Phase III Trial of Anaplastic Glioma Without 1p/19q Loss of Heterozygosity (LOH): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase III Trial of Anaplastic Glioma Without 1p/19q Loss of Heterozygosity (LOH) (NCT00626990), a randomized phase 3 study. The primary outcome is Overall Survival as Measured From the Day of Randomization.

2. Clinical Question

The trial rationale states that it was not known whether giving temozolomide during and/or after radiotherapy was more effective than radiotherapy alone in treating anaplastic glioma. The reported analyses address the timing questions by combining arms according to whether each component was present. This framing matters: the results are not four independent pairwise comparisons, and the presence-versus-absence comparisons should not be mistaken for direct estimates of every possible regimen-versus-regimen contrast.

3. Trial Design

This was a randomized phase III parallel-group trial with no masking. The registry records enrollment of 751, a start date of 2007-12, primary completion on 2018-09-05, and status as active, not recruiting. The sponsor is the European Organisation for Research and Treatment of Cancer - EORTC. The registry record does not report the trial's original sample-size justification or its original statistical design assumptions.

Four randomized arms varied the timing of temozolomide around radiotherapy: radiotherapy alone; radiotherapy with concurrent chemotherapy; radiotherapy followed by adjuvant chemotherapy; and radiotherapy with concurrent chemotherapy plus adjuvant chemotherapy. The primary outcome analysis recombined those arms in two ways. One comparison contrasted absence with presence of concomitant temozolomide; the other contrasted absence with presence of adjuvant temozolomide. Arm-level median survival summaries are also reported for those combined groups.

The primary outcome was overall survival from randomization to death due to any cause. The registry says patients not reported dead or lost to follow-up were censored at their last follow-up examination. The recorded timeframe states that time to death was followed up to 10.9 years after enrollment. Secondary outcomes listed include progression-free survival, quality of life, and neurological deterioration-free survival. Numerical results are posted for progression-free survival; detailed numerical results for the other listed secondary outcomes are not reported in the registry record.

The four original arms make it possible to examine the two treatment components separately, but the combined comparisons do not by themselves establish whether the effects of concurrent and adjuvant treatment interact. An interaction analysis or a more detailed account of the model's handling of both factors is not reported in the registry record. The trial's allocation concealment, exact randomization procedure, and analysis-population definitions are also 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: The patients were first registered to the trial by authorized sites. For all patients, tumor and blood samples had to be sent for histology review, 1p/19q analysis and O6-Methylguanine-DNA Methyltransferase (MGMT) assay. If inclusion was based on central pathology and 1p/19q diagnosis the patient could be randomized into the trial once found eligible at central assessment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: This trial studied 2 questions 1) effect of concomitant Temozolomide (TMZ) 2) effect of adjuvant TMZ. The 4 randomized arms were combined. Absence of concomitant TMZ including Arm Radiotherapy (RT) alone \& Arm RT followed by adjuvant TMZ. Presence of concomitant TMZ including Arm TMZ/RT \& Arm TMZ/RT followed by adjuvant TMZ. Absence of adjuvant TMZ including Arm RT alone \& Arm TMZ/RT. Presence of adjuvant TMZ including Arm RT followed by adjuvant TMZ \& Arm TMZ/RT followed by adjuvant TMZ

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 Overall Survival as Measured From the Day of Randomization, measured over from date from enrollment till the date of death (time till death is up to 10.9 years after patient enrollment in the study).

Overall survival is the time from randomization until death from any cause. The registry defines the event broadly, rather than restricting it to death attributed to the tumor or to treatment. Patients who had not been reported dead or lost to follow-up were censored at the date of their last follow-up examination. Censoring means that the exact later survival time is not known for those individuals, but their observed follow-up contributes information up to the recorded last contact.

The endpoint is reported as median survival in months, with a 95% confidence interval for each combined treatment group. A median is the time at which the estimated survival distribution reaches its midpoint; it is a summary of the survival curve, not a statement that every individual survives for that duration. The registry also reports Cox-model hazard ratios for the presence-versus-absence comparisons. A hazard ratio summarizes the relative event rate over follow-up under the model; it is not a ratio of median survival times and should not be read as an individual patient's probability of survival.

Overall survival is commonly used in clinical trials because death is a clear event and the endpoint captures survival regardless of the cause of death. General statistical context: time-to-event methods can use information from participants with different follow-up durations while accounting for censoring, provided the method's assumptions are suitable. The registry record does not report detailed checks of those assumptions or the exact censoring rules beyond the description above.

7. Results Posted in the Registry

• Absence of Concomitant Temozolomide (TMZ) vs Presence of Concomitant Temozolomide (TMZ): hazard ratio 0.97 (95% CI 0.73–1.28), reported p = 0.760 (as reported in the registry).
• Absence of Adjuvant Temozolomide (TMZ) vs Presence of Adjuvant Temozolomide (TMZ): hazard ratio 0.64 (95% CI 0.52–0.79), reported p < 0.001 (as reported in the registry).

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Progression-free Survival

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Regression, Cox); it was not recalculated.
• Result:
  • Absence of Concomitant TMZ vs Presence of Concomitant TMZ: hazard ratio 0.86 (95% CI 0.72–1.03), reported p = 0.110 (as reported in the registry).
  • Absence of Adjuvant TMZ vs Presence of Adjuvant TMZ: hazard ratio 0.59 (95% CI 0.49–0.70), reported p < 0.001 (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 Absence of Concomitant TMZ vs Presence of Concomitant TMZ and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Absence of Adjuvant TMZ vs Presence of Adjuvant TMZ than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcome information gives a broad pattern that is consistent with the primary comparison: the adjuvant treatment contrast has stronger statistical evidence than the concomitant contrast, while the concomitant contrast is less conclusive. This provides supporting context for the primary overall-survival interpretation, but it does not make the secondary outcomes equivalent to the primary endpoint. Progression-free survival captures a different event definition from overall survival, and outcomes based on quality of life or neurological deterioration would address other aspects of patients' experience.

The registry lists progression-free survival, quality of life, and neurological deterioration-free survival as secondary outcomes. Numerical results are posted for progression-free survival, while detailed numerical results for quality of life and neurological deterioration-free survival are not reported in the registry record. Thus, the available secondary evidence cannot be treated as a complete picture across all outcomes listed for the trial.

General statistical context: secondary outcomes can help assess whether a primary result is accompanied by effects on related outcomes, but testing many outcomes or contrasts increases the chance of at least one apparently notable result arising by chance. Unadjusted secondary p-values do not account for that multiplicity. Their interpretation depends on whether endpoints and analyses were prespecified, how many tests were performed, and what hierarchy or adjustment strategy was planned. Those details are not reported in the registry record.

Secondary results can also differ from the primary result because they measure distinct events, have different event frequencies, or are affected by different patterns of follow-up and missingness. They should therefore be read as complementary evidence, not as a basis to override the primary overall-survival analysis. The registry does not report a formal combined assessment across the primary and secondary outcomes or a multiplicity-adjusted conclusion.

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

• Progression-free Survival
• Quality of Life of the Patient
• Neurological Deterioration Free Survival

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is time-to-event data: each participant contributes a time from randomization until death or, if death was not observed, until the last recorded follow-up used for censoring. General statistical context: censoring allows an analysis to retain partial follow-up rather than treating an individual without an observed event as if the event never occurred. This approach depends on assumptions about the relationship between censoring and the event process. How those assumptions were assessed is not reported in the registry record.

What does a Cox hazard ratio describe?

The registry reports Cox regression hazard ratios for both primary comparisons. General statistical context: a hazard is the instantaneous event rate among people still at risk at a given point in follow-up. A hazard ratio compares those rates between groups over time, subject to the model's assumptions. A value below 1 indicates a lower estimated hazard in the group in the numerator relative to the comparison group, while a value above 1 indicates a higher estimated hazard. The ordering of groups matters, so the reported contrast must be read as absence versus presence of the treatment component, as labeled by the registry.

A hazard ratio is not a direct comparison of median survival durations, and it is not a probability that an individual will survive. It also generally summarizes the relative event rate rather than describing how the absolute survival difference changes at every follow-up time. The proportional-hazards assumption, under which the relative hazard is treated as stable over time, is a general feature commonly associated with the Cox model. Whether that assumption was checked or met in this trial is not reported in the registry record.

Why report medians as well as hazard ratios?

The posted median survival values offer a time-based description of each combined group. The accompanying confidence intervals convey uncertainty in those medians. The Cox hazard ratio instead summarizes a relative contrast using the event-time information over follow-up. Both can help describe the outcome, but each has a different interpretation. A difference between medians should not be calculated from the posted values when the explanatory text is restricted to registry-reported results, and a median is not a replacement for the model-based comparison.

The registry's primary survival intervals for the combined groups are reported as 95% confidence intervals. The sponsor-reported Cox analysis for concomitant temozolomide has a 99.1% confidence interval, while the adjuvant comparison has a 95% interval. Different confidence levels produce intervals with different degrees of coverage and should be kept distinct when reading the results. The reported p-values are 0.76 for the concomitant comparison and < 0.001 for the adjuvant comparison.

What does a confidence interval contribute?

General statistical context: a confidence interval describes the range of estimates reasonably compatible with the observed data and the method's assumptions at its stated confidence level. It communicates precision more directly than a point estimate alone. An interval that includes the null value of 1 for a hazard ratio indicates that values consistent with no relative hazard difference remain within that interval. An interval entirely below 1 indicates estimates favoring a lower hazard for the group in the numerator, while still leaving uncertainty about the exact magnitude.

In the concomitant comparison, the reported interval includes 1 and extends on either side of it. In the adjuvant comparison, the reported interval is below 1. These statements describe the posted estimates and intervals; they do not determine whether an effect is clinically important. Clinical importance requires context about outcomes, tradeoffs, and prespecified thresholds, which is not reported in the registry record.

Why can over-dispersion matter?

General statistical context: over-dispersion occurs when observed variability is greater than a model expects. It can make standard errors too small if the extra variability is ignored, producing intervals that are too narrow and tests that appear more conclusive than warranted. The primary endpoint is a time-to-event outcome analyzed with Cox regression, not a count outcome for which a basic over-dispersion adjustment would automatically apply. The registry does not report whether any model diagnostics, robust variance estimates, or alternative approaches were used, so no claim about over-dispersion handling can be made for this trial.

The trial's two primary treatment-component contrasts are based on arms combined by presence or absence of each component. General statistical context: when factors are studied in a multi-arm design, the estimated main effect of one factor can depend on how the other factor is handled and on whether interaction is present. A treatment-by-treatment interaction would ask whether the association for one component differs according to the status of the other component. Such a result is not reported in the registry record. The posted hazard ratios should therefore be interpreted as the registry's stated combined comparisons, not as evidence that the two components act independently.

The registry reports the sponsor's own Cox results. The explanatory material also contains planning illustrations based on observed effects, but these are not additional trial outcomes and do not change the interpretation of the sponsor's primary analysis. The trial's complete prespecified statistical analysis plan, details of covariate adjustment, randomization stratification, and multiplicity procedure are not reported in the registry record.

11. Confidence Intervals and Precision

Precision is conveyed by the intervals around estimates, not just by the point estimates or p-values. For concomitant temozolomide, the reported hazard ratio is 0.97 and its 99.1% confidence interval runs from 0.73 to 1.28. The interval includes 1 and spans estimates on either side of the null hazard ratio. This means the point estimate alone should not be read as a precise indication of a lower hazard. The interval allows appreciable uncertainty about both the direction and size of the relative hazard contrast.

For adjuvant temozolomide, the reported hazard ratio is 0.64 and the 95% confidence interval runs from 0.52 to 0.79. This interval is below 1, which gives clearer evidence of a lower estimated hazard in the presence group under the reported model. It still covers a range of possible relative effects; it does not identify one exact effect size. The corresponding p-value is < 0.001, but the interval is more useful for seeing the range of estimates compatible with the analysis.

The median overall-survival confidence intervals also show uncertainty. The absence and presence groups for the concomitant comparison have intervals from 45.7 to 71.49 months and from 48.53 to 82.33 months, respectively. For the adjuvant comparison, the intervals are 37.88 to 56.94 months and 67.19 to 116.6 months. These are separate intervals for group medians; their overlap or separation is not a substitute for the reported Cox comparison.

General statistical context: an interval's width reflects precision under the analysis method and assumptions, while the confidence level describes the method's long-run coverage property. Precision does not establish clinical importance, and a narrow interval cannot remove bias from design or analysis limitations. The registry record does not report all details needed to assess model assumptions, missing-data handling, or the effects of analysis choices.

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 planning illustrations use observed treatment effects as starting points and show event targets for time-to-event outcomes. In these time-to-event entries, each figure is the total number of events required, not a number of participants per arm. The columns for 80% power and 90% power describe planning targets: under the stated assumptions, the former targets an 80% chance and the latter a 90% chance of detecting the specified effect if that effect is present. These are prospective planning quantities, not observed results from the trial.

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 as the stated fraction of the observed effect. The calculation is described as an illustration based on the observed effect, not the trial's own original design. The trial's original sample-size justification and design assumptions are not reported in the registry record. The planning entries use total events, following the Schoenfeld approach.

Rows that use 100% of the observed effect plan around the full estimated effect; rows using 75% or 50% plan around smaller effects. Smaller effects are harder to distinguish from no effect, so they require more information and generally more events to reach the same power target. A higher power target also requires more information under otherwise comparable assumptions. The event targets can be very large when the observed effect is small, so they should be read as conditional planning illustrations rather than as a description of the trial's actual enrollment or event count.

These figures are not observed power. Observed power is not a useful substitute for interpreting the estimated effect and its confidence interval after a trial has been completed; the posted result and its uncertainty are the relevant evidence. The planning calculations should not be interpreted as proof that the trial was designed with these assumptions, nor as evidence that the trial did or did not have adequate power.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

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

Worked calculation for this trial

1. Hazard ratio HR = 0.970, so ln(HR) = -0.0305.
2. d = 4 × (1.96 + 0.842)² / [ln(0.970)]² = 33,840.0, rounded up to 33,841 events in total for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
hr <- 0.97                              # 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(626990)
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 = 33,841 (80% power), 45,303 (90% power)
 75% of observed effect: events = 60,161 (80% power), 80,538 (90% power)
 50% of observed effect: events = 135,361 (80% power), 181,209 (90% power)
Simulated power with 33,841 events: 0.800 (20000 replicates)

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

14. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 135,787 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 751 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 60.42          # control-arm median, months (converted from the posted unit)
hr          <- 0.97          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
      Analysis               Value Efficacy
     IA 1: 33%                   Z   3.7103
     N: 134090         p (1-sided)   0.0001
 Events: 15280        ~HR at bound   0.9417
     Month: 24    P(Cross) if HR=1   0.0001
               P(Cross) if HR=0.97   0.0338
     IA 2: 67%                   Z   2.5114
     N: 135788         p (1-sided)   0.0060
 Events: 30559        ~HR at bound   0.9717
     Month: 38    P(Cross) if HR=1   0.0060
               P(Cross) if HR=0.97   0.5603
         Final                   Z   1.9930
     N: 135788         p (1-sided)   0.0231
 Events: 45839        ~HR at bound   0.9816
     Month: 60    P(Cross) if HR=1   0.0250
               P(Cross) if HR=0.97   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

The sponsor-reported analysis uses Cox regression and provides hazard ratios, confidence intervals, and p-values for the two treatment-component comparisons. The registry also presents a recalculated unadjusted comparison based on posted summary statistics. Such a recalculation is a simple check derived from summary data, not a replacement for the sponsor's full analysis, which may rely on participant-level follow-up and the model specified by the investigators.

For this record, the discrepancy indicator is false, so there is no identified disagreement requiring a separate qualification. The sponsor's own analysis remains the primary result to interpret. It reports a hazard ratio of 0.97 for the concomitant comparison, with a 99.1% confidence interval from 0.73 to 1.28 and p = 0.76. For the adjuvant comparison, it reports a hazard ratio of 0.64, with a 95% confidence interval from 0.52 to 0.79 and p < 0.001.

Recalculated estimates should be treated as checks on the posted summaries only. They cannot reproduce all aspects of a Cox model from medians and marginal confidence intervals, including participant-level event timing and censoring. The registry record does not report additional details about covariates or model diagnostics. Accordingly, the sponsor-reported Cox analysis takes precedence when describing the trial's inferential results, and the recalculation should not be presented as an independent confirmation from a separate dataset.

16. Multiplicity, Missing Data and Analysis Populations

The record reports two primary Cox comparisons, one for concomitant temozolomide and one for adjuvant temozolomide, alongside median estimates for the combined groups. It also lists secondary endpoints. When several hypotheses are tested, the chance of a false-positive result across the set can be higher than for a single test. General statistical context: multiplicity can be addressed through a prespecified hierarchy, an adjustment to significance thresholds, or other procedures suited to the analysis plan. Whether and how multiplicity was handled is not reported in the registry record.

The reported p-values should therefore be interpreted as the values attached to the sponsor's stated analyses, not as evidence that all endpoint and contrast testing has been controlled as a family. The registry does not report whether the two primary comparisons were co-primary, whether a particular result was required before testing another, or whether the secondary analyses were confirmatory or exploratory. Those features affect how strongly a reader should interpret isolated statistical significance.

The primary outcome description explains that participants not reported dead or lost to follow-up were censored at the last follow-up examination. It does not provide detailed counts or reasons for loss to follow-up, missing outcome data, or exclusions from the analysis. The exact analysis population—such as whether all randomized participants were included according to assigned arm—is not reported in the registry record. The handling of missing covariates, protocol deviations, and any sensitivity analyses is also not reported in the registry record.

General statistical context: censoring and missingness are not automatically equivalent to a complete absence of information. Their impact depends on why follow-up ends and whether that process is related to the outcome. Transparent reporting of follow-up completeness and sensitivity analyses helps readers judge how robust survival estimates are to those issues. The registry summary does not report enough detail to assess those matters.

The record also does not report an interaction analysis for the concurrent and adjuvant treatment components, or the model details needed to determine how both components were represented together. These omissions limit conclusions about whether treatment timing effects are independent, whether one component modifies the other, or how the reported combined contrasts should be generalized to each specific regimen.

17. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: neurology (8.1%), infection (4.4%), constitutional symptoms (3.8%), blood (2.7%), gastrointestinal (2.7%).

The registry reports serious adverse event counts by original randomized arm, with an at-risk count and an affected count for each. The serious-event figures are 13 of 186 for radiotherapy alone, 27 of 185 for radiotherapy with concurrent chemotherapy, 32 of 183 for radiotherapy with adjuvant chemotherapy, and 32 of 185 for radiotherapy with both concurrent and adjuvant chemotherapy. These counts describe the posted groups; they are not adjusted comparisons and should not be used to infer that a treatment component caused a difference.

The listed leading serious-event categories are neurology at 8.1%, infection at 4.4%, constitutional symptoms at 3.8%, blood at 2.7%, gastrointestinal at 2.7%, and pain at 2.2%. The registry also reports counts for other adverse events and deaths by arm, but these categories and counts alone do not describe event severity beyond the recorded classification, timing, duration, or attribution. The record does not report the criteria used for determining relatedness to treatment in this summary.

Safety counts should be considered with the size of each at-risk group, follow-up, event definitions, and completeness of reporting. A serious adverse event count is not the same as a count of all adverse events, and the registry's “other” event counts are a separate summary. No causal conclusion about safety can be drawn from these descriptive figures alone.

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

The primary registry analysis shows a distinction between the two treatment-component comparisons. The concomitant-temozolomide comparison has a hazard-ratio interval that includes 1 and a p-value of 0.76, so the posted analysis does not provide clear statistical evidence of a difference in overall survival for that contrast. The adjuvant-temozolomide comparison has a hazard-ratio interval below 1 and p < 0.001, indicating a lower estimated hazard in the presence group under the sponsor's Cox model. The estimate remains uncertain within the bounds of its confidence interval.

These results describe group-level statistical comparisons in a randomized trial. They do not guarantee a particular survival time for an individual, and they do not show that every participant benefits or experiences the same course. The median estimates are population summaries, not predictions for a specific patient. The registry record also does not establish the biological mechanism behind the observed associations.

The two comparisons combine the four randomized arms according to whether concurrent or adjuvant temozolomide was present. They should not be read as direct estimates for every pair of regimens. Without a reported interaction analysis, the results do not establish whether the effect associated with one component changes depending on the other component. The record also does not supply the full statistical analysis plan, multiplicity procedures, detailed missing-data information, or analysis-population definition.

The secondary progression-free-survival pattern gives additional context but does not replace overall survival. Other listed secondary outcomes lack detailed numerical results in the registry record. Serious adverse event counts are descriptive and do not by themselves establish treatment causation. A balanced interpretation considers the estimated effects, their confidence intervals, the endpoint definitions, trial design, and limits of registry reporting together.

19. Why This Trial Matters Statistically

This trial is statistically informative because it randomized patients among regimens that varied two treatment components: temozolomide given with radiotherapy and temozolomide given after radiotherapy. Combining the four arms by component allows the registry to report separate presence-versus-absence contrasts rather than only a single comparison of one regimen against radiotherapy alone. That structure can address distinct treatment-timing questions, while also making it important to understand exactly how the arms were combined.

The primary results illustrate why a point estimate should not be read alone. The concomitant comparison has an estimated hazard ratio close to 1, but the interval is broad enough to include values below and above 1. The adjuvant comparison has a hazard-ratio interval below 1 and a small reported p-value. The difference in the strength of evidence is clearer when estimates, uncertainty, and endpoint definitions are considered together.

The record also demonstrates a core issue in time-to-event analysis: follow-up can vary, and some participants are censored rather than observed to have the event. Survival models use the available time-to-event information, but interpretation depends on assumptions and transparent reporting. The posted summary gives the basic censoring description and model type, while leaving several analysis details unreported.

Finally, the planning illustrations distinguish event targets from observed trial outcomes. They can show how design targets depend on the assumed effect size and power goal, but they are not the trial's original sample-size rationale. Keeping planning calculations separate from the sponsor's results helps prevent a common misreading: a design illustration cannot establish whether the observed analysis was adequately powered or whether a result is clinically important.

20. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several trial-specific limits affect interpretation. The posted primary comparisons combine arms by presence or absence of concurrent and adjuvant temozolomide. This framing does not provide every direct regimen-to-regimen contrast, and an interaction analysis explaining whether the two components modify each other's effects is not reported in the registry record.

The record identifies Cox regression but does not report the full model specification, covariates, stratification variables, proportional-hazards diagnostics, or sensitivity analyses. The extent to which the estimates depend on those choices therefore cannot be evaluated from the registry summary alone. The trial's prespecified analysis plan and multiplicity strategy are also not reported in the registry record.

The registry describes censoring at last follow-up for people not reported dead or lost to follow-up, but detailed follow-up completeness, reasons for loss, missing-data handling, and analysis-population definitions are not reported. These details are relevant to assessing potential bias in time-to-event estimates. The record does not provide detailed numerical results for all listed secondary outcomes, limiting comparison across the full outcome set.

Safety reporting is summarized by counts and categories, without enough detail in this record to assess timing, duration, attribution, or the criteria applied to serious events. The registry also labels the trial active, not recruiting, while reporting a primary completion date; the record does not explain the relationship between those status fields. These limitations do not negate the posted results, but they constrain how much can be concluded from the registry summary alone.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Phase III Trial of Anaplastic Glioma Without 1p/19q Loss of Heterozygosity (LOH) — NCT00626990.
• Registry record last updated 2023-09-11; status last verified 2023-08.
• Results first posted 2023-09-11.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

24. Record Summary

The registry describes a randomized, unmasked, parallel-group phase III trial of temozolomide timing with radiotherapy for anaplastic glioma without 1p/19q deletion or loss of heterozygosity. Four arms varied concurrent and adjuvant temozolomide. The primary endpoint was overall survival from randomization to death from any cause, with censoring at last follow-up for patients not reported dead or lost to follow-up.

The sponsor-reported Cox analysis shows less conclusive evidence for the concomitant comparison and stronger evidence of a lower estimated hazard for the adjuvant comparison. Progression-free survival is also reported, with a broadly similar pattern. Quality of life and neurological deterioration-free survival are listed, but detailed numerical results for those outcomes are not reported in the registry record.

Interpretation is limited by unreported details about multiplicity, analysis populations, missing data, model specification, and interaction between treatment components. The planning illustrations are based on observed effects and stated assumptions, not the trial's original design rationale or observed power.