Bevacizumab With or Without Anti-Endoglin Monoclonal Antibody TRC105 in Treating Patients With Recurrent Glioblastoma Multiforme: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Bevacizumab With or Without Anti-Endoglin Monoclonal Antibody TRC105 in Treating Patients With Recurrent Glioblastoma Multiforme (NCT01648348), a randomized phase 1, phase 2 study. The primary outcome is Maximum Tolerated Dose (MTD) (Phase I) as Measured by the Number of Participants With Dose Limiting Toxicities.

2. Clinical Question

The central statistical question differs by trial phase. Phase I was intended to characterize dose-limiting toxicity and identify a maximum tolerated dose. The registry defines that dose using observed toxicity at a dose level and at the next higher dose level. The posted primary results are counts of participants with dose-limiting toxicities across the listed phase I dose cohorts.

Phase II addressed a comparative question: how outcomes in the bevacizumab-plus-TRC105 arm compare with outcomes in the bevacizumab-alone arm. The registry identifies the trial as randomized and unmasked, with a parallel-group model. The phase II overall-survival result is the principal posted comparative estimate available in the results. It is reported as medians for each arm and as a hazard ratio with a confidence interval and p-value.

The record also lists toxicity, progression-free survival at 6 months, and quality-of-life outcomes. Those outcomes can add context, but they do not replace the phase I dose-finding endpoint or the phase II survival comparison. The registry record does not report a single combined primary question that treats these different phases and outcomes as interchangeable.

3. Trial Design

NCT01648348 is a completed phase I/phase II trial sponsored by the National Cancer Institute. The registry describes it as partially randomized, with a randomized parallel-group phase II comparison and no masking. The overall enrollment is recorded as 116; that figure should not be treated as the analysis population for any individual endpoint, because the posted phase-specific results use endpoint-specific populations.

Phase I examined bevacizumab with TRC105 at dose levels identified as Dose 0, Dose 1, and Dose 2. The posted primary endpoint counts are 0 participants with dose-limiting toxicity among 3 at Dose 0, 0 among 3 at Dose 1, and 1 among 6 at Dose 2. The record describes the maximum tolerated dose definition in terms of toxicity at a dose level and at the next higher dose level. These cohort counts describe dose-finding observations; they are not a randomized comparison of the combination against bevacizumab alone.

In phase II, Arm I received bevacizumab plus TRC105 and Arm II received bevacizumab alone. The treatment descriptions specify intravenous administration and courses repeated every 14 days in the absence of disease progression or unacceptable toxicity. The phase II overall-survival results list 52 participants in Arm I and 49 in Arm II. The safety results use different numbers at risk for the corresponding phase II arms, so the endpoint populations should be read separately rather than assumed to be identical.

The registry lists 11 outcome measures and 16 analyses, including overall survival, progression-free survival at 6 months, toxicity, and several quality-of-life assessments. The record does not report the protocol’s statistical analysis plan or a detailed explanation of how phase I enrollment flowed into the randomized phase II portion. Those design details should not be inferred from the posted summary.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All phase I patients who completed the study were eligible and analyzed for MTD.

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 Maximum Tolerated Dose (MTD) (Phase I) as Measured by the Number of Participants With Dose Limiting Toxicities, measured over 28 days.

The primary endpoint is maximum tolerated dose in phase I, measured by the number of participants with dose-limiting toxicities during the stated 28-day timeframe. The registry defines the maximum tolerated dose as the highest safely tolerated dose level where at most 1 out of 6 patients experience dose-limiting toxicity, with at least 2 patients out of a maximum of 6 experiencing dose-limiting toxicity at the next higher dose. The registry further states that 6 patients treated at the maximum tolerated dose will be sufficient to identify common toxicities at that dose.

This endpoint is used in dose-finding because early studies need to characterize tolerability before a dose is taken forward for further evaluation. The count of participants with dose-limiting toxicity is a direct description of the observed event burden in each listed cohort. It is not a measure of comparative treatment efficacy, and it does not by itself describe every possible adverse effect or longer-term safety concern.

The posted counts are 0 of 3 at Dose 0, 0 of 3 at Dose 1, and 1 of 6 at Dose 2. These observations need to be interpreted in light of the registry’s definition, the cohort sizes, and the dose levels actually reported. The record identifies phase I patients who completed the study as eligible and analyzed for maximum tolerated dose. Details about the handling of participants who did not complete the study are not reported in the registry record.

The endpoint is distinct from the phase II secondary outcomes. In particular, a phase I dose-finding count does not answer whether adding TRC105 changes survival, progression, quality of life, or overall toxicity relative to bevacizumab alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the phase I count measure?

The primary endpoint is a count of participants with dose-limiting toxicity during the 28-day period. A count is a direct tally, rather than a model-based estimate of a treatment effect. The registry specifies a dose-selection rule involving both the current dose and the next higher dose. The posted counts are 0 of 3 at Dose 0, 0 of 3 at Dose 1, and 1 of 6 at Dose 2. Because the cohorts are small, each participant’s event status has substantial influence on the observed count. General statistical context: dose-finding counts are useful for describing what occurred in the enrolled cohorts, but their precision is limited when only a small number of participants are observed.

How should the survival summaries be read?

Overall survival is defined as time from the start of study therapy to death from any cause, assessed up to 2 years. The posted median is 9.7 months for bevacizumab plus TRC105 and 7.4 months for bevacizumab alone. The reported confidence intervals are 6.7 to 11.5 months and 6.5 to 12.7 months, respectively. A median gives a central summary of the observed time-to-event distribution; it does not describe the entire curve or all variation in survival times.

The comparison is reported as a hazard ratio of 1.06, with a 95% confidence interval of 0.66 to 1.69, and p = 0.82. The method is identified as log-rank. General statistical context: a hazard ratio compares event rates over time under the assumptions of the analysis; it is not a ratio of medians. A confidence interval conveys uncertainty in the estimate, while a p-value assesses compatibility of the data with a null comparison. The confidence interval’s inclusion of 1 and the reported p-value do not support a precise claim of a difference, but they also do not demonstrate equivalence.

What do the planning illustrations mean?

The registry’s sample-size illustrations are based on the observed effect and are not the trial’s own design justification. They assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The stated effect to detect is a specified fraction of the observed effect. For time-to-event calculations, the displayed quantities are total events, using the Schoenfeld approach; they are not participants per arm and are not observed power.

For planning at 100% of the observed effect, the illustration gives 9,247 total events for 80% power and 12,379 total events for 90% power. At 75% of the observed effect, it gives 16,439 and 22,007 total events, respectively. At 50% of the observed effect, it gives 36,988 and 49,516 total events, respectively. These values are planning illustrations derived from the observed estimate, not evidence that the trial enrolled or observed those numbers of events. The registry record does not report the original sample-size justification.

What can be inferred from the unmasked design?

The registry identifies the trial as unmasked. General statistical context: when participants or study personnel know treatment assignment, expectations or treatment decisions can influence some outcomes, particularly subjective assessments or care processes. Overall survival is a comparatively objective endpoint, but its analysis can still depend on follow-up and endpoint handling. Quality-of-life outcomes can be more sensitive to assessment and response processes. The record does not describe procedures intended to limit these influences.

How should different analysis populations be treated?

The overall enrollment is 116, while the posted phase II survival analysis lists 52 participants in the combination arm and 49 in the bevacizumab-alone arm. The phase II safety summaries use 49 and 43 participants at risk, respectively. These endpoint-specific denominators differ, and should not be replaced with the overall enrollment figure. The record says phase I patients who completed the study were eligible and analyzed for maximum tolerated dose. It does not report a detailed analysis-population definition for the phase II survival result, censoring rules, missing-data procedures, or the number of participants excluded from each analysis.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Phase I: Bev + TRC105 (Dose 0, Cohort A)", "Phase I: Bev + TRC105 (Dose 1, Cohort A)", "Phase I: Bev + TRC105 (Dose 2, Cohort A + B)")
n      <- c(3, 3, 6)
events <- c(0, 0, 1)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Phase I: Bev + TRC105 (Dose 0, Cohort A)  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Phase I: Bev + TRC105 (Dose 1, Cohort A)  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Phase I: Bev + TRC105 (Dose 2, Cohort A + B)  6  1  16.7  0.4  64.1  3.0  56.4  3.7  57.9

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 11 outcome measures and 16 analyses, along with 6 posted secondary outcomes. It does not report a multiplicity strategy, such as an adjustment of significance thresholds, a prespecified testing order, or a rule for interpreting the collection of secondary analyses. Without that information, individual secondary p-values, if reported elsewhere in the results, should not automatically be treated as independent confirmatory tests. General statistical context: examining many outcomes can increase the chance that at least one result appears noteworthy even when there is no underlying effect.

Missing-data methods are not reported in the registry record. This matters for outcomes such as survival and quality of life in different ways. Survival analysis commonly needs rules for censoring participants whose event time is not observed during follow-up; the registry summary does not describe those rules. Questionnaire outcomes can have incomplete responses, but the record does not explain whether incomplete forms were excluded, imputed, or handled by another method. No method should be assumed.

The analysis populations are not fully described across endpoints. For phase I maximum tolerated dose, the record states that all phase I patients who completed the study were eligible and analyzed. For the phase II survival result, the arm counts are 52 and 49. For phase II safety, the numbers at risk are 49 and 43. The record does not explain these differences or provide a complete participant flow for each analysis. The overall enrollment of 116 is not a substitute for the endpoint-specific analysis populations.

The registry describes the trial as partially randomized. The phase II arms are presented as a randomized comparison, while the phase I cohorts are dose-finding groups. The record does not report the allocation procedure in further detail, the statistical analysis plan, or whether any interim analyses were planned or conducted. These items are not reported in the registry record and should not be inferred from the posted results.

12. 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: anemia (33.3%), hydrocephalus (33.3%), epistaxis (33.3%), thromboembolic event (33.3%), nausea (25.0%).

The registry separates serious adverse events, other adverse events, and deaths in its safety summaries. For phase II, the serious-adverse-event figures list 15 affected among 49 at risk in the bevacizumab-plus-TRC105 arm and 8 among 43 at risk in the bevacizumab-alone arm. The death counts are 2 in each arm. These are descriptive counts; the record does not provide a randomized comparative estimate for these safety outcomes in the information.

In phase I, the serious-event counts are listed by dose cohort: 3 affected among 4 at risk at Dose 0, 0 among 3 at Dose 1, 1 among 4 in Dose 2 Cohort A, and 2 among 3 in Dose 2 Cohort B. Other adverse events are also reported by cohort. These small cohort summaries should not be read as precise comparative safety estimates.

The registry lists anemia, hydrocephalus, epistaxis, and thromboembolic event at 33.3% each among the top serious-event terms, and nausea and vomiting at 25.0% each. The term-specific denominators and further details are not reported in the registry record. A serious adverse event is not necessarily judged related to treatment; the displayed serious-event counts should not be conflated with the separate secondary outcome concerning grade 3 or higher events considered at least possibly treatment-related.

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

The posted results describe two different statistical tasks. In phase I, the study recorded dose-limiting toxicities across TRC105 dose cohorts given with bevacizumab. In phase II, it compared overall survival for bevacizumab plus TRC105 against bevacizumab alone. The former is a dose-finding assessment; the latter is a randomized time-to-event comparison. Neither should be used as a substitute for the other.

The phase I counts show 0 participants with dose-limiting toxicity at Dose 0, 0 at Dose 1, and 1 at Dose 2. These observations are limited by the cohort sizes and by the fact that toxicity counts cannot characterize all possible adverse events. The registry’s maximum tolerated dose definition also refers to the next higher dose level. The posted summary does not report every detail needed to independently reconstruct the dose-selection decision.

The phase II overall-survival estimate is a hazard ratio of 1.06, with a 95% confidence interval of 0.66 to 1.69 and p = 0.82. The median survival estimates are 9.7 months for the combination arm and 7.4 months for bevacizumab alone. These data do not establish that one treatment is better, worse, or equivalent. A nonsignificant p-value is not proof of no effect, and a confidence interval spanning 1 indicates that the estimate is compatible with different directions of the relative hazard under the analysis assumptions.

The findings also do not establish the result for every secondary outcome. The record lists toxicity, progression-free survival, and quality-of-life outcomes, but the reported information does not support a complete comparative interpretation across all of them. Nor do the results establish what would happen in people outside the trial population or under different treatment conditions.

General statistical context: clinical trial estimates are evidence about outcomes in a defined study setting, subject to uncertainty and the quality of measurement and follow-up. They are not individualized predictions. The registry record does not report details such as missing-data procedures, the full analysis plan, or the basis for the original sample-size decision. Those omissions limit independent assessment of how robustly the reported analyses address the trial questions.

14. Why This Trial Matters Statistically

This trial is statistically interesting because it combines early dose finding with a randomized comparative phase. The phase I endpoint is a count-based safety criterion, while phase II uses time-to-event analysis for overall survival. Recognizing this shift is essential: a dose cohort is not equivalent to a randomized treatment arm, and a maximum tolerated dose is not itself an estimate of comparative benefit.

The posted survival result illustrates why effect estimates should be read with their uncertainty. The hazard ratio is 1.06, and its 95% confidence interval extends from 0.66 to 1.69. The p-value is 0.82. The estimate does not provide a precise distinction between the treatment strategies, and the reported statistics do not establish equivalence. Considering only the median values, or only the point estimate, would omit important information about uncertainty.

The record also illustrates the importance of endpoint-specific denominators. The overall enrollment is 116, but the phase II survival and safety results list different numbers at risk in their respective arms. Those differences matter for interpretation and cannot be resolved by substituting overall enrollment for an analysis-specific population. The registry summary does not explain the denominator differences.

Finally, the secondary outcomes and planning illustrations need separate treatment. Outcomes such as quality of life and progression-free survival may address meaningful questions, but multiplicity and incomplete reporting affect how strongly they can support conclusions. The sample-size illustrations are calculations based on the observed effect under stated assumptions; they are not the trial’s design justification or observed power. Keeping these distinctions clear helps prevent descriptive results, planning scenarios, and confirmatory evidence from being conflated.

15. 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 phase I cohorts are small, with posted group sizes of 3, 3, and 6 for the primary dose-limiting toxicity result. Counts in cohorts of this size provide limited precision, particularly for uncommon or delayed toxicities. The registry’s dose-selection definition relies on observations at a dose and the next higher dose; the posted count summary does not by itself explain the complete dose-selection process.

The phase II survival comparison has a confidence interval that spans values below and above 1 for the hazard ratio. This leaves substantial uncertainty about the relative hazard. The record describes the trial as partially randomized and unmasked. It does not report further detail about allocation procedures, masking safeguards, or the statistical analysis plan.

There are differences between the arm counts listed for phase II overall survival and those at risk in the safety summaries. The registry record does not explain these differences or provide full participant flow by endpoint. Missing-data handling, censoring rules, and the analysis population definitions beyond the phase I completion criterion are not reported in the registry record.

The record lists multiple secondary outcomes, but does not report a multiplicity strategy. It also notes that binary endpoint values are missing or not estimable for a two-arm comparison. Accordingly, the posted information does not support a comprehensive assessment of every registered outcome. Quality-of-life instruments are named, but detailed comparative results and response-handling methods are not reported in the information.

The sample-size illustrations are based on the observed effect, not on the trial’s own design justification. They should not be treated as evidence about the trial’s power or as a retrospective validation of its design. The original sample-size rationale and any interim-analysis plans are not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Bevacizumab With or Without Anti-Endoglin Monoclonal Antibody TRC105 in Treating Patients With Recurrent Glioblastoma Multiforme — NCT01648348.
• Registry record last updated 2018-05-23; status last verified 2018-04.
• Results first posted 2018-05-23.
• 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.

19. Record Summary

NCT01648348 is a completed phase I/phase II study of bevacizumab with or without TRC105 in recurrent glioblastoma. It is described as partially randomized, unmasked, and parallel-group, with overall enrollment of 116.

The phase I primary endpoint was maximum tolerated dose, measured by participants with dose-limiting toxicities over 28 days. The posted counts are 0 of 3 at Dose 0, 0 of 3 at Dose 1, and 1 of 6 at Dose 2. Phase II compared bevacizumab plus TRC105 with bevacizumab alone. Overall survival was reported as median time, with a hazard ratio of 1.06, 95% confidence interval of 0.66 to 1.69, and p = 0.82 by log-rank analysis.

The registry also lists toxicity, progression-free survival, and quality-of-life outcomes. Missing-data handling, multiplicity methods, the full phase II analysis-population definitions, and the original sample-size justification are not reported in the registry record.