Testing the Use of the Immunotherapy Drugs Ipilimumab and Nivolumab Plus Radiation Therapy Compared to the Usual Treatment (Temozolomide and Radiation Therapy) for Newly Diagnosed MGMT Unmethylated Glioblastoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Testing the Use of the Immunotherapy Drugs Ipilimumab and Nivolumab Plus Radiation Therapy Compared to the Usual Treatment (Temozolomide and Radiation Therapy) for Newly Diagnosed MGMT Unmethylated Glioblastoma (NCT04396860), a randomized phase 2, phase 3 study. The primary outcome is Progression-free Survival (PFS) (Phase II).

2. Clinical Question

3. Trial Design

This was a randomized, open-label, parallel-group phase II/III study sponsored by the National Cancer Institute. The registry record lists enrollment of 159 and a completed status. The recorded start date is 2020-09-01, and the primary completion date is 2023-01-13. Masking was none, so the study was not described as blinded.

Participants were assigned to radiation therapy with temozolomide or radiation therapy with ipilimumab and nivolumab. The posted phase II primary outcome was progression-free survival. For the standard-treatment arm, radiation was scheduled 5 days per week for 30 fractions over 6 weeks, alongside daily temozolomide for 6 weeks. The other arm received radiation on the same schedule, ipilimumab every 4 weeks for 4 doses, and nivolumab every 2 weeks until disease progression. The registry record also lists imaging, quality-of-life assessment, questionnaires, and other interventions.

The primary outcome is specifically labeled as applying to phase II. The record also lists progression-free survival for the entire cohort as a secondary outcome, along with overall survival at 2 years, adverse events, symptom assessment, neurocognitive function, and patient-reported outcomes. Numerical results for those secondary outcomes are not reported in the registry record in the information summarized for this comparison. The trial’s original design assumptions, including its planned sample-size justification and any interim-analysis plan, are not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: Screening required central pathology review confirmation of glioblastoma histology and unmethylated MGMT promotor status in order to proceed to randomization. Of 374 patients registered, 159 patients were randomized.

Overall Study

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

Population: Randomized participants

6. Primary Endpoint

The registry lists the primary outcome as Progression-free Survival (PFS) (Phase II), measured over From randomization to date of progression, death, or last known follow-up, whichever occurs first. Maximum follow-up time was 19.3 months..

Progression-free survival (PFS) measures the time from randomization until disease progression or death. In this trial, follow-up could also end at the last progression assessment without documented progression, in which case the observation was censored. The posted timeframe specifies progression, death, or last known follow-up, whichever occurred first, and gives a maximum follow-up time of 19.3 months.

Progression was confirmed by central review and defined using an adaptation of Response Assessment in Neuro-Oncology criteria. The adaptation made the first MRI after radiotherapy the baseline MRI for assessing radiographic disease. The registry record says PFS rates were estimated using the Kaplan-Meier method.

General explanation: PFS is a time-to-event endpoint because it concerns the time until an event occurs, rather than simply whether an event occurred by a single visit. It is often used to describe how long participants remain alive without documented progression during follow-up. Censoring allows an analysis to retain information from participants whose event status is not observed through the end of follow-up, provided the analysis method and censoring assumptions are appropriate. The registry record does not report further details about those assumptions.

7. Results Posted in the Registry

• Arm I (Radiation Therapy, Temozolomide) vs Arm II (Radiation Therapy, Ipilimumab, Nivolumab): hazard ratio 1.47 (95% CI 0.98–2.21), reported p = 0.960 (as reported in the registry).

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome analysis uses methods for time-to-event data. The registry record identifies the Kaplan-Meier method for estimating PFS rates and the log-rank method for the reported hazard-ratio comparison. The outcome definition includes censoring at the last progression assessment when progression has not been documented.

What does Kaplan-Meier estimation describe?

General explanation: Kaplan-Meier estimation uses the observed event times and the follow-up information available before censoring to estimate the probability of remaining event-free over time. It can accommodate participants whose follow-up ends before an event is observed, rather than treating every censored participant as if an event had occurred. Its interpretation depends on the event definition, the follow-up process, and assumptions about censoring. The registry record describes the endpoint’s censoring rule but does not report diagnostics or additional assumptions.

The median PFS is the estimated time at which the Kaplan-Meier curve reaches the halfway point in event-free probability. It is a summary of the curve, not a statement that all participants experienced progression or death by that time. The confidence intervals around the arm-specific medians communicate uncertainty in those estimates. They do not directly test whether the two medians differ.

What does the log-rank analysis contribute?

General explanation: A log-rank test compares event-time patterns between groups across follow-up, taking account of the timing of events and censoring. A hazard ratio expresses a relative event rate for a comparison, conditional on the coding and model convention used. The registry record reports a hazard ratio of 1.47 and a p-value of 0.96 for the log-rank method, but does not report the reference coding or further model details. It is therefore important not to infer the hazard ratio’s direction from the arm ordering alone.

A p-value describes how compatible the observed test result is with a specified null model, given the test’s assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The posted p-value is 0.96. The registry record does not report additional details about the log-rank test’s specification, such as stratification or handling of any planned interim analyses.

Why do confidence intervals matter?

General explanation: A confidence interval gives a range of estimates compatible with the data under the analysis procedure and its assumptions. A wider interval indicates less precision than a narrower interval on the same scale. For the hazard ratio, the posted interval is 0.98 to 2.21; it includes the null value and also allows for a range of departures from it. The interval is not a guarantee that the true effect lies within its endpoints, and it does not resolve questions about model assumptions or bias.

What is over-dispersion?

General explanation: Over-dispersion occurs when observed variation is greater than a statistical model expects. It matters for analyses that model counts or proportions because standard errors can otherwise be too small and uncertainty understated. The primary outcome is time-to-event, and the registry record does not report an over-dispersion analysis. No specific over-dispersion conclusion can therefore be drawn from this endpoint description.

The registry record describes central review of progression and an adaptation of the RANO criteria, including use of the first MRI after radiotherapy as the baseline for radiographic disease. Such endpoint rules define what counts as an event and can affect event timing. The record does not report the number of events, detailed adjudication procedures, or additional sensitivity analyses, so those aspects cannot be assessed from the posted primary result.

10. Confidence Intervals and Precision

Precision is limited by the width of the reported uncertainty intervals. For the radiation therapy and temozolomide arm, median PFS is 8.5 months, with a 95% confidence interval from 7.1 to 10.4 months. For the radiation therapy, ipilimumab, and nivolumab arm, the median is 7.7 months, with a 95% confidence interval from 6.5 to 8.5 months. These intervals describe uncertainty in each median separately; overlapping intervals are not, by themselves, a formal test of the between-arm contrast.

The reported hazard ratio is 1.47, with a 95% confidence interval from 0.98 to 2.21. The interval includes the null value for a hazard ratio and extends to values representing a larger relative difference. This range leaves substantial uncertainty about the comparative effect. The point estimate should not be read in isolation, particularly because the reference coding is not reported in the registry record.

General explanation: A confidence interval is a measure of statistical uncertainty under the method used to calculate it. It does not account automatically for every possible source of uncertainty, such as outcome measurement, missing follow-up, or differences in treatment delivery. Nor does a 95% interval mean there is a 95% probability that the fixed true effect lies inside the interval. The registry record does not provide enough detail to assess all assumptions behind these intervals.

The maximum follow-up time reported for the primary outcome is 19.3 months. The registry record does not give the number of events or a fuller description of follow-up distribution. Precision and maturity of the time-to-event estimates therefore cannot be judged from the maximum follow-up alone.

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 expressed as total events, not participants per arm. The illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its starting point, then considers planning for 100%, 75%, or 50% of that observed effect. For 100% of the observed effect, the figures are 212 total events for 80% power and 284 total events for 90% power. For 75% of the observed effect, they are 377 and 504 total events, respectively. For 50% of the observed effect, they are 847 and 1,133 total events, respectively.

General explanation: The 80% power column describes the event total that the illustration estimates would be needed to have an 80% chance of detecting the assumed effect under its planning assumptions. The 90% power column uses a higher chance of detection and therefore requires a larger event total for each effect scenario. Smaller effects are harder to distinguish from random variation, so the illustration calls for more events when planning for a smaller fraction of the observed effect.

These are planning illustrations based on the observed effect, not the trial’s original design assumptions. The trial’s own sample-size justification is not reported in the registry record. This is not observed power: it does not describe the probability that the completed trial “had power” after its results were known, and it is not an estimate of the probability that the finding is true.

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

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

Worked calculation for this trial

1. Hazard ratio HR = 1.470, so ln(HR) = 0.3853.
2. d = 4 × (1.96 + 0.842)² / [ln(1.470)]² = 211.5, rounded up to 212 events in total for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
hr <- 1.47                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(4396860)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 212 (80% power), 284 (90% power)
 75% of observed effect: events = 377 (80% power), 504 (90% power)
 50% of observed effect: events = 847 (80% power), 1,133 (90% power)
Simulated power with 212 events: 0.788 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

13. Analyses Reported by the Sponsor

The posted comparison gives the sponsor-reported log-rank hazard ratio of 1.47, with a 95% confidence interval from 0.98 to 2.21 and a p-value of 0.96. The statistical presentation uses the reported hazard ratio rather than a separate recalculation from arm-level summary statistics. The registry record does not flag a discrepancy between a recalculated result and the sponsor’s analysis.

There is therefore no alternative recalculated hazard ratio to set against the posted analysis. The reported result remains the relevant comparison in the registry record, while its interpretation remains constrained by the interval width, the p-value, and the absence of reported reference coding. The registry record does not provide additional analysis details that would resolve the unusual combination of the reported hazard ratio and p-value, so it is best to state both as posted rather than try to reconcile them by assumption.

General explanation: A recalculation from summary statistics can be useful as a check, but it may not reproduce an analysis based on participant-level time-to-event data, censoring, or other model choices. The sponsor’s own analysis is the primary reported result when no discrepancy is identified; a summary-statistic calculation should not be presented as though it were the trial’s prespecified analysis.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists several secondary outcomes, including PFS for the entire cohort, overall survival at 2 years, adverse events, symptom change, neurocognitive function, and selected patient-reported outcomes. It reports six secondary outcomes in the record, but numerical results for those secondary outcomes are not reported in the registry record in the information summarized for the primary comparison. The primary PFS result is labeled phase II.

Multiplicity arises when a study makes several statistical comparisons or evaluates multiple outcomes. General explanation: As the number of tests increases, the chance of at least one apparently unusual result can increase if each test is interpreted separately. A prespecified approach may control that risk, but the registry record does not report whether multiplicity adjustment was planned or used. It also does not report a hierarchy of endpoints or an allocation of statistical significance across analyses.

The record identifies randomized participants as the population for the primary outcome. It does not report whether the analysis followed an intention-to-treat principle, whether participants were excluded after randomization, or how protocol deviations were handled. The enrollment value is 159, and the arm-level primary outcome entries list 80 and 79; the record does not explain any difference between enrollment and the analysis population.

Missing-data methods beyond the stated censoring rule are not reported in the registry record. The record does not describe missing progression assessments, missing patient-reported outcomes, withdrawal handling, or sensitivity analyses for missingness. Those omissions limit assessment of whether follow-up and outcome availability could have influenced the estimates. The registry record also does not report interim analyses, stopping rules, or the original statistical analysis plan.

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: seizure (12.8%), thromboembolic event (7.7%), encephalopathy (6.4%), colitis (5.1%), diarrhea (5.1%).

The posted serious-adverse-event figures are 29 affected among 74 at risk in the radiation therapy and temozolomide arm, and 45 affected among 78 at risk in the radiation therapy, ipilimumab, and nivolumab arm. For other adverse events, the record gives 70 affected among 74 at risk in the first arm and 70 among 78 at risk in the second. Death counts are 41 and 37, respectively. These counts should be read with their stated at-risk denominators and should not be treated as comparative rates calculated elsewhere.

The listed serious-event terms include seizure at 12.8%, thromboembolic event at 7.7%, encephalopathy at 6.4%, colitis at 5.1%, diarrhea at 5.1%, and infections and infestations - Other at 5.1%. The registry record does not report the denominators for these term-specific percentages or clarify whether participants could contribute to more than one term. It also does not give timing, duration, attribution, or a detailed comparison of event severity by arm in the information.

General explanation: Adverse-event counts describe events recorded during the study and do not, by themselves, establish that treatment caused an event. Differences in exposure time, follow-up, and event definitions can affect comparisons. The figures should therefore be interpreted descriptively rather than as a causal safety comparison.

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

The registry record reports median PFS of 8.5 months for radiation therapy with temozolomide and 7.7 months for radiation therapy with ipilimumab and nivolumab. It also reports a hazard ratio of 1.47, a 95% confidence interval from 0.98 to 2.21, and a p-value of 0.96. Together, these estimates describe the posted phase II analysis, but they do not establish that one regimen is better, worse, or equivalent to the other.

The confidence interval is wide enough to leave uncertainty around the comparative effect, and the record does not give the reference coding needed to assign a direction to the hazard ratio. A non-significant result does not prove that the treatments have the same effect. Equivalence or non-inferiority would require an appropriate design and prespecified margin; those details are not reported in the registry record.

The endpoint is progression-free survival, not overall survival. Although overall survival at 2 years is listed as a secondary outcome, its numerical result is not reported in the registry record in the information summarized for this comparison. The maximum follow-up for the primary outcome was 19.3 months, so the posted primary result should not be used to make claims about outcomes beyond the information recorded.

The trial addresses a defined population with newly diagnosed MGMT-unmethylated glioblastoma and compares two multi-part treatment strategies. The observed findings apply to the trial population and the endpoint definition; they do not automatically answer questions about other populations, treatment schedules, or outcomes. General explanation: Statistical estimates are not individual predictions. They summarize group-level data and remain subject to uncertainty, measurement choices, and study limitations.

These results are a statistical description of the posted registry record, not medical advice. The available information does not support a claim that either treatment works or does not work.

17. Why This Trial Matters Statistically

This trial is statistically notable because it compares two treatment strategies in a randomized phase II/III design while reporting a phase II time-to-event endpoint. Random allocation can support a fair comparison by reducing systematic differences between groups at baseline, but randomization alone does not resolve uncertainty from limited event information, incomplete follow-up, or outcome measurement.

The endpoint requires careful interpretation. PFS depends on how progression is defined and assessed, and the registry record describes central review and an adapted RANO-based definition. It also specifies that the first MRI after radiotherapy became the baseline for radiographic disease. Such rules make the event definition more explicit, while also making clear that endpoint interpretation depends on the assessment process.

The contrast between the arm medians and the reported hazard ratio demonstrates why several summaries matter. Medians describe a point on each estimated event-time curve; the hazard ratio summarizes a relative event rate under a comparison convention. The confidence interval and p-value add information about uncertainty, but do not replace clear coding and analysis details. Here, those details are not fully reported, so the safest reading is descriptive and restrained.

Finally, the trial record lists multiple secondary outcomes and a broader phase II/III framework, but the available numerical result is for the phase II PFS outcome. Distinguishing the posted result from the full set of trial questions is important: one outcome cannot stand in for unreported results on survival, adverse events, symptoms, or neurocognitive function.

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 reported primary result is specific to phase II, whereas the trial is described as phase II/III. The registry record also lists PFS for the entire cohort as a secondary outcome, but its numerical result is not reported in the registry record in the information summarized for this comparison. This limits conclusions about the whole-trial outcome from the phase II result alone.

The maximum follow-up time for the primary outcome was 19.3 months. The record does not provide the number of events, the distribution of follow-up, or further detail on progression assessments. It is therefore not possible to judge the maturity of the event data from the maximum follow-up alone.

The hazard ratio’s reference coding is not reported, making its direction unclear from the registry summary. The reported hazard ratio, confidence interval, and p-value are stated together, but the registry record gives no further explanation of their relationship or detailed model specification. The confidence interval is also broad, leaving substantial uncertainty about the magnitude of the comparative effect.

The record does not report the original sample-size justification, interim analyses, missing-data procedures beyond censoring, multiplicity adjustment, or full analysis-population rules. It also does not provide numerical results for the listed secondary outcomes in the information. The study was open-label, and no masking was used; the registry record does not describe any steps to limit potential effects of that design on assessment or reporting.

Safety figures are descriptive and lack details such as term-specific denominators, event timing, and attribution in the information summarized for this comparison. These limitations constrain interpretation of both the comparative efficacy estimate and the safety counts.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Testing the Use of the Immunotherapy Drugs Ipilimumab and Nivolumab Plus Radiation Therapy Compared to the Usual Treatment (Temozolomide and Radiation Therapy) for Newly Diagnosed MGMT Unmethylated Glioblastoma — NCT04396860.
• Registry record last updated 2026-04-22; status last verified 2026-04.
• Results first posted 2024-05-01.
• 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

NCT04396860 is a completed randomized, open-label phase II/III parallel-group trial sponsored by the National Cancer Institute. It enrolled 159 and compared radiation therapy with temozolomide against radiation therapy with ipilimumab and nivolumab in newly diagnosed MGMT-unmethylated glioblastoma.

The posted phase II primary outcome was PFS. Median PFS was 8.5 months in the temozolomide arm and 7.7 months in the immunotherapy arm. The reported hazard ratio was 1.47, with a 95% confidence interval from 0.98 to 2.21 and p-value 0.96. The confidence interval indicates limited precision, and the registry record does not report reference coding for the hazard ratio.

The record lists secondary outcomes covering PFS, survival, adverse events, symptoms, neurocognitive function, and patient-reported outcomes. Their numerical results are not reported in the registry record in the information summarized for this comparison.