of the observed effect. For 75% of the observed effect, the listed totals are 818 and 1,094; for 50%, they are 1,839 and 2,462. These are illustrations of planning requirements, not a judgment about the original trial's design.

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 = 0.770, so ln(HR) = -0.2614.
2. d = 4 × (1.96 + 0.842)² / [ln(0.770)]² = 459.6, rounded up to 460 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 <- 0.77                              # 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(2038647)
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 = 460 (80% power), 616 (90% power)
 75% of observed effect: events = 818 (80% power), 1,094 (90% power)
 50% of observed effect: events = 1,839 (80% power), 2,462 (90% power)
Simulated power with 460 events: 0.783 (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 645 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 178 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 <- 2.16986301369863          # control-arm median, months (converted from the posted unit)
hr          <- 0.77          # 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: 306         p (1-sided)   0.0001
 Events: 208        ~HR at bound   0.5974
   Month: 11    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.77   0.0338
   IA 2: 67%                   Z   2.5114
      N: 524         p (1-sided)   0.0060
 Events: 415        ~HR at bound   0.7815
   Month: 19    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.77   0.5603
       Final                   Z   1.9930
      N: 646         p (1-sided)   0.0231
 Events: 623        ~HR at bound   0.8523
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.77   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 log-rank hazard ratio of 0.77, a 95% confidence interval of 0.557 to 1.067, and p = 0.113. The registry also reports that method and estimate as the primary comparison. No discrepancy between the sponsor's reported analysis and the posted primary estimate is flagged.

Recalculated results, where reported for binary secondary outcomes, are simple unadjusted calculations from posted summary statistics. They are not substitutes for the sponsor's own analyses, which may use a different statistical method or account for the trial's analysis structure. For overall response rate, the sponsor reports a weighted Cochran-Mantel-Haenszel odds ratio of 0.74, with a 95% confidence interval of 0.35 to 1.55 and p = 0.406. The unadjusted recalculation reports p = 0.455 for its risk-difference comparison. These p-values answer related but not identical statistical comparisons and should not be treated as interchangeable.

For overall survival, the sponsor reports a Cox proportional-hazards estimate of 0.93, with a 95% confidence interval of 0.652 to 1.341 and p = 0.714. The posted primary hazard ratio is likewise the sponsor's reported result, rather than an independently recalculated estimate. When a discrepancy is flagged, the sponsor's own analysis takes precedence; no discrepancy is flagged in this record. Any recalculated secondary result remains a simple unadjusted check, not a replacement for the sponsor's model-based analysis.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study assesses several outcomes or conducts several statistical comparisons. Even if each test is evaluated using a prespecified threshold, repeated testing increases the chance that at least one result will appear statistically notable by chance. Researchers may address this through an outcome hierarchy, adjusted significance thresholds, or other prespecified procedures. The registry lists 15 outcome measures and 5 analyses, but a familywise error strategy, hierarchy, or other multiplicity plan is not reported in the registry record. Therefore, secondary p-values should not be read as if they necessarily had confirmatory status.

Missing data and censoring are especially important for time-to-event outcomes. The registry describes censoring participants without progression who were last known alive at their last response assessment showing stable disease or better. The description is truncated, and the full censoring rules, reasons for missing assessments, and sensitivity analyses are not reported in the registry record. Those details matter because the interpretation of a time-to-event analysis depends on how follow-up ends and whether censoring is informative.

The detailed primary outcome description defines progression-free survival from the start of study treatment, whereas the brief summary defines it from randomization. These are different time origins. The record does not explain the wording difference or specify which description governs the posted analysis. The registry states that the primary population was intent-to-treat, defined as all randomized participants, but it does not give complete endpoint-specific analysis populations or disclose all exclusions and missing outcome data.

The trial's statistical analysis plan, prespecified handling of missing outcomes, interim-analysis procedures, and adjustment variables are not reported in the registry record. No assumptions about these procedures can be inferred from the reported estimates alone. Consequently, the posted results support a description of the reported analyses, while leaving uncertainty about some decisions that affect robustness and reproducibility.

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: febrile neutropenia (10.3%), small cell lung cancer (6.7%), neutropenia (5.7%), diarrhoea (5.7%), stomatitis (4.6%).

The registry reports serious adverse events in 39 of 87 participants at risk in the alisertib-plus-paclitaxel group and 30 of 89 participants at risk in the placebo-plus-paclitaxel group. Other adverse events are reported in 85 of 87 and 80 of 89 participants at risk, respectively. Deaths are listed for 12 participants in the alisertib group and 11 in the placebo group. These are counts as posted; they do not by themselves establish that treatment caused an event.

The most frequently listed serious adverse-event terms include febrile neutropenia at 10.3%, small cell lung cancer at 6.7%, neutropenia at 5.7%, diarrhoea at 5.7%, stomatitis at 4.6%, and pneumonia at 3.4%. The terms are not a complete description of the safety profile, and the registry excerpt does not provide event-level timing, duration, severity distributions, or a full account of attribution.

Safety comparisons need care because the numbers at risk differ by outcome group, and exposure duration can affect the opportunity to observe adverse events. The registry does not report exposure-adjusted event rates or detailed safety-analysis methods in the record. These figures describe reported events; they should not be interpreted as proof of causation or as a comprehensive assessment of comparative safety.

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

The posted primary result describes a numerical difference in median progression-free survival between the randomized groups and a hazard-ratio estimate below 1. However, the 95% confidence interval for the hazard ratio includes 1 and the reported p-value is 0.113. The data therefore do not establish a statistically significant difference in the primary outcome under the reported analysis. The interval also shows that the estimate is not precise enough to rule out a range of relative effects.

This does not prove that the regimens are identical. A non-significant result can reflect uncertainty as well as an absence of a detectable difference. The estimate and its interval should be considered together, rather than interpreting a p-value as a binary verdict on whether an effect exists. Nor can the observed median values be translated into an individual participant's expected course.

The trial comparison was between the treatment regimens as recorded, including different paclitaxel doses across groups. It cannot be described as a controlled test of alisertib added to an otherwise identical paclitaxel regimen. The registry's two different wordings for the progression-free survival time origin also warrant caution in interpreting the exact endpoint implementation.

Overall survival and response were secondary outcomes. Their posted estimates do not establish a clear difference in those outcomes, and the registry does not report whether multiple-testing procedures were used. Symptom outcomes are listed, but their numeric results are not reported in the registry record summarized in the registry record. Safety counts describe adverse events observed and do not establish which events were caused by treatment.

The findings apply to the participants, regimens, follow-up, and methods represented in this phase 2 trial. They do not by themselves establish efficacy, lack of efficacy, safety in all settings, or a definitive treatment recommendation. The registry record does not provide enough detail to resolve every analytic limitation, including the complete missing-data rules and original sample-size justification.

19. Why This Trial Matters Statistically

This trial illustrates why a time-to-event result should not be reduced to its point estimate. The primary hazard ratio is below 1, but its confidence interval includes 1 and its p-value is 0.113. Reading the estimate, uncertainty interval, and test result together makes the evidence more transparent than treating the direction of the point estimate as a conclusion.

It also illustrates the difference between a median and a hazard ratio. Medians summarize the time-to-event distribution at a particular point, while a hazard ratio compares event rates over follow-up under the model used. The two measures are related to the same endpoint but are not interchangeable, and the median estimates do not supply a direct estimate of the overall relative hazard.

The registry record provides a useful example of why endpoint definitions and treatment contrasts matter. Its brief and detailed descriptions use different time origins for progression-free survival, and paclitaxel doses differ between the treatment groups. These are features of the recorded comparison that need to be made explicit before interpreting its estimates.

Finally, the sample-size illustrations show how planning requirements depend on the effect one aims to detect and the desired power. They are not evidence of the power achieved in this trial, and they do not replace the trial's original design assumptions, which are not reported in the registry record. Separating planning calculations from observed results helps prevent overstatement. The combination of effect estimates, confidence intervals, analysis methods, and limitations provides a more informative statistical reading than any single number alone.

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 has an inconsistency in the primary endpoint's time origin: the detailed outcome description starts follow-up at study treatment, while the brief summary starts it at randomization. The record does not resolve which definition was implemented in the analysis. The detailed censoring description is also incomplete, so the full rules for participants without a documented event are not available.

The two arms did not receive the same paclitaxel dose. The alisertib-plus-paclitaxel regimen used 60 mg/m^2, while the placebo-plus-paclitaxel regimen used 80 mg/m^2. As a result, the trial estimates compare the stated treatment regimens, not the isolated contribution of alisertib under an identical chemotherapy dose.

The primary result's confidence interval includes 1, and the posted p-value is 0.113. This leaves uncertainty about the size and direction of the relative hazard. The record does not report the original sample-size justification, detailed analysis plan, interim-analysis procedures, or full model diagnostics. It also does not provide complete endpoint-specific missing-data information or the full multiplicity strategy.

Some listed secondary outcomes have no numeric result in the registry record summarized in the registry record. The safety information includes selected counts and common serious-event terms, but does not provide a complete account of event timing, exposure-adjusted rates, severity, or causality. These limits constrain what can be inferred from the registry results and should be kept distinct from the results that are reported.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Phase 2 Study of Alisertib (MLN8237) in Combination With Paclitaxel Versus Placebo in Combination With Paclitaxel as Second Line Therapy for Small Cell Lung Cancer (SCLC) — NCT02038647.
• Registry record last updated 2018-12-27; status last verified 2018-11.
• Results first posted 2018-12-27.
• 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

NCT02038647 is a completed randomized phase 2 trial of alisertib plus paclitaxel versus placebo plus paclitaxel as second-line therapy for small cell lung cancer. The registry describes a parallel, placebo-controlled, multicenter study with quadruple masking. Enrollment is listed as 178.

The primary outcome was investigator-assessed progression-free survival. The posted median values were 101.0 days and 66.0 days, and the reported hazard ratio was 0.77, with a 95% confidence interval of 0.557 to 1.067 and p = 0.113. The interval includes 1, leaving uncertainty about the relative effect.

The record lists overall survival, response, symptoms, and adverse events among secondary outcomes. Some numeric secondary results are posted, while numeric findings for symptom outcomes are not reported in the registry record summarized in the registry record. The registry also leaves important details unreported, including the trial's original sample-size justification, full analysis plan, and complete missing-data procedures.