Food Effect Study of Alisertib (MLN8237) in Participants With Advanced Solid Tumors or Lymphomas: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Food Effect Study of Alisertib (MLN8237) in Participants With Advanced Solid Tumors or Lymphomas (NCT01898078), a randomized phase 1 study. The primary outcome is Cmax: Maximum Observed Plasma Concentration of Alisertib.

2. Clinical Question

3. Trial Design

This was a randomized, open-label, Phase 1, two-way crossover study of alisertib in participants with advanced solid tumors or lymphomas. The registry lists 26 enrolled participants and describes the study as completed. Its stated purpose was to evaluate the effect of food on the single-dose pharmacokinetics of alisertib administered as an enteric-coated tablet.

In a crossover design, participants receive the conditions being compared in different periods, and the order can be randomized. In this study, one sequence began with fed administration and then switched to fasted administration; the other began fasted and then switched to fed administration. The posted regimen descriptions also include alisertib twice daily on Days 4 through 10 and a 14-day rest period in Cycle 1, described as a 24-day cycle. The primary pharmacokinetic measurement was collected on Day 1 before dosing and at multiple time points up to 72 hours after dosing in Cycles 1 and 2.

Randomizing the sequence can help distribute order-related influences across the two sequences. Open-label means the study was not masked. The registry record does not report further detail about sequence-generation procedures, allocation concealment, or how period or sequence effects were handled in the analysis. The pharmacokinetic population was defined by having sufficient dosing and pharmacokinetic concentration-time data to estimate the parameters reliably. The posted Cmax summaries include 22 participants in each condition.

The crossover structure is important when interpreting the comparison: the fed and fasted measurements are associated with participants who received both conditions, but the unadjusted recalculation uses the posted condition-level means and standard deviations. The registry record reports a separate least-squares mean ratio analysis. Its statistical model details 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.

Before assignment: Participants with a diagnosis of advanced solid tumors or lymphomas were enrolled to crossover fashion to receive alisertib 50 mg enteric-coated tablet (ECT), orally in fasted or fed state in Cycles 1 and 2.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Pharmacokinetic (PK) population included participants with sufficient dosing data and PK concentration-time data to reliably estimate the PK parameters.

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: Safety population included all enrolled participants who received at least one dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Cmax: Maximum Observed Plasma Concentration of Alisertib, measured over Cycles 1 and 2, Day 1 pre-dose and at multiple time points (up to 72 hours) post-dose.

The primary endpoint is Cmax, the maximum observed plasma concentration of alisertib, measured in nM. It is a continuous pharmacokinetic measure: rather than recording a clinical event, it describes the highest concentration observed in plasma after administration. The registry specifies measurement before dosing and at multiple post-dose time points up to 72 hours in Cycles 1 and 2.

In general, a maximum concentration measure is used to characterize the peak of a drug’s observed concentration-time profile. Comparing this measure across fed and fasted conditions addresses whether the observed peak concentration differs with the dosing condition. Cmax is only one part of a pharmacokinetic profile; it does not, by itself, describe the full course of concentrations over time.

The registry reports the endpoint as a mean, with standard deviation as the measure of spread. For the fed condition, the posted mean is 1339.0 nM and the standard deviation is 487.2; for the fasted condition, the mean is 1354.0 nM and the standard deviation is 432.2. These are condition-specific summaries in the pharmacokinetic population. They should be read alongside the reported least-squares mean ratio and its confidence interval, which describe the registry’s own comparative analysis.

7. Results Posted in the Registry

• Alisertib 50 mg Fasted: mean 1,354 versus 1,339 on control; difference 15.90 (95% CI -256.25 to 288.05), standardized effect d = 0.03; z-test p = 0.909.

8. Statistical Interpretation

9. Statistical Methods Explained

For Cmax, the registry reports a mean and standard deviation for each dosing condition. A mean describes the central value of the observed measurements, while a standard deviation describes how spread out individual measurements are around that mean. The fed-state mean is 1339.0 nM with standard deviation 487.2, and the fasted-state mean is 1354.0 nM with standard deviation 432.2. The analysis population is the pharmacokinetic population, defined in the record by adequate dosing and concentration-time data for reliable parameter estimation.

In general, a mean difference expresses the distance between two group means on the original measurement scale. The unadjusted recalculation reports a mean difference of 15.9 nM for fasted versus fed administration. Its 95% confidence interval is -256.3 to 288.1, and its unadjusted p-value is 0.909. The interval is not a range containing a known fixed effect with certainty; it reflects the uncertainty associated with the estimate under the calculation’s assumptions. Because this interval includes zero, it does not indicate a clearly separated mean difference on this unadjusted scale.

In general, a p-value describes how compatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, nor is it a measure of clinical importance. The unadjusted p-value is 0.909. The registry-reported analysis has no posted p-value, so the recalculated p-value should not be attributed to the sponsor’s least-squares mean ratio analysis.

The registry reports the comparative result as a least-squares mean ratio of 0.97 with a 90% confidence interval from 0.85 to 1.12. In general, a ratio compares estimated values on a relative scale, with 1 representing equal values in the ratio’s numerator and denominator. The registry record does not specify the direction of the ratio in enough detail to assign its numerator and denominator beyond the reported groups, and the model method is not reported in the registry record. The interval spans 1, but interpretation as equivalence would require a defined equivalence margin; such a margin is not reported in the registry record.

The study’s crossover model means the same participants were intended to receive both fed and fasted conditions, in randomized order. In general, paired analysis can use within-participant information and can account for the fact that two measurements from one participant may be related. The unadjusted recalculation uses posted group-level means and standard deviations rather than participant-level paired observations. The registry record does not report the correlation between paired measurements or enough detail to reconstruct a paired analysis. The recalculation is therefore a simple unadjusted comparison, not a recreation of the sponsor’s model.

Randomized sequence assignment and crossover treatment periods are features of the study design, but the registry record does not report whether the sponsor’s analysis adjusted for period, sequence, or other design terms. The reported method is not specified. This limits the ability to assess whether the posted ratio accounts for the crossover structure in a particular way. It also means that the unadjusted result should be interpreted as a separate summary-based calculation rather than as the definitive analysis of the randomized crossover design.

Multiplicity concerns arise when several outcomes, comparisons, or analyses are evaluated. In general, making multiple statistical tests can increase the chance of observing at least one apparently notable result by chance if no adjustment or prespecified hierarchy is used. The registry lists six outcome measures and three analyses, but multiplicity procedures, the prespecified analysis hierarchy, and any adjustment are not reported in the registry record. The record therefore does not establish whether multiplicity was addressed.

Missing data and analysis populations also affect interpretation. The primary endpoint population is described as participants with sufficient dosing and pharmacokinetic concentration-time data to estimate parameters reliably. The posted condition summaries each include 22 participants. The registry record does not report how many participants were excluded from the pharmacokinetic population, why any data were missing, how missing measurements were handled, or whether sensitivity analyses were performed. Those details matter because the available summaries may not represent all enrolled participants.

In general, over-dispersion means that observed variability is greater than a model assumes. It matters when a statistical model relies on a variance assumption that does not fit the data, because uncertainty estimates can then be inaccurate. The posted summaries show standard deviations, but the registry record does not report an over-dispersion assessment or indicate that over-dispersion was part of the analysis.

The study also reports adverse-event outcomes, but these are not the primary pharmacokinetic endpoint. Safety summaries are reported by randomized sequence groups, while the primary result compares fed and fasted conditions. The sequence-group safety figures cannot be read as a direct fed-versus-fasted safety comparison from the information reported.

10. Confidence Intervals and Precision

The registry-reported least-squares mean ratio is 0.97, with a 90% confidence interval from 0.85 to 1.12. This interval expresses uncertainty around the ratio estimate under the sponsor’s analysis. In general, a confidence interval gives a range of values compatible with the observed data and the method’s assumptions; it is not a guarantee that the true value lies in that range.

The interval includes 1, the value representing equal quantities on a ratio scale. Its lower and upper limits lie on opposite sides of 1, so the estimate does not tightly pin down the direction or size of a relative difference. The registry record does not report an equivalence margin or a decision threshold for this ratio. Consequently, the interval alone cannot establish that fed and fasted administration are equivalent.

The unadjusted recalculation gives a mean difference of 15.9 nM with a 95% confidence interval from -256.3 to 288.1. This interval includes zero and covers a broad span of possible differences on the concentration scale. Its limits indicate that the point estimate is not precise enough to identify a narrow range of plausible mean differences from the posted summary statistics.

The two intervals should not be compared as though they were produced by the same analysis: the sponsor’s result is a least-squares mean ratio with a 90% interval, whereas the recalculation is an unadjusted mean difference with a 95% interval. The registry record does not report the model details needed to assess the sponsor’s interval construction. The uncertainty is therefore best described separately for each result.

11. 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 participants per arm, not total events. They are illustrations based on the observed effect and are not the trial’s original sample-size calculation. The listed assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; each planning scenario targets the stated fraction of the observed effect.

The 80% power column describes the planned number of participants per arm under the stated assumptions for a design intended to have an 80% chance of detecting the chosen effect if that effect is real and the assumptions hold. The 90% power column describes the corresponding number per arm for a design intended to have a 90% chance. These are planning targets, not guarantees of a result.

The scenarios are planning for 100% of the observed effect, planning for 75% of the observed effect, and planning for 50% of the observed effect. The listed participants per arm for 80% power are 13,170, 23,413, and 52,678, respectively; for 90% power, they are 17,631, 31,343, and 70,521, respectively. The pattern reflects a general design principle: smaller effects are harder to detect and require more participants under otherwise unchanged assumptions.

These figures are not observed power. They are planning illustrations derived from the observed effect, not evidence about the probability that the completed study detected an effect. The registry record does not report the trial’s original sample-size justification or its original design assumptions.

12. 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

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 15.900; pooled standard deviation σ = 460.534.
2. n = 2 × (1.96 + 0.842)² × 460.534² / 15.900² = 13,169.4, rounded up to 13,170 participants per arm 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
delta <- 15.9000000000001                      # observed difference in means
sigma <- 460.53418515893                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1898078)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 13,170 (80% power), 17,631 (90% power)
 75% of observed effect: n per arm = 23,413 (80% power), 31,343 (90% power)
 50% of observed effect: n per arm = 52,678 (80% power), 70,521 (90% power)
power.t.test: 13170.4 
Simulated power at n = 13,170 per arm: 0.805 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The registry’s own comparative analysis is a least-squares mean ratio of 0.97, with a 90% confidence interval from 0.85 to 1.12. The analysis method and a p-value for this result are not reported in the registry record. The ratio and interval are the sponsor-reported result available in the record.

The unadjusted recalculation instead compares the posted condition-level means using summary statistics. It gives a mean difference of 15.9 nM, a 95% confidence interval from -256.3 to 288.1, and an unadjusted p-value of 0.909. It is a simple unadjusted recalculation, not a reproduction of the least-squares mean ratio analysis. In particular, it does not incorporate participant-level paired information from the crossover design.

The estimates use different scales and confidence levels, so they are not direct substitutes. The sponsor’s own analysis takes precedence when describing the sponsor-reported result. The recalculation is useful as a separate check on the posted condition summaries, but it cannot establish what model produced the sponsor’s result or why the estimates differ in form.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists six outcome measures and three analyses. The primary endpoint is Cmax, while the listed secondary outcomes include adverse events, serious adverse events, clinically significant laboratory changes reported as adverse events, and clinically significant vital-sign changes reported as adverse events. The registry record does not report a multiplicity adjustment, a prespecified hierarchy of hypotheses, or how the analyses were selected for emphasis.

In general, multiplicity refers to drawing conclusions across several statistical tests or outcomes. If many tests are conducted, the chance of at least one apparently notable result can increase unless the analysis plan accounts for that testing. The registry information does not establish whether that issue was addressed in the sponsor’s analysis. No particular adjustment or testing strategy can be inferred from the posted result.

The primary endpoint population is described as participants with sufficient dosing and pharmacokinetic concentration-time data to estimate pharmacokinetic parameters reliably. The condition-specific summaries each include 22 participants, while the registry reports 26 enrolled participants. The registry record does not report the reasons for any difference between enrollment and the pharmacokinetic summaries, the number with missing measurements, the handling of missing data, or sensitivity analyses for missingness.

The record also does not report whether all randomized participants were included in the primary analysis, whether any participants were excluded after randomization, or how incomplete crossover periods were handled. These details matter because availability of both condition measurements can affect the information contributed by a participant in a crossover comparison.

The registry describes a randomized crossover design and reports a least-squares mean ratio, but the model method, treatment of period and sequence, and any assumptions about carryover are not reported in the registry record. The extent to which the analysis accounts for the study structure therefore cannot be evaluated from the posted statistical details alone.

15. 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: abdominal pain (15.4%), rectal haemorrhage (7.7%), cellulitis (7.7%), bacteraemia (7.7%), pneumonia streptococcal (7.7%).

The registry reports safety figures by randomized sequence group, not separately by fed and fasted treatment condition. In each sequence group, 13 participants were at risk for serious adverse events, and 7 were reported as affected. For other adverse events, 13 participants were at risk and 13 were affected in each sequence group. The record reports zero deaths in the fed-then-fasted sequence and one death in the fasted-then-fed sequence.

The listed serious adverse event terms include abdominal pain at 15.4%, rectal haemorrhage at 7.7%, cellulitis at 7.7%, bacteraemia at 7.7%, pneumonia streptococcal at 7.7%, and pyelonephritis at 7.7%. These percentages are reported as posted and should not be converted into counts or attributed to a particular dosing condition on the basis of this record.

These summaries describe events observed in the sequence groups. They do not provide a direct fed-versus-fasted comparison of safety, and the record does not report a causal assessment linking individual events to either condition. The small sequence-group denominators also mean that the figures are descriptive rather than a basis for precise comparative safety estimates.

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

The results address a narrow pharmacokinetic question: how Cmax after alisertib administration compares between fed and fasted conditions in the study population. The condition-specific mean values are 1339.0 nM in the fed state and 1354.0 nM in the fasted state. The registry’s least-squares mean ratio is 0.97, with a 90% confidence interval from 0.85 to 1.12. The unadjusted recalculation gives a mean difference of 15.9 nM, with a 95% confidence interval from -256.3 to 288.1.

These estimates do not establish that the conditions produce identical Cmax values. The confidence intervals retain uncertainty about the magnitude and direction of a difference, and the record does not report an equivalence margin. Nor do the results establish that any concentration difference is clinically important or unimportant. Cmax is a pharmacokinetic measure, not a clinical outcome.

The findings also do not answer broader questions about the effects of alisertib in advanced solid tumors or lymphomas. This study was designed to evaluate the effect of food on single-dose pharmacokinetics; the registry does not describe it as an efficacy comparison. No conclusion about tumor response, survival, or treatment benefit can be drawn from the Cmax results.

The safety summaries are reported by treatment sequence and do not provide a direct comparison of adverse events between fed and fasted administration. The number of participants and the descriptive format also limit what can be inferred from those figures. Overall, the posted results characterize a pharmacokinetic comparison and offer limited descriptive safety information, while leaving uncertainty about the comparative estimate and its clinical significance.

17. Why This Trial Matters Statistically

This study illustrates how the choice of statistical scale affects interpretation. The registry reports a least-squares mean ratio, whereas the unadjusted recalculation reports a mean difference. A ratio describes relative contrast, while a difference describes contrast in the original measurement units. Both can help summarize a comparison, but they answer the question in different forms and their intervals cannot be read as if they were the same statistic.

The crossover design also highlights the importance of using the structure of the data. Participants received both fed and fasted conditions in randomized order, so the measurements are potentially paired within participants. A summary-based unadjusted calculation cannot make full use of that participant-level pairing. The registry’s ratio analysis is distinct, but its model method is not reported, limiting assessment of how the design was handled.

Precision is central to the interpretation. The point estimates are near the values representing little difference on their respective scales, but the confidence intervals remain broad enough to include a range of possible differences. A result near a null value is not, by itself, evidence of equivalence. That conclusion requires a defined margin and an analysis designed to assess it; such a margin is not reported in the registry record.

Finally, this case separates pharmacokinetic evidence from clinical outcomes. Cmax is useful for describing peak plasma concentration, but it does not establish the clinical consequences of a concentration pattern. Reading the endpoint, design, analysis scale, and uncertainty together helps keep the interpretation aligned with the question the study was designed to address.

18. 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 principal limitation for interpretation is the incomplete reporting of the comparative statistical method. The registry provides a least-squares mean ratio and confidence interval, but the method is not specified. It does not report the model terms, the handling of period or sequence, or how the crossover structure was incorporated. This prevents an independent assessment of the assumptions behind the reported ratio.

The pharmacokinetic summaries include 22 participants per condition, while the registry reports 26 enrolled. The reasons for this difference, the amount and pattern of missing pharmacokinetic data, and the handling of incomplete observations are not reported in the registry record. The endpoint population requires sufficient dosing and concentration-time data, which means that the analysis is based on a defined subset rather than necessarily every enrolled participant.

The unadjusted recalculation uses group-level means and standard deviations, not participant-level paired measurements. It therefore does not reflect the within-participant association that may be relevant in a crossover study. The registry record does not report paired data or the correlation needed to assess that contribution from the posted summaries.

The reported confidence intervals use different scales and confidence levels: the registry analysis is a 90% interval for a ratio, while the recalculation is a 95% interval for a mean difference. They should not be treated as equivalent evidence. No equivalence margin, multiplicity approach, or original sample-size justification is reported in the registry record.

Safety results are presented by sequence group rather than by fed and fasted condition, limiting condition-specific comparisons. The study’s purpose and primary endpoint concern single-dose pharmacokinetics, so the findings do not answer clinical efficacy questions.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Food Effect Study of Alisertib (MLN8237) in Participants With Advanced Solid Tumors or Lymphomas — NCT01898078.
• Registry record last updated 2019-06-21; status last verified 2019-03.
• Results first posted 2019-06-21.
• 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.

22. Record Summary

NCT01898078 is a completed, randomized, open-label, Phase 1 crossover study of the food effect on alisertib pharmacokinetics in participants with advanced solid tumors or lymphomas. The registry reports 26 enrolled participants. The primary endpoint is Cmax, the maximum observed plasma concentration, in nM.

The posted condition means are 1339.0 nM for fed administration and 1354.0 nM for fasted administration, with 22 participants in each summary. The registry reports a least-squares mean ratio of 0.97 and a 90% confidence interval from 0.85 to 1.12. An unadjusted recalculation reports a mean difference of 15.9 nM with a 95% confidence interval from -256.3 to 288.1.

The record also includes adverse-event summaries by randomized sequence group. Important details of the sponsor’s statistical method, missing-data handling, multiplicity approach, and original sample-size justification are not reported in the registry record.