Isotretinoin With or Without Dinutuximab, Aldesleukin, and Sargramostim Following Stem Cell Transplant in Treating Patients With Neuroblastoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Isotretinoin With or Without Dinutuximab, Aldesleukin, and Sargramostim Following Stem Cell Transplant in Treating Patients With Neuroblastoma (NCT00026312), a randomized phase 3 study. The primary outcome is Event-Free Survival (EFS).

2. Clinical Question

3. Trial Design

This is a phase III, randomized, parallel-group trial that the registry describes as partially randomized. Masking is listed as none. The trial studied treatment following myeloablative therapy and autologous stem cell rescue in patients with neuroblastoma. Its title and summary frame the central comparison as isotretinoin with dinutuximab, aldesleukin, and sargramostim versus isotretinoin alone.

The registry reports enrollment of 1449. That figure describes the overall trial record and should not be confused with the numbers in the posted primary outcome comparison: 112 participants in Regimen A and 114 in Regimen B. The primary outcome population is eligible, randomized patients. The record also describes outcomes for patients from the non-randomized portion, but those outcomes are separate from the randomized comparison considered in the primary result.

The trial began on 2001-10-26, had a primary completion date of 2012-01-12, and is listed as completed. The record names the National Cancer Institute (NCI) as sponsor. It lists isotretinoin as the drug in the comparator regimen and dinutuximab, aldesleukin, and sargramostim as biological interventions in the combination regimen. The available arm descriptions are incomplete; the full protocol schedule and treatment details are not reported in the registry record.

The design information establishes the broad randomized comparison, but does not establish how allocation was generated or concealed, or how the partially randomized component related operationally to the randomized component. Those details are not reported in the registry record. The primary endpoint description identifies a three-year event-free survival comparison, while the registry also lists overall survival, toxicity, number of therapy courses, and outcomes in the non-randomized portion as secondary outcomes.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Eligible, randomized patients.

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 Event-Free Survival (EFS), measured over Three years.

The primary endpoint is event-free survival (EFS), reported as a percentage of participants at three years. In general, event-free survival describes the proportion of participants who remain free of a defined event by a specified time. It is useful in clinical trials when events that mark progression or failure are relevant before death, but the exact event definition used in this trial is not reported in the registry record.

The registry states that the primary comparison is intended to determine whether the combination of retinoic acid and immunotherapy improves EFS compared with retinoic acid only. In the arm descriptions, isotretinoin is the retinoic acid treatment. The primary population is eligible, randomized patients. A fixed-time percentage gives a clinically interpretable snapshot of outcome at the stated time, while a time-to-event analysis can incorporate information about the timing of events and follow-up.

The record posts percentages with 95% confidence intervals for both regimens and also reports a log-rank test. These are related ways of describing EFS, but they are not interchangeable summaries. The event definition, censoring rules, and detailed construction of the posted percentages are not reported in the registry record. That limits the ability to determine why the posted percentages and the sponsor’s log-rank result differ.

7. Results Posted in the Registry

• Regimen B: 63.2% (72/114) versus 48.2% (54/112) on control; risk difference 14.9 percentage points (95% CI 2.1 to 27.8), risk ratio 1.31 (1.03–1.66); unadjusted two-proportion z-test p = 0.024.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports the primary outcome as a percentage of participants with an event-related status summarized at three years, with a 95% confidence interval for each regimen. In general, a confidence interval describes uncertainty in an estimate under the analysis assumptions; it is not a range containing every participant’s outcome. The intervals around the two regimen estimates communicate that the percentages are estimated rather than known exactly.

The percentage summaries can be contrasted using a risk difference. In general, a risk difference expresses the absolute separation between two proportions, while a relative risk expresses their ratio and an odds ratio compares their odds. The posted recalculation gives a risk difference of 0.149, a relative risk of 1.31, and an odds ratio of 1.841. Each measure frames the contrast differently. The risk difference is on an absolute scale; relative risk and odds ratio are relative scales. The posted confidence intervals are 0.021 to 0.278 for the risk difference, 1.033 to 1.661 for the relative risk, and 1.082 to 3.132 for the odds ratio.

These estimates are labelled recalculated because they are simple unadjusted calculations from posted summary statistics. They do not necessarily reproduce a trial’s own time-to-event analysis. The recalculated unadjusted p-value is 0.024, but the registry also reports the sponsor’s log-rank test with p = 0.102. Since the discrepancy flag is true, the sponsor’s own analysis takes precedence. The registry record does not state why the results differ, and the recalculation should not be treated as a correction to the sponsor’s result.

A log-rank test is a general method for comparing time-to-event experience between groups. It uses information about when events occur during follow-up, rather than reducing all follow-up to a single fixed-time comparison. The registry reports a log-rank test statistic of 2.680 and a p-value of 0.102 for the two groups. The record does not report the test’s detailed implementation, the number of events included, the censoring rules, or whether any adjustment was used. Those details are important for reproducing or fully assessing the analysis.

The primary endpoint is EFS, but the registry does not supply the operational definition of an event. For time-to-event outcomes, event definitions and censoring rules determine how participant histories contribute to the analysis. In general, censoring occurs when the event status is not observed through the full period of interest, and time-to-event methods account for such incomplete follow-up under assumptions. The assumptions and handling used in this trial are not reported in the registry record.

The record lists 18 outcome measures and 5 analyses. Secondary outcomes include EFS and overall survival for patients from the non-randomized portion, incidence of toxicities assessed using Common Terminology Criteria for Adverse Events Version 4.0, number of courses of therapy delivered, and overall survival in the randomized comparison. The registry does not give enough information to determine the full set of statistical tests, whether the primary analysis was adjusted for other variables, or how the five analyses correspond to the listed outcomes.

Multiplicity refers, in general, to the increased chance of at least one apparently notable result when many outcomes or comparisons are examined. Whether and how a trial accounts for multiplicity depends on the analysis plan. The registry record does not report a multiplicity adjustment or describe a testing hierarchy. It also does not report interim analyses, stopping rules, missing-data methods, or the trial’s original statistical analysis plan. No claims about these design or analysis features can be made from the posted information.

The comparison is described as randomized, but allocation generation, allocation concealment, and stratification are not reported in the registry record. Masking is listed as none. Lack of masking can matter in general when knowledge of treatment assignment could influence outcome assessment or care, but the registry does not establish whether any such influence occurred. For EFS, the importance of this issue depends in part on how events were defined and assessed, details that are not reported in the registry record.

The reported group counts for the primary outcome are 112 for Regimen A and 114 for Regimen B. These are the denominators associated with the posted arm summaries, not a complete account of participant flow. Numbers screened, excluded, randomized, treated, lost to follow-up, or included in each analysis beyond the posted counts are not reported in the registry record. Without that information, the extent of missing outcome data and any differences between randomized and analyzed populations cannot be assessed.

10. Confidence Intervals and Precision

The EFS estimate for Regimen A is 48.1%, with a 95% confidence interval from 38.4% to 57.8%. For Regimen B, the estimate is 62.9%, with a 95% confidence interval from 54.0% to 71.9%. In general, a confidence interval reflects statistical uncertainty around an estimate under the method’s assumptions. The intervals indicate that the exact underlying EFS percentages are not determined by the point estimates alone.

The recalculated risk difference is 0.149, with a 95% confidence interval from 0.021 to 0.278. Its interval expresses uncertainty in the unadjusted contrast calculated from the posted summary statistics. The relative risk and odds ratio also have intervals: 1.31, with a 95% confidence interval from 1.033 to 1.661, and 1.841, with a 95% confidence interval from 1.082 to 3.132, respectively. These relative measures describe contrasts on different scales and should not be read as interchangeable with the risk difference.

Confidence intervals do not account for every possible source of uncertainty. They do not resolve uncertainty about the event definition, missing data, analysis population, or the discrepancy between the recalculation and the sponsor’s log-rank result. The registry does not provide the information needed to assess those sources in detail. Because the sponsor’s own analysis takes precedence when the discrepancy flag is true, the recalculated interval should be understood as a descriptive unadjusted calculation rather than the definitive trial inference.

The confidence intervals for the arm percentages overlap. That visual feature is not a replacement for the trial’s formal comparison, especially because the reported sponsor analysis is a log-rank test and the recalculation uses posted fixed-time summary statistics. Precision should therefore be judged in relation to the specific estimand and method being considered, while keeping the unexplained difference between analyses in view.

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 sample-size figures are planning illustrations based on the observed effect, not the original trial’s design calculations. The listed 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 set to the stated fraction of the observed effect. The trial’s own original sample-size assumptions and justification are not reported in the registry record.

For the binary outcome planning figures, the counts are participants per arm, not total participants across both arms and not total events. The 80% power column describes the number of participants per arm that the planning calculation estimates would be needed for an 80% chance of detecting the selected effect under those assumptions, if that effect and assumptions held. The 90% power column gives the corresponding per-arm planning number for 90% power. These are planning concepts, not guarantees about what a future trial would observe.

The rows use 100%, 75%, and 50% of the observed effect as planning targets. As the target effect becomes smaller, the per-arm participant requirement increases in the listed illustrations: for 80% power, the figures are 173, 310, and 701; for 90% power, they are 231, 414, and 938. This pattern reflects a general principle: smaller differences are harder to distinguish from random variation, so more participants are needed under otherwise similar assumptions.

These values are not observed power. They are hypothetical planning illustrations derived from the observed effect, and they do not describe the trial’s actual enrollment decision or establish the probability that its result is correct.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.482; treatment proportion p₁ = 0.632; average p̄ = 0.557.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7025 = 1.3769.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6945 = 0.5845.
4. n = (1.3769 + 0.5845)² / (0.1494)² = 172.3, rounded up to 173 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 <- 0.482142857142857                          # control proportion
p1 <- 0.631578947368421                          # 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(26312)
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 = 173 (80% power), 231 (90% power)
 75% of observed effect: n per arm = 310 (80% power), 414 (90% power)
 50% of observed effect: n per arm = 701 (80% power), 938 (90% power)
power.prop.test: 172.3 
Simulated power at n = 173 per arm: 0.795 (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.

13. Analyses Reported by the Sponsor

The registry’s sponsor-reported comparison is a log-rank test, with a test statistic of 2.680 and p = 0.102. The recalculation from posted summary statistics gives an unadjusted p-value of 0.024 and a risk difference of 0.149, with a 95% confidence interval from 0.021 to 0.278. These results do not agree. The registry discrepancy flag is true, so the sponsor’s own analysis takes precedence.

The recalculation is based on the posted arm-level percentages and counts and is not the same procedure as the log-rank analysis. In general, a fixed-time contrast and a time-to-event test can summarize different aspects of outcome data. The registry does not report enough detail to establish what specifically produced the difference in this trial. Potential explanations should not be treated as findings about the record.

The appropriate reading is therefore to report both results with their distinct origins and give priority to the sponsor’s log-rank analysis when describing the trial’s own inferential result. The recalculated estimate remains useful as an unadjusted description of the posted summary statistics, but it should not replace, override, or be presented as a correction to the sponsor’s analysis. The registry record does not explain the discrepancy or provide the detailed analysis plan needed to reconcile the results.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists 18 outcome measures and 5 analyses. Alongside the primary EFS outcome, it names secondary outcomes that include overall survival, toxicity, number of courses of therapy delivered, and EFS and overall survival for the non-randomized portion of the trial. The presence of multiple outcomes and analyses makes it important to know which comparisons were primary, which were secondary, and whether any adjustment or hierarchy was planned.

In general, multiplicity arises when several hypotheses are tested: the chance of finding at least one apparently notable result can increase as the number of tests grows. Trials may address this through prespecified testing sequences or statistical adjustments. Whether either approach was used, and how it applied across the outcomes and analyses, is not reported in the registry record. The posted p-values should not be assumed to be adjusted for multiplicity.

The registry does not report the protocol’s statistical analysis plan, prespecified interim analyses, stopping rules, missing-data procedures, or the primary analysis population beyond describing eligible, randomized patients. It also does not report participant flow in enough detail to assess whether everyone randomized was included in the posted analyses, how incomplete follow-up was handled, or whether exclusions occurred after randomization. These omissions limit the ability to assess risks associated with selective analysis or missing outcomes.

The overall trial is described as partially randomized, and the registry explicitly lists separate outcome measures for the non-randomized portion. The available information does not describe how patients entered the randomized versus non-randomized components, or whether the same analysis rules applied to both. Results from the non-randomized portion should not be merged with the randomized comparison without a clear basis, which is not reported in the registry record.

Masking is listed as none. In general, open-label treatment can affect outcome assessment or other trial processes, but the registry does not report how EFS events were assessed or whether any independent review was used. The event definition and adjudication procedures are also not reported in the registry record. These gaps prevent a full assessment of the analysis population, missing data, and multiplicity handling.

15. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hypotension (18.8%), capillary leak syndrome (11.6%), anaphylaxis (11.5%), allergic reaction (10.6%), fever (9.1%).

The registry posts serious adverse event figures of 10 affected participants among 112 at risk for Regimen A, and 785 among 1328 at risk for Regimen B. For other adverse events, it posts 92 affected among 112 at risk for Regimen A, and 1244 among 1328 at risk for Regimen B. These are registry counts with their stated denominators; they should not be interpreted as comparable treatment effects without additional information about how the groups and observation periods were defined.

The adverse-event denominators differ substantially from the primary EFS comparison counts of 112 and 114, particularly for Regimen B. The registry record does not explain the relationship between these safety populations and the randomized outcome population. It also does not report exposure duration or participant-level details needed to interpret the counts as rates over a common period. The death figures are not reported in the registry record.

The registry lists hypotension at 18.8%, capillary leak syndrome at 11.6%, anaphylaxis at 11.5%, allergic reaction at 10.6%, fever at 9.1%, and infections and infestations - Other specify at 8.5% among the top serious adverse event terms. The record does not identify these term percentages by regimen in the listed results, so they should not be attributed to one arm. Safety figures describe recorded events; they do not, by themselves, establish causation or comparative risk.

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

The posted results describe three-year EFS percentages of 48.1% for Regimen A and 62.9% for Regimen B, each with a 95% confidence interval. An unadjusted recalculation from the posted summaries gives a risk difference of 0.149, with an interval from 0.021 to 0.278. These values describe the summaries and the recalculation, but the recalculation is not the sponsor’s time-to-event analysis.

The sponsor-reported log-rank test has p = 0.102. Because the registry identifies a discrepancy between the reported analysis and recalculated results, the sponsor’s own analysis takes precedence. The record does not explain the reason for the difference. The two sets of results should not be combined into a single conclusion or selectively quoted to support a preferred interpretation.

These results do not establish that one regimen works or does not work. They do not settle how well the findings apply outside the eligible randomized population, or what the results mean for an individual patient. The registry record does not provide enough detail on the event definition, missing-data handling, or the full analysis plan to resolve those questions. The overall enrollment of 1449 also should not be mistaken for the primary comparison’s posted group counts of 112 and 114.

The secondary outcome list and safety counts add context, but the available record does not provide a complete account of the statistical methods, participant flow, exposure, or adverse-event group definitions. Interpretation is therefore most defensible when it stays close to the identified population, endpoint, time point, and analysis method, and when the unreported details and discrepancy are made explicit.

17. Why This Trial Matters Statistically

This trial is statistically instructive because it places a fixed-time endpoint summary beside a time-to-event test. The EFS percentages and their confidence intervals are readily interpretable as estimates at three years, while the log-rank test compares event-time experience across follow-up. These summaries answer related questions but rely on different statistical procedures. A reader should not assume that a recalculation from percentages reproduces the trial’s own analysis.

The difference between the sponsor’s p = 0.102 and the recalculated unadjusted p-value of 0.024 shows why the provenance of a statistic matters. The recalculation is based on posted summary statistics; the registry identifies a discrepancy, making the sponsor’s analysis the priority for reporting the trial’s inferential result. Without the analysis plan and fuller outcome data, the reason for the difference cannot be determined from the registry record.

The record also illustrates the importance of keeping denominators and populations distinct. Overall enrollment is 1449, the posted randomized EFS groups are 112 and 114, and the safety denominators include 112 and 1328. These figures refer to different reported contexts, and their relationship is not fully explained in the registry record. Treating them as interchangeable would obscure the scope of each result.

Finally, the registry identifies a partially randomized design, multiple outcome measures, and safety outcomes, yet leaves important analysis details unreported. This makes the separation of randomized and non-randomized results, attention to multiplicity, and transparent handling of missing information central to sound interpretation. The statistical lesson is not to infer more than the record can support.

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

The registry record does not state the operational definition of EFS, the censoring rules, or how the three-year percentages were estimated. It reports a log-rank test but does not provide the detailed analysis plan or enough information to reproduce that result. The reason the sponsor’s result differs from the unadjusted recalculation is not reported in the registry record.

The record describes the trial as partially randomized, but does not detail how participants entered randomized and non-randomized portions or how their analyses were conducted. Although eligible, randomized patients are named as the primary population, participant flow from overall enrollment of 1449 to the posted EFS groups of 112 and 114 is not explained. This limits assessment of selection, exclusions, and the population represented by the posted estimates.

Masking is listed as none, but event assessment procedures and any independent review are not reported. The registry also does not report missing-data methods, interim analyses, stopping rules, multiplicity adjustments, allocation concealment, or the original sample-size justification. The planning illustrations are based on the observed effect and do not reveal the trial’s actual design assumptions.

The safety results use denominators that differ from those in the primary EFS comparison, and the registry does not explain the safety population or exposure time. The top serious adverse event term percentages are not linked to a specific regimen in the posted list. Death counts are not reported in the registry record. These limitations restrict comparisons of safety counts and interpretation of their relationship to treatment exposure.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Isotretinoin With or Without Dinutuximab, Aldesleukin, and Sargramostim Following Stem Cell Transplant in Treating Patients With Neuroblastoma — NCT00026312.
• Registry record last updated 2026-03-05; status last verified 2026-01.
• Results first posted 2017-03-29.
• 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

NCT00026312 is a completed phase III, partially randomized, open-label, parallel-group trial sponsored by the National Cancer Institute (NCI). It examined isotretinoin alone versus isotretinoin with dinutuximab, aldesleukin, and sargramostim following stem cell transplant in neuroblastoma. The registry reports overall enrollment of 1449 and identifies eligible, randomized patients as the primary outcome population.

The primary endpoint is three-year EFS. Posted percentages are 48.1% for Regimen A and 62.9% for Regimen B; the sponsor-reported log-rank test has p = 0.102. A recalculation from the posted summaries differs, and the registry record does not explain the discrepancy. The sponsor’s own analysis takes precedence. Important details about the endpoint definition, analysis plan, missing data, multiplicity, participant flow, and safety denominators are not reported in the registry record.