This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results and design facts on this page are restricted to the ClinicalTrials.gov record.
1. Trial at a Glance
RTOG 0424 was a phase 2, single-group oncology study of temozolomide and radiation therapy in patients with brain and central nervous system tumors. The registry reports 136 participants, no allocation because the study used a single-group design, and no masking.
| Feature | RTOG 0424 |
|---|---|
| Trial name | RTOG 0424 |
| Brief title | Temozolomide and Radiation Therapy in Treating Patients With Gliomas |
| Phase | Phase 2 |
| Status | Completed |
| Study design | Single-group |
| Masking | None |
| Allocation | Not applicable |
| Primary purpose | Treatment |
| Enrollment | 136 |
| Interventions | Temozolomide; radiation therapy |
| Primary endpoint types | Binary; time-to-event |
| Results posted | Yes |
| Statistical analyses posted | 3 |
| ClinicalTrials.gov | NCT00114140 |
2. Clinical Question
The trial evaluates temozolomide and radiation therapy in patients with gliomas within the registered condition category of Brain and Central Nervous System Tumors. Because the study is single-group, its principal statistical structure differs from a randomized two-arm trial: there is no concurrent randomized comparator in the ClinicalTrials.gov record.
Population
Patients with gliomas, within the registered condition category of Brain and Central Nervous System Tumors.
Intervention
Temozolomide and radiation therapy.
Comparator
No concurrent comparator arm is registered. The design model is single-group.
Primary statistical questions
What is the reported 3-year overall survival rate, and how are survival and progression-free survival associated with MGMT methylation status?
3. Trial Design
Temozolomide + Radiation Therapy
- Temozolomide
- Radiation therapy
- Primary purpose: treatment
- Masking: none
- Design model: single-group
Interpretive consequence
- There is no concurrent randomized comparator in the registered design.
- The MGMT analyses compare methylated and unmethylated status within eligible patients who have MGMT status.
- Those MGMT comparisons are prognostic comparisons within the treated study population rather than randomized treatment-effect comparisons.
4. Trial Timing and Registry Profile
Study start
The registry lists the study start as 2005-01.
Primary completion
The registry lists primary completion as 2013-02.
Registry status
The trial is listed as completed, with results posted on ClinicalTrials.gov.
| Registry feature | Reported value |
|---|---|
| Lead sponsor | Radiation Therapy Oncology Group |
| Sponsor type | Network |
| Therapeutic area | Oncology |
| Condition | Brain and Central Nervous System Tumors |
| Enrollment | 136 |
| Arms | 1 |
| Allocation | Not applicable |
| Design model | Single-group |
| Masking | None |
5. Endpoints
The registry lists five primary endpoints. They span time-to-event outcomes and assessments classified in the ClinicalTrials.gov record as binary endpoints. The definitions below preserve the registry wording available in the ClinicalTrials.gov record.
| Primary endpoint | Time frame | Registry definition / assessment |
|---|---|---|
| Overall Survival Rate at 3 Years | Registration to 3 years | Survival time is defined as time from registration to date of death from any cause and is estimated by the Kaplan-Meier method. Patients last known to be alive are censored at the date of last contact. This analysis was planned to occur when all patients had been potentially followed for at least 3 years. |
| Progression-free Survival | From registration to last follow-up, up to 7.1 years. Analysis occurs after all patients have been on study for at least 3 years. | Progressive Disease (PD) is defined as 25% or > increase in the cross-sectional area of enhancing or non-enhancing tumor on consecutive MRI scans, or any new area(s) of tumor. Under exceptional circumstances, disease progression may be declared in the absence of an increase in tumor size based on "clinical deterioration" including the need for increasing doses of steroid and/or a worsening Karnofsky Performance Status(KPS) / Neurologic Function Score(NFS). Progression-free survival time is defined as time from registration to date of progressive disease or death from any cause and is estimated by the Kaplan-Meier method. Patients last known to be alive are censored at the date of last contact. Median survival time is reported.. |
| Survival and Progression-free Survival by O(6)-Methylguanine-DNA Methyltransferase (MGMT) Methylation Status | Registration to 3 years | Survival time is defined as time from registration to date of death from any cause. Progressive Disease (PD) is defined as 25% or > increase in the cross-sectional area of enhancing or non-enhancing tumor on consecutive MRI scans, or any new area(s) of tumor. Under exceptional circumstances, disease progression may be declared in the absence of an increase in tumor size based on "clinical deterioration" including the need for increasing doses of steroid and/or a worsening Karnofsky Performance Status(KPS) / Neurologic Function Score(NFS). Survival and progression-free survival are estimated by the Kaplan-Meier method. Patients last known to be alive are censored at the date of last contact.. |
| Quality of Life as Measured by the Functional Assessment of Cancer Therapy Scale With Brain Module (FACT-BR) | Baseline, 6 months, and 12 months. | Functional Assessment of Cancer Therapy Scale with brain module (FACT-BR): a 50-question self-report questionnaire contains the following domains (scales): Physical well-being (7 questions totalling 0-28), social/family well-being (7 questions totalling 0-28), emotional well-being (6 questions totalling 0-24), functional well-being (7 questions totalling 0-28) and brain cancer subscale which contains concerns relevant to patients with brain tumors (19 questions totalling 0-76). Each question has a value 0-4. For some questions a higher indicates better outcome and others are the opposite. The former are summed as is, the latter are reversed in value before adding, such that each domain ranges from 0 to 4 multiplied by the number of questions in the domain, with 0 indicating worst and the highest possible value indicating best outcome. The FACT-Br total (0-184) is obtained by adding all domains together if the overall question response rate is greater than 80%. |
| Neurocognitive Function | Baseline, 6 months, and 12 months. | Hopkins Verbal Learning Test (HVLT) is a test measuring learning memory retrieval, and memory consolidation processes.; Controlled Oral Word Association Test (COWAT) is a test of phonemic verbal fluency. The patient produces as many words as possible in 1 min. (each) for a specific letter (C, F, L or P, R, W).; Trail Making Test (TMT) is a measure of visuospatial scanning, attention, sequencing, and speed in Part A (TMT A) and executive function in Part B (TMT B). Patients must "connect the dots" either in a numbered sequence or alternating letters and numbers. Difference between pre-treatment baseline and follow-up assessment scores determined by the reliable change (RC) index, using a 90% confidence interval to designate statistically significant change. |
6. Statistical Methodology
Kaplan-Meier estimation
The registered overall-survival endpoint explicitly specifies the Kaplan-Meier method. This is appropriate for a time-to-event endpoint because not every participant necessarily experiences the event during the observation period. Participants who are last known to be alive are censored at the date of last contact according to the registered definition.
Here, di represents events at an observed event time and ni represents the number at risk immediately before that time. The registry specifically identifies Kaplan-Meier estimation for the 3-year overall-survival endpoint.
Log-rank testing
The registry reports the log-rank test for the MGMT methylation analyses. A log-rank test compares the observed pattern of event times between groups while accounting for censoring. In this study, the compared groups were MGMT-methylated and MGMT-unmethylated patients within the treated cohort.
Hazard ratios
The effect measure reported for the MGMT analyses is the hazard ratio. Because the analysis notes identify methylated status as the reference level, the reported hazard ratios describe the relative event hazard associated with the comparison against that reference.
With methylated status as the reference level, an HR above 1 indicates a higher estimated event hazard for the comparison group relative to the methylated group, under the fitted time-to-event analysis.
Wald / z-test
The 3-year overall-survival-rate analysis is reported using a Z-test, normalized in the ClinicalTrials.gov record to the Wald / z-test family. This analysis produces a P-value rather than a reported hazard ratio or confidence interval in the ClinicalTrials.gov record.
Superiority hypotheses
The ClinicalTrials.gov record classifies all three posted statistical analyses as superiority hypotheses. For the MGMT analyses, superiority refers to the statistical comparison of outcomes by methylation status; it should not be confused with a randomized treatment-versus-control superiority claim because the trial itself is single-group.
7. Results: Overall Survival Rate at 3 Years
The primary overall-survival endpoint was defined as time from registration to death from any cause, with Kaplan-Meier estimation and censoring at the date of last contact for patients last known to be alive. The registry reports a formal Z-test analysis in eligible patients.
3-year overall survival rate
Method: Z-test · Hypothesis type: Superiority
The registry does not report an estimate or confidence interval for this statistical analysis in the ClinicalTrials.gov record.
The analysis notes state that, assuming an exponential distribution of survival times, the null hypothesis was a 43% increase in median survival time (40.5 mo. to 57.9 mo.) and a 21% increase in 3-year survival rate (54% to 65%). Assuming 5% ineligibility, 72 patients were required to be accrued over 3 years with 3-years of follow-up resulting in 80% power and 1-sided 0.10 significance level. The registry-reported text notes that N was later increased to 135 to accommodate additional requirements.
What the P-value means: the registry reports P < 0.001 for the Z-test of the 3-year overall-survival-rate endpoint. Under the statistical framework used for this analysis, this is evidence against the specified null hypothesis.
What it does not mean: a P-value is not the magnitude of the treatment effect, does not give the probability that the treatment works, and does not establish how much longer an individual patient will live.
Precision: no confidence interval for the overall-survival-rate estimate is reported in the ClinicalTrials.gov record, so precision of the observed 3-year survival estimate cannot be assessed from the reported analysis alone.
Important design caution: because RTOG 0424 is a single-group study, this result is not a randomized comparison against a concurrent control arm. The analysis notes instead define a historical/statistical benchmark through the specified survival assumptions. The distinction matters when translating a one-group survival result into a treatment-effect claim.
8. Results: Survival by MGMT Methylation Status
The registry reports a primary endpoint analysis comparing eligible patients with MGMT status according to methylation status. The groups are listed as Temozolomide + Radiation Therapy (RT) - Methylated vs Temozolomide + Radiation Therapy (RT) - Unmethylated, with methylated status identified as the reference level.
Overall survival hazard ratio
95% CI: 1.64–7.56 · P = 0.0006
Method: Log-rank · Analysis population: eligible patients with MGMT status
What the estimate means: with methylated status as the reference level, the reported HR of 3.52 indicates a substantially higher estimated hazard of death for the comparison group relative to the methylated reference group, within this treated cohort and under the reported time-to-event analysis.
What it does not mean: it is not a hazard ratio for temozolomide and radiation therapy versus no treatment or versus another randomized treatment. The study has one treatment group. The analysis is instead a comparison of outcomes by MGMT methylation status.
What the confidence interval says: the two-sided 95% CI is 1.64–7.56. It quantifies uncertainty around the estimated hazard ratio under the statistical model and sampling framework. It is not a range in which individual patients' effects must fall.
Why the P-value is different from the effect size: P = 0.0006 describes the strength of evidence against the statistical null used by the analysis; it does not measure the size of the association. The HR and its confidence interval provide the effect estimate and its precision.
Model caution: hazard-ratio interpretation relies on the assumptions of the underlying time-to-event analysis. The ClinicalTrials.gov record does not provide a diagnostic assessment of proportional hazards, so the HR should be treated as a model-based summary rather than a complete description of how hazards changed over time.
9. Results: Progression-free Survival by MGMT Methylation Status
The second statistical analysis of the MGMT endpoint evaluated progression-free survival from registration to the specified 3-year time frame. The same methylated reference level was used.
Progression-free survival hazard ratio
95% CI: 1.55–6.04 · P = 0.0007
Method: Log-rank · Analysis population: eligible patients with MGMT status
What the estimate means: with methylated status as the reference level, the reported PFS HR of 3.06 indicates a higher estimated hazard of progression or the relevant PFS event in the comparison group relative to the methylated reference group, within the treated study population.
What it does not mean: the HR does not say that an individual patient's progression risk is exactly 3.06 times higher at every point in time, nor does it represent a randomized treatment effect.
Confidence interval: the two-sided 95% CI of 1.55–6.04 expresses uncertainty around the estimated HR. Its width reflects that the estimate is not known with unlimited precision.
P-value: P = 0.0007 indicates strong statistical evidence under the reported analysis, but the P-value itself does not quantify clinical magnitude or patient-level benefit.
Censoring and endpoint definition: PFS incorporates the registry's definition of progressive disease, including the stated MRI-based criteria and the provision for clinical deterioration under exceptional circumstances. Consequently, the statistical result depends not only on the survival-analysis method but also on how progression events were defined and assessed.
| MGMT analysis | Estimate | 95% CI | P-value | Method |
|---|---|---|---|---|
| Overall survival | HR 3.52 | 1.64–7.56 | 0.0006 | Log-rank |
| Progression-free survival | HR 3.06 | 1.55–6.04 | 0.0007 | Log-rank |
10. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm as 37/129 for the temozolomide plus radiation therapy group. Because the study has one registered treatment group, there is no second randomized arm in the ClinicalTrials.gov record for a comparative serious-adverse-event analysis.
| Safety measure | Temozolomide + Radiation Therapy (RT) |
|---|---|
| Serious adverse events | 37/129 affected / at risk |
11. Quality of Life and Neurocognitive Endpoints
The registry lists quality of life and neurocognitive function among the primary endpoints, both assessed at baseline, 6 months, and 12 months. However, the ClinicalTrials.gov record does not contain formal statistical analyses for these endpoints. The page therefore does not assign an outcome estimate or P-value to them.
FACT-BR
The Functional Assessment of Cancer Therapy Scale with brain module is registered as a 50-question self-report questionnaire covering physical, social/family, emotional, functional, and brain-cancer domains in the registry description.
Neurocognitive function
The registry identifies the Hopkins Verbal Learning Test, Controlled Oral Word Association Test, and Trail Making Test among the neurocognitive assessments.
12. Statistical Methods Explained
Why was Kaplan-Meier estimation used?
Overall survival is a time-to-event endpoint: both whether death occurred and when it occurred matter. Kaplan-Meier estimation incorporates different follow-up times and right-censoring rather than requiring every participant to have an observed death by the analysis date. The registry explicitly specifies Kaplan-Meier estimation for the 3-year overall-survival endpoint.
What does an HR of 3.52 mean in the MGMT analysis?
The registry identifies methylated MGMT status as the reference level. Therefore, the reported HR of 3.52 is a relative hazard estimate for the comparison against that reference. It is a model-based summary of the difference in event hazards, not a statement that every comparison-group patient has exactly 3.52 times the event probability.
Why is the confidence interval important?
The HR of 3.52 is only an estimate. Its two-sided 95% CI of 1.64–7.56 communicates uncertainty around that estimate. A confidence interval is therefore more informative than a P-value alone because it combines the estimated effect with information about precision.
Why doesn't the P-value measure the effect size?
A P-value addresses statistical evidence under a specified null hypothesis. It does not tell the reader how large the association is. For the MGMT overall-survival analysis, the effect size is represented by HR 3.52, while P = 0.0006 addresses statistical evidence. The same distinction applies to PFS, where HR 3.06 is accompanied by P = 0.0007.
Why is the MGMT comparison different from a randomized treatment comparison?
RTOG 0424 is a single-group study. All participants received the registered treatment approach, and the MGMT analysis separates patients according to biomarker status. Consequently, MGMT status is not a randomized intervention. The observed difference can be interpreted as an association between MGMT status and outcome in the treated cohort, not as evidence that changing MGMT status would cause the observed difference.
What does the log-rank test contribute?
The log-rank test provides a formal comparison of the time-to-event experience between the MGMT groups while accounting for censoring. It complements the hazard ratio: the test addresses statistical evidence against the null comparison, while the HR quantifies the estimated relative hazard and the confidence interval communicates its precision.
Why was a Z-test used for the 3-year survival endpoint?
The registry reports a Z-test for the 3-year overall-survival-rate analysis. Its analysis notes specify the survival assumptions and benchmark values used for the design, including a 1-sided 0.10 significance level and 80% power under the stated planning assumptions. The ClinicalTrials.gov record does not provide a confidence interval or an effect estimate for this Z-test.
13. Interpreting the Three Primary Statistical Analyses Together
The three posted analyses address related but distinct questions. The 3-year overall-survival analysis evaluates the study cohort against the statistical framework specified for the overall-survival rate. The two MGMT analyses compare time-to-event outcomes according to methylation status among eligible patients with MGMT status.
| Analysis | Question addressed | Effect measure | Statistical test |
|---|---|---|---|
| Overall Survival Rate at 3 Years | Does the observed 3-year survival experience differ from the specified planning benchmark? | No estimate reported in registry-reported analysis | Z-test |
| OS by MGMT status | Do overall-survival times differ by MGMT methylation status? | HR 3.52 (95% CI 1.64–7.56) | Log-rank |
| PFS by MGMT status | Do progression-free survival times differ by MGMT methylation status? | HR 3.06 (95% CI 1.55–6.04) | Log-rank |
The distinction between these analyses is essential. The first is a one-group survival-rate analysis tied to the trial's design assumptions. The latter two are within-cohort biomarker comparisons. None of the analyses posted on ClinicalTrials.gov is a randomized comparison between temozolomide plus radiation therapy and a concurrent control treatment.
14. Confidence Intervals and P-values
Overall survival by MGMT
HR 3.52 with a two-sided 95% CI of 1.64–7.56 and P = 0.0006. The interval gives a range of plausible values for the modeled hazard ratio under the specified inferential framework.
Progression-free survival by MGMT
HR 3.06 with a two-sided 95% CI of 1.55–6.04 and P = 0.0007. The estimate and interval indicate the magnitude and precision of the reported association.
The fact that both confidence intervals lie above 1 is consistent with the corresponding small P-values under the reported analyses. But statistical significance and clinical importance are different concepts. Clinical interpretation requires consideration of the endpoint definition, the magnitude of the estimated effect, uncertainty, study design, and the absence of a randomized concurrent comparator.
15. Multiplicity, Interim Analysis, Crossover, and Bayesian Methods
The ClinicalTrials.gov record does not report a multiplicity-adjustment strategy, a crossover procedure, a Bayesian analysis, or an interim-analysis procedure for the three posted statistical analyses. These topics are therefore not assigned methods or interpretations beyond what can be supported by the registry extract.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Multiplicity | No multiplicity procedure is reported in the ClinicalTrials.gov record. |
| Interim analysis | No interim-analysis method is reported in the ClinicalTrials.gov record. |
| Crossover | No crossover procedure is reported in the ClinicalTrials.gov record. |
| Bayesian methods | No Bayesian method is reported in the ClinicalTrials.gov record. |
| Missing-data imputation | No specific imputation method is reported in the ClinicalTrials.gov record. |
| Stratification | No stratification factors are reported for the posted statistical analyses in the ClinicalTrials.gov record. |
| Non-inferiority margin | No non-inferiority margin is reported; the analyses posted on ClinicalTrials.gov are classified as superiority hypotheses. |
16. Sample Size and Power
The registry-reported analysis notes provide the original planning framework for the overall-survival analysis. Assuming an exponential distribution of survival times, the stated null hypothesis involved a 43% increase in median survival time from 40.5 mo. to 57.9 mo. and a 21% increase in 3-year survival rate from 54% to 65%.
The analysis notes state that, assuming 5% ineligibility, 72 patients were required to be accrued over 3 years with 3-years of follow-up. The sample size was later increased to 135 to accommodate additional requirements, while the registered enrollment is 136.
17. Analysis Populations and Reference Levels
| Analysis | Analysis population | Reference / comparison |
|---|---|---|
| Overall Survival Rate at 3 Years | Eligible patients | Temozolomide + Radiation Therapy (RT) |
| OS by MGMT status | Eligible patients with MGMT status | Methylated reference level versus unmethylated comparison |
| PFS by MGMT status | Eligible patients with MGMT status | Methylated reference level versus unmethylated comparison |
Analysis-population definitions matter because a result applies to the population actually included in that analysis. The two MGMT results, for example, are explicitly restricted to eligible patients with MGMT status. They should not automatically be generalized to every enrolled participant.
18. What the Hazard Ratio Does — and Does Not — Mean
The reported HR 3.52 is a relative time-to-event measure comparing the MGMT groups with methylated status as the reference. In the context of this analysis, the estimate indicates a higher modeled hazard in the comparison group.
It does not mean that the comparison group has exactly 3.52 times the probability of death, nor does it quantify the causal effect of temozolomide and radiation therapy.
The reported HR 3.06 similarly summarizes the relative progression-free event hazard between the MGMT groups. It is not a direct measure of absolute progression probability and should not be read as a patient-level multiplier that is necessarily constant over time.
A hazard ratio is a relative measure. For a complete clinical interpretation, readers ideally want absolute survival probabilities, median event times, and their uncertainty. The ClinicalTrials.gov record does not provide those additional MGMT-specific quantities, so the page does not reconstruct them.
19. Important Limitations and Interpretation Issues
- Single-group design: the study has one registered treatment group and no concurrent randomized comparator. This limits the extent to which the observed survival experience can be attributed causally to the intervention rather than to the underlying patient population or other factors.
- MGMT is not randomized: the methylated versus unmethylated analyses are within-cohort biomarker comparisons. They describe an association between MGMT status and outcome rather than a randomized treatment effect.
- Hazard-ratio assumptions: a hazard ratio is a model-based summary of time-to-event experience. The ClinicalTrials.gov record does not report an assessment of the proportional-hazards assumption.
- Censoring: overall survival explicitly censors patients last known to be alive at the date of last contact. The validity of Kaplan-Meier inference depends on the appropriateness of the censoring mechanism.
- Endpoint complexity: progression-free survival depends on the registered definition of progressive disease, including MRI-based tumor changes and the stated exceptional clinical-deterioration criterion.
- Analysis population: the MGMT analyses include eligible patients with MGMT status, which is narrower than the overall eligible population.
- Quality-of-life amendment: the registry notes that the quality-of-life endpoint was added through an amendment while the study was in progress.
- Limited reported statistical detail: the ClinicalTrials.gov record reports three statistical analyses but does not provide formal statistical analyses for the other posted outcome measures.
- No reported multiplicity framework: the ClinicalTrials.gov record does not specify how the multiple primary endpoints or analyses were incorporated into a familywise type-I-error framework.
20. Why This Trial Matters Statistically
RTOG 0424 is a useful teaching case because it illustrates how statistical interpretation changes when a trial has a single treatment group rather than randomized concurrent controls. It also connects a prespecified survival-rate analysis with biomarker-based time-to-event comparisons.
| Concept | How it appears in RTOG 0424 |
|---|---|
| Single-group design | The registry identifies the design model as single-group with 1 arm. |
| Kaplan-Meier estimation | Explicitly specified for the 3-year overall-survival endpoint. |
| Time-to-event endpoints | Overall survival and progression-free survival are registered as time-to-event endpoints. |
| Log-rank test | Used for the posted MGMT survival and progression-free survival analyses. |
| Hazard ratio | Reported for both MGMT overall survival and MGMT progression-free survival. |
| Confidence interval | Two-sided 95% CIs accompany both reported MGMT hazard ratios. |
| P-values | Reported as 0.0006 and 0.0007 for the MGMT analyses and <0.001 for the 3-year survival-rate Z-test. |
| Wald / z-test | Used for the 3-year overall-survival-rate analysis. |
| Biomarker comparison | MGMT methylated versus unmethylated status is evaluated within the treated cohort. |
| Planning assumptions | The overall-survival analysis notes specify survival-rate, median-survival, power, significance-level, and accrual assumptions. |
| Endpoint amendment | The registry notes that the quality-of-life endpoint was added while the study was in progress. |
21. A Statistical Reading of the Trial
The most important statistical distinction is between estimating outcomes in a single treatment cohort and estimating a randomized treatment effect. RTOG 0424 provides the former. Its overall-survival analysis asks whether the cohort's survival experience is statistically different from the specified planning benchmark, while the MGMT analyses ask whether outcomes differ between biomarker-defined groups within that cohort.
The MGMT findings are expressed particularly clearly through their hazard ratios and confidence intervals: HR 3.52 for overall survival and HR 3.06 for progression-free survival, with both two-sided 95% confidence intervals above 1 and small reported P-values. These statistics quantify the observed association under the registered analysis. They do not convert the single-group trial into a randomized comparison.
This distinction is central to clinical-trial statistics. A statistically strong biomarker association can be scientifically informative while still answering a different question from whether a treatment improves outcomes relative to an alternative treatment. The design determines which causal conclusions are supported.
22. Related Tutorials
Learn more about the methods used in this trial:
23. Related Statistical Calculators
24. Limitations of the Reported Evidence
The ClinicalTrials.gov record supports three formal statistical analyses, but it does not provide a complete statistical-analysis dataset for every registered primary endpoint. In particular, formal statistical results are not posted on ClinicalTrials.gov for the FACT-BR quality-of-life endpoint or the neurocognitive-function endpoint.
The single-group design is the central interpretive limitation. Without a concurrent randomized comparator, observed survival cannot be separated as cleanly from patient selection, disease characteristics, background prognosis, and other temporal or clinical factors as it could be in a randomized controlled comparison. The trial's overall-survival analysis instead relies on the specified statistical benchmark in the planning assumptions.
The MGMT results have a different purpose. They compare patients within the same treated cohort according to methylation status. Such analyses can be useful for understanding prognostic associations, but they do not demonstrate that the treatment effect differs by MGMT status because there is no randomized treatment-by-MGMT comparison in the ClinicalTrials.gov record.
25. Sources
- ClinicalTrials.gov: RTOG 0424, NCT00114140.
- Linked publication: PubMed record for PMID 29955793.
Continue through Clinical Biostats
Use the related statistical tutorials and calculators to explore the time-to-event methods, hazard ratios, confidence intervals, and hypothesis tests used in this trial.
26. Record Summary
RTOG 0424 is a phase 2, single-group study of temozolomide and radiation therapy with 136 registered participants. Its posted statistical analyses combine a Z-test for the 3-year overall-survival-rate endpoint with log-rank analyses of overall survival and progression-free survival according to MGMT methylation status. The reported MGMT hazard ratios were 3.52 for overall survival (95% CI 1.64–7.56; P = 0.0006) and 3.06 for progression-free survival (95% CI 1.55–6.04; P = 0.0007). The 3-year overall-survival-rate analysis reported P < 0.001.
The key statistical lesson is that these results answer different questions. The overall-survival-rate analysis evaluates a single treatment cohort against a specified statistical benchmark, whereas the MGMT analyses compare time-to-event outcomes between biomarker-defined groups within that cohort. The reported confidence intervals, hazard ratios, P-values, endpoint definitions, censoring rules, and single-group design all need to be considered together when interpreting the evidence.