A Study of Abemaciclib (LY2835219) Alone or in Combination With Other Agents in Participants With Previously Treated Pancreatic Ductal Adenocarcinoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Abemaciclib (LY2835219) Alone or in Combination With Other Agents in Participants With Previously Treated Pancreatic Ductal Adenocarcinoma (NCT02981342), a randomized phase 2 study. The primary outcome is Stage 1: Disease Control Rate (DCR): Percentage of Participants With a Best Overall Response of Complete Response (CR), Partial Response (PR) or Stable Disease (SD).

2. Clinical Question

3. Trial Design

This was a randomized, open-label, parallel-group phase 2 study. The registry describes the study as adaptive in its official title, but the details of any adaptations are not reported in the registry record. Masking was listed as none. The study was sponsored by Eli Lilly and Company and is marked completed. The recorded start date is 2017-01-12, and the primary completion date is 2018-04-20.

The registry records enrollment of 106 participants. The primary disease control results cover three groups, each with 33 participants: 200mg abemaciclib, 150mg abemaciclib plus 150mg LY3023414, and gemcitabine or capecitabine. The primary outcome population is described as all randomized participants. How the enrolled total relates to the analyzed group counts is not reported in the registry record.

The trial compared abemaciclib monotherapy and the abemaciclib plus LY3023414 combination with a choice of standard of care. The registry also lists a safety lead-in group receiving 150mg abemaciclib plus 150mg galunisertib; this group appears in the safety data but not among the three primary outcome groups. Details of allocation procedures, stratification, protocol changes, or the rationale for the adaptive design are not reported in the registry record.

The stated purpose was to evaluate safety and efficacy in participants with previously treated metastatic pancreatic ductal adenocarcinoma. The registry lists 15 outcome measures and 64 analyses. The full statistical analysis plan and the original trial design assumptions 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: Per protocol, no efficacy analysis was planned for safety lead in. Purpose of safety lead in was only safety evaluation. All efficacy was done on randomized pts.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants.

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.

6. Primary Endpoint

The registry lists the primary outcome as Stage 1: Disease Control Rate (DCR): Percentage of Participants With a Best Overall Response of Complete Response (CR), Partial Response (PR) or Stable Disease (SD), measured over Baseline to Measured Progressive Disease or Start of New Anticancer Therapy (Up to 6 Months).

The primary endpoint was disease control rate (DCR), defined as the percentage of participants with a best overall response of complete response, partial response, or stable disease. The registry says response was assessed using RECIST v1.1, with the timeframe extending from baseline to measured progressive disease or start of new anticancer therapy, up to 6 months.

Complete response requires disappearance of all target and non-target lesions and no new lesions. Partial response requires at least a 30% decrease in the sum of the longest diameters of target lesions relative to baseline, without progression of non-target lesions or new lesions. Stable disease is neither enough shrinkage for partial response nor enough increase for progressive disease, with no progression of non-target lesions or new lesions. Progressive disease includes at least a 20% increase in the sum of target lesion diameters from the smallest sum on study, with an absolute increase of at least 5 mm, unequivocal progression of non-target lesions, or one or more new lesions.

As a general statistical concept, DCR combines tumor shrinkage with disease that meets the stable-disease definition. It is used to describe the proportion meeting a prespecified response category within a defined assessment period. It does not, by itself, describe how long disease control lasted for each participant. The registry lists a separate clinical benefit rate outcome with a duration requirement for stable disease, as well as duration of response and overall survival outcomes.

7. Results Posted in the Registry

• 150mg Abemaciclib + 150mg LY3023414: 12.1% (4/33) versus 15.2% (5/33) on control; risk difference -3.0 percentage points (95% CI -19.6 to 13.5), risk ratio 0.80 (0.24–2.72); unadjusted two-proportion z-test p = 0.720.
• Gemcitabine or Capecitabine: 36.4% (12/33) versus 15.2% (5/33) on control; risk difference 21.2 percentage points (95% CI 0.7 to 41.7), risk ratio 2.40 (0.95–6.05); unadjusted two-proportion z-test p = 0.049.

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.

Stage 1: Objective Response Rate (ORR): Percentage of Participants With a Best Overall Response (BOR) of CR or PR

• 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:
  • 150mg Abemaciclib + 150mg LY3023414: 0.0% (0/33) versus 3.0% (1/33) on control; risk difference -3.0 percentage points (95% CI -8.9 to 2.8); unadjusted two-proportion z-test p = 0.314.
  • Gemcitabine or Capecitabine: 3.0% (1/33) versus 3.0% (1/33) on control; risk difference 0.0 percentage points (95% CI -8.3 to 8.3), risk ratio 1.00 (0.07–15.33); unadjusted two-proportion z-test p = 1.000.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 150mg Abemaciclib + 150mg LY3023414 and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Gemcitabine or Capecitabine and the control. The interval gives the range of effects that remain compatible with the data.

Stage 2: Clinical Benefit Rate (CBR): Percentage of Participants With Best Overall Response of CR, PR, or SD With Duration of SD for at Least 6 Months

• 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:
  • 150mg Abemaciclib + 150mg LY3023414: 0.0% (0/33) versus 3.0% (1/33) on control; risk difference -3.0 percentage points (95% CI -8.9 to 2.8); unadjusted two-proportion z-test p = 0.314.
  • Gemcitabine or Capecitabine: 3.0% (1/33) versus 3.0% (1/33) on control; risk difference 0.0 percentage points (95% CI -8.3 to 8.3), risk ratio 1.00 (0.07–15.33); unadjusted two-proportion z-test p = 1.000.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 150mg Abemaciclib + 150mg LY3023414 and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Gemcitabine or Capecitabine and the control. The interval gives the range of effects that remain compatible with the data.

Stage 2: 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:
  • 200mg Abemaciclib vs Gemcitabine or Capecitabine: hazard ratio 1.60 (95% CI 0.78–3.27), reported p = 0.194 (as reported in the registry).
  • 150mg Abemaciclib + 150mg LY3023414 vs Gemcitabine or Capecitabine: hazard ratio 1.53 (95% CI 0.75–3.15), reported p = 0.248 (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 200mg Abemaciclib vs Gemcitabine or Capecitabine and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between 150mg Abemaciclib + 150mg LY3023414 vs Gemcitabine or Capecitabine and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover response, clinical benefit, pharmacokinetics, and survival. The registry reports objective response rate and clinical benefit rate results with very low percentages across the three main groups. Their overall pattern is consistent with the fact that the primary disease control endpoint is a broader category than objective response: objective response requires complete or partial response, while disease control also includes stable disease. Clinical benefit has a further duration requirement for stable disease. These outcomes address related but distinct questions and should not be treated as interchangeable confirmations of the primary result.

The overall survival results are also uncertain. The registry reports median overall survival for the two abemaciclib-containing groups, but not a median value for standard of care; its confidence interval has a lower limit of 2.53 and no reported upper limit. The reported hazard-ratio confidence intervals are broad and include 1, and the log-rank p-values are 0.194 and 0.248. These analyses do not give a precise estimate of the survival comparisons.

The registry lists pharmacokinetic outcomes for abemaciclib and its metabolites and for LY3023414. Numerical results for those pharmacokinetic outcomes are not reported in the registry record. Pharmacokinetic measurements describe drug exposure rather than clinical response, so they answer a different question from DCR, ORR, or survival.

Secondary outcomes can help describe the range of observed results, but unadjusted secondary tests carry less weight than a prespecified primary analysis. When many outcomes or comparisons are examined, the chance of at least one apparently notable result can increase. Whether the sponsor adjusted for multiplicity, and how, is not reported in the registry record. The secondary results therefore add context but do not remove the uncertainty in the primary comparison.

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

• Stage 1: Objective Response Rate (ORR): Percentage of Participants With a Best Overall Response (BOR) of CR or PR
• Stage 1: Pharmacokinetics (PK): Mean Steady State Exposure of Abemaciclib and Its Metabolites (LSN2839567 (M2), LSN3106726 (M20))
• Stage 1: PK: Area Under the Curve (AUC) (AUC[Tau]) of LY3023414
• Stage 1: PK: Maximum Concentration (Cmax) at Steady State of LY3023414
• Stage 2: Disease Control Rate (DCR): Percentage of Participants With a Best Overall Response of CR, PR, and SD
• Stage 2: Clinical Benefit Rate (CBR): Percentage of Participants With Best Overall Response of CR, PR, or SD With Duration of SD for at Least 6 Months
• Stage 2: Duration of Response (DoR)
• Stage 2: Overall Survival (OS)

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports the primary endpoint as a percentage of participants and gives 95% confidence intervals. The posted comparison calculations also include unadjusted risk differences and p-values. The sponsor’s analyses are identified as Cochran-Mantel-Haenszel tests. For overall survival, the sponsor reports log-rank analyses and hazard ratios. These methods address different types of outcomes and should not be interpreted as interchangeable.

What does a risk difference describe?

As a general statistical concept, a risk difference subtracts one group’s event proportion from another group’s event proportion. A value of zero represents equal observed proportions; positive or negative values indicate which group had the higher observed proportion according to the subtraction order. In the posted recalculations, the comparison is between the stated pair of groups, and the interval gives a range of values compatible with the data under the calculation’s assumptions. The registry gives the arm percentages and the recalculated risk differences; it does not report an adjusted risk difference for the primary endpoint.

How should a p-value be read?

As a general explanation, a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the test assumptions held. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. The registry identifies the sponsor’s primary comparison method as Cochran-Mantel-Haenszel, but further details of the test setup are not reported in the registry record. The sponsor’s p-values and the unadjusted recalculated p-values differ for one primary comparison, so the sponsor’s analysis takes precedence where they disagree.

Why use a confidence interval?

As a general statistical concept, a confidence interval communicates estimation uncertainty as well as a central estimate. Wider intervals indicate less precision. For a risk difference, an interval that includes zero is compatible with no difference under the interval procedure, as well as with positive or negative differences when those values are included. An interval that does not include zero still does not establish the size, importance, or cause of a difference. The posted primary endpoint intervals are 95% confidence intervals.

What does the Cochran-Mantel-Haenszel method add?

As a general explanation, the Cochran-Mantel-Haenszel method can compare categorical outcomes while accounting for defined strata. The registry names this method for the sponsor’s primary outcome comparisons, but the strata and their role in the analysis are not reported in the registry record. It would therefore be inappropriate to infer which factors were controlled for or how the reported p-values were calculated in detail.

How does the survival analysis differ?

Overall survival is a time-to-event outcome, with death from any cause as the event in the registry description. The sponsor reports a log-rank test and hazard ratios for comparisons with standard of care. In general, the log-rank test compares event-time patterns between groups, while a hazard ratio expresses a relative rate of events under the assumptions of the survival model. A hazard ratio is not a difference in median survival. The registry reports 95% confidence intervals and p-values for these comparisons, but the detailed survival analysis plan is not reported in the registry record.

Why can small groups produce uncertain estimates?

As a general statistical concept, estimates based on limited numbers of participants or events can vary substantially from one study to another. This is reflected in confidence intervals, which may be wide even when a point estimate appears distinct. Here, each primary outcome group lists 33 participants, and the confidence intervals for the group percentages extend across broad ranges. The observed percentages should therefore be read alongside their uncertainty rather than as exact underlying rates.

What does over-dispersion mean?

As a general concept, over-dispersion occurs when observed variability is greater than a statistical model expects. It can make standard errors too small if a model does not account for the extra variability. The primary outcome is reported as a binary percentage, and no over-dispersion assessment or model-based variance adjustment is reported in the registry record. It is not possible to determine from the registry whether over-dispersion was evaluated in the sponsor’s analysis.

What is the distinction between the posted analysis and recalculation?

The recalculated comparisons use the posted group summaries to produce simple, unadjusted estimates. They are not a substitute for the sponsor’s own analysis, which may use analysis details not included in the registry summary. The record marks a discrepancy, and the sponsor’s reported analysis takes precedence where results differ. The recalculations are useful as a transparent check on the posted summaries, but they should not be read as the trial’s prespecified primary analysis.

11. Confidence Intervals and Precision

The primary group percentages have broad 95% confidence intervals. For 200mg abemaciclib, the interval is 2.9% to 27.4%; for 150mg abemaciclib plus 150mg LY3023414, it is 1.0% to 23.3%; and for gemcitabine or capecitabine, it is 20.0% to 52.8%. These ranges convey that the estimated percentages are not precise point values. They also show that the underlying values compatible with the data span a meaningful range within each group.

For the unadjusted comparison of the combination with abemaciclib alone, the risk difference is -0.030 and its 95% confidence interval runs from -0.196 to 0.135. The interval includes zero and extends in both directions. For standard of care compared with abemaciclib alone, the risk difference is 0.212, with a 95% confidence interval from 0.007 to 0.417. This interval is wide relative to its estimate, so the precise size of the observed difference remains uncertain.

As a general statistical explanation, a confidence interval reflects sampling uncertainty under the method used; it is not a guarantee that the true value lies within the particular interval. A point estimate alone can conceal that uncertainty. The comparison intervals also do not incorporate unreported sources of uncertainty, such as the consequences of missing outcomes or multiple testing, if any. Details of those issues are not reported in the registry record.

For overall survival, the reported hazard ratio for 200mg abemaciclib versus standard of care is 1.600, with a 95% confidence interval from 0.782 to 3.272. For the combination versus standard of care, it is 1.533, with a 95% confidence interval from 0.746 to 3.150. Both intervals are broad and include 1. The registry also reports p-values of 0.194 and 0.248, respectively. Together, these results do not provide a precise estimate of the relative survival comparison.

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 calculations are planning illustrations based on observed effects, not the trial’s original design justification. The registry’s planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. For time-to-event calculations, the entries are total events. For the other calculations, the entries are participants per arm. The primary endpoint planning figures are therefore participants per arm, while the overall-survival planning figures are total events.

The 80% power column describes the number of participants per arm, or total events for time-to-event outcomes, that the illustration estimates would be needed to have an 80% chance of detecting the specified effect under the stated assumptions if that effect were present. The 90% power column expresses the corresponding requirement for a 90% chance. Greater power generally requires more information, which can mean more participants or events.

The rows planning for 100%, 75%, and 50% of the observed effect vary the assumed effect size. The smaller-effect rows require more participants per arm or more total events. For example, the comparison of standard of care with abemaciclib alone lists 66 participants per arm for 80% power when planning for 100% of the observed effect, and 226 participants per arm when planning for 50% of that effect. These are hypothetical planning quantities, not participants enrolled in this study. The original trial’s sample-size assumptions and justification are not reported in the registry record.

As a general concept, power is a design characteristic tied to an assumed effect and analysis, not a measure of whether a completed study found a result. These calculations are not observed power. Their purpose is to illustrate how required sample size changes under the listed planning assumptions and effect scenarios. The results depend on those assumptions and should not be treated as a recommendation for a future study without a separate design assessment.

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

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.152; treatment proportion p₁ = 0.121; average p̄ = 0.136.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4853 = 0.9512.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4848 = 0.4081.
4. n = (0.9512 + 0.4081)² / (-0.0303)² = 2,012.1, rounded up to 2,013 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
p0 <- 0.151515151515152                          # control proportion
p1 <- 0.121212121212121                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(2981342)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 2,013 (80% power), 2,694 (90% power)
 75% of observed effect: n per arm = 3,662 (80% power), 4,901 (90% power)
 50% of observed effect: n per arm = 8,425 (80% power), 11,278 (90% power)
power.prop.test: 2012.1 
Simulated power at n = 2,013 per arm: 0.803 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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.

14. Analyses Reported by the Sponsor

The sponsor’s primary outcome comparisons are reported as Cochran-Mantel-Haenszel analyses. For 200mg abemaciclib versus gemcitabine or capecitabine, the registry reports p = 0.050. For 150mg abemaciclib plus 150mg LY3023414 versus gemcitabine or capecitabine, it reports p = 0.023. The registry does not provide a sponsor-reported effect estimate or confidence interval for these primary comparisons.

The unadjusted recalculation based on posted summary counts gives p = 0.049 for standard of care compared with abemaciclib alone, and p = 0.72 for the combination compared with abemaciclib alone. The combination comparison therefore differs materially from the sponsor’s reported result. The registry marks a discrepancy between the sponsor analysis and the recalculation. The sponsor’s own analysis takes precedence where the results disagree; the recalculation is a simple unadjusted check and cannot reproduce analysis details that are not reported in the registry record.

The reported sponsor p-values do not, by themselves, give the size or precision of the primary comparisons. The registry also does not report enough detail to establish how any strata, missing data, or multiplicity were handled. These omissions limit the ability to reconstruct or fully assess the sponsor’s inferential analysis from the registry summary.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 15 outcome measures and 64 analyses. It includes a primary endpoint and multiple secondary endpoints, with comparisons among treatment groups. Examining multiple outcomes and comparisons raises the general statistical issue of multiplicity: the more hypotheses tested, the greater the chance of obtaining at least one result that appears unusual by chance alone if no effect exists.

The registry record does not report whether multiplicity was controlled, which testing hierarchy may have been used, or whether the reported p-values were adjusted. The sponsor’s primary comparisons are labeled as superiority analyses, but details of any family-wise error or false-discovery procedure are not reported in the registry record. It is therefore not possible to interpret the displayed p-values as necessarily accounting for all the analyses conducted.

The primary endpoint population is described as all randomized participants. Beyond that statement, the registry record does not report details of analysis populations, exclusions, or how participants with missing response assessments were handled. The number and reasons for missing primary outcome data are not reported in the registry record. No imputation method, sensitivity analysis, or missing-data assumption is reported in the registry record.

The registry also does not report the full statistical analysis plan, the original sample-size justification, or details of interim analyses. Although the study is described as adaptive in its official title, the adaptation procedures are not reported in the registry record. These gaps matter because planned adaptations, multiplicity control, and outcome handling can affect how estimates and tests should be interpreted. Without those details, the reported results can be described, but the full inferential framework cannot be reconstructed from the registry summary.

16. 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: vomiting (15.4%), anaemia (14.3%), gastric ulcer haemorrhage (14.3%), asthenia (14.3%), fatigue (14.3%).

The registry reports safety groups with counts at risk and counts affected. For serious adverse events, the group titled 200mg abemaciclib lists 17 affected among 32 at risk; the 150mg abemaciclib plus 150mg LY3023414 group lists 18 among 33; and the gemcitabine or capecitabine group lists 15 among 26. A separate 150mg abemaciclib plus 150mg galunisertib safety lead-in group lists 4 affected among 7 at risk. The safety group sizes differ from the primary outcome group sizes.

For other adverse events, the registry lists 30 affected among 32 at risk in the 200mg abemaciclib group, 33 among 33 in the abemaciclib plus LY3023414 group, 25 among 26 in the standard-of-care group, and 7 among 7 in the galunisertib lead-in group. The record also lists deaths by group, but these counts should not be interpreted as deaths caused by treatment; the registry does not provide that attribution in these figures.

The most frequent serious adverse event terms listed include vomiting at 15.4%, and anaemia, gastric ulcer haemorrhage, asthenia, fatigue, and systemic inflammatory response syndrome at 14.3% each. These are registry-listed terms and percentages. The figures do not, on their own, establish comparative safety or causation, and differences in group sizes and follow-up should be considered. Further detail about event definitions and exposure-adjusted analyses is not reported in the registry record.

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

The posted results describe disease control proportions observed in three randomized groups during a defined assessment period. The registry reports the highest DCR percentage in the gemcitabine or capecitabine group, with lower percentages in the two abemaciclib-containing groups. The group estimates have broad confidence intervals, and the sponsor’s reported analyses and unadjusted recalculations do not fully agree. Taken together, the results should be read as estimates with meaningful uncertainty, not as exact measurements of what would occur in other groups of participants.

The endpoint combines complete response, partial response, and stable disease. It does not convey the duration of stable disease, duration of response, or overall survival by itself. The registry reports separate secondary outcomes addressing some of those questions, but the survival comparisons are also imprecise. Pharmacokinetic outcomes, where reported, would describe exposure rather than clinical benefit; numerical results for those measures are not reported in the registry record.

The comparison is limited to the study groups, eligibility context, outcome definitions, and follow-up periods described in the registry. It does not establish why the group percentages differed. The open-label design and the absence of reported details about missing data, multiplicity, and the full analysis plan constrain interpretation. The record marks a discrepancy between the sponsor’s primary analysis and a simple recalculation, and the sponsor’s analysis takes precedence where they disagree.

These statistical results are not medical advice and do not establish that a treatment should or should not be used. They also do not establish that the same percentages or comparisons would apply to people outside the trial or to different care settings. The appropriate statistical conclusion is limited to describing the posted estimates, their uncertainty, and the analysis information available in the registry record.

18. Why This Trial Matters Statistically

This trial illustrates why percentages should be read alongside their denominators and uncertainty. Each main primary outcome group lists 33 participants, and the group estimates have broad 95% confidence intervals. A displayed percentage can look definitive even when the range compatible with the data is wide. Reporting both the estimate and interval makes that uncertainty visible.

It also illustrates the distinction between a sponsor’s reported analysis and a transparent unadjusted recalculation. For one primary comparison, the recalculated p-value differs from the sponsor’s reported p-value. The registry identifies this discrepancy, so the recalculation should not be mistaken for the sponsor’s analysis. Clear separation between adjusted or method-specific sponsor results and simple summary-statistic checks is important for reproducibility and interpretation.

The trial’s multiple outcomes and analyses raise questions about multiplicity and analysis populations. The registry record does not describe how multiple testing or missing outcomes were handled. Those omissions do not erase the reported results, but they limit the strength of conclusions that can be drawn from individual p-values, especially among secondary outcomes.

Finally, the sample-size illustrations show how planning requirements depend on the effect size chosen and the outcome being planned. The original design assumptions are not reported, so these illustrations should not be confused with the rationale for the completed trial. Statistically, the study is a useful example of why design assumptions, endpoint definitions, uncertainty intervals, and analysis transparency all matter when interpreting a randomized clinical comparison.

19. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary outcome groups are small, with 33 participants listed in each group, and the confidence intervals for the group percentages are broad. This limits precision. The registry records total enrollment of 106, but the primary outcome groups account for fewer participants; the accounting for this difference is not reported in the registry record.

The trial was open-label. The registry does not report details about blinded outcome assessment, if any, or procedures intended to address potential assessment bias. The endpoint uses RECIST v1.1 response categories, but the registry summary does not provide the assessment schedule beyond the stated timeframe or additional information on independent review.

The sponsor’s reported primary analyses provide p-values but no sponsor-reported primary effect estimates or confidence intervals in the record. The recalculated unadjusted comparisons are not a replacement for those analyses, and one comparison shows a notable discrepancy. The full analysis plan, stratification details, multiplicity procedures, missing-data handling, and analysis-set details beyond the all-randomized primary population statement are not reported in the registry record.

The registry’s safety data include a separate galunisertib lead-in group and different at-risk counts across safety groups. This means those safety figures should not be collapsed into a simple comparison of the three primary efficacy groups. The record does not provide exposure-adjusted event rates in the safety summary. Finally, the study’s adaptive description is present in its official title, but the adaptation process and its statistical implications are not reported in the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Abemaciclib (LY2835219) Alone or in Combination With Other Agents in Participants With Previously Treated Pancreatic Ductal Adenocarcinoma — NCT02981342.
• Registry record last updated 2019-11-20; status last verified 2019-08-18.
• Results first posted 2019-06-25.
• 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.

23. Record Summary

This completed randomized, open-label phase 2 study evaluated abemaciclib alone and with LY3023414 against gemcitabine or capecitabine in previously treated metastatic pancreatic ductal adenocarcinoma. The primary endpoint was disease control rate through measured progression or start of new anticancer therapy, up to 6 months. The registry lists 106 enrolled participants and 33 participants in each of the three primary outcome groups.

The posted DCR percentages were 15.2% for 200mg abemaciclib, 12.1% for the combination, and 36.4% for standard of care. Confidence intervals are broad. Sponsor-reported Cochran-Mantel-Haenszel p-values differ from an unadjusted recalculation for one comparison; the sponsor’s own analysis takes precedence where they disagree. Secondary outcomes include response, clinical benefit, pharmacokinetics, and survival. Details of multiplicity, missing data, and the full statistical plan are not reported in the registry record.