QUILT-3.011 Phase 2 Yeast-Brachyury Vaccine Chordoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for QUILT-3.011 Phase 2 Yeast-Brachyury Vaccine Chordoma (NCT02383498), a randomized phase 2 study. The primary outcome is Overall Response Rate.

2. Clinical Question

The central estimand is a difference in the proportion of participants with an overall response. Complete response was defined as disappearance of all target lesions and no new lesions; partial response was defined as at least a 30% decrease in the sum of the greatest diameter. These definitions specify which tumor changes counted toward the primary endpoint.

The registry describes random assignment and a double-blind phase 2 trial; its masking field also records quadruple masking. The recorded comparison is between two groups receiving radiotherapy, one with vaccine and the other with placebo. Thus, the primary question is the incremental effect associated with adding GI-6301 under the trial conditions. It is not a comparison of radiotherapy with another radiation strategy, and it does not directly answer questions about other chordoma populations or treatments.

3. Trial Design

This was a randomized, double-blind, phase 2 trial of GI-6301 plus definitive radiotherapy versus placebo plus definitive radiotherapy in chordoma. The registry records a crossover model and quadruple masking. It does not identify which roles were masked, or report the sequence, timing, or operational details of any crossover. Those details are not reported in the registry record.

The registry lists 24 participants enrolled. The primary outcome results are presented for 11 participants in the GI-6301 Vaccine arm and 13 in the Placebo arm. Both arms also include an Actigraph device in their descriptions. The enrolled population was described as patients with inoperable or residual chordoma, no metastases, and planned definitive radiotherapy above 70Gy. The brief summary describes the disease as localized chordoma, while the official title specifies locally advanced, unresectable chordoma.

The study status is completed. Its recorded start was 2015-05-01 and its primary completion was 2019-09-01. The primary endpoint was overall response rate at the irradiated tumor site over up to 24 months, defined using RECIST 1.1 complete or partial response criteria.

Enrollment and outcome group sizes are small, and the registry record lists one outcome measure and no secondary outcome measures. It records no analysis entries. The details of the original sample-size justification, trial-specific statistical analysis plan, randomization implementation, crossover procedures, and analysis population are not reported in the registry record. The design labels alone do not establish how these elements were carried out.

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

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 Response Rate, measured over Up to 24 months.

The primary endpoint was overall response rate, expressed as the count of participants who had a complete or partial response at the irradiated tumor site over up to 24 months. The registry defines complete response as disappearance of all target lesions with no new lesions, and partial response as at least a 30% decrease in the sum of the greatest diameter. The response criteria are attributed to RECIST 1.1.

Overall response rate reduces tumor-response assessments to whether a participant met either response definition. It is used to summarize tumor shrinkage in a defined period and provides a direct way to compare the proportion responding in the two randomized groups. In this trial, the endpoint addresses the question of whether response differed when GI-6301 was added to radiotherapy, compared with radiotherapy plus placebo.

The measure does not, by itself, describe how long a response lasted, whether tumor growth was prevented after the assessment period, or other clinical outcomes. Those outcomes are not reported as secondary measures in the registry record. The registry also does not report further details about assessment timing, independent review, or how incomplete assessments were handled. The response count therefore needs to be read alongside the limited information available about endpoint ascertainment and analysis.

7. Results Posted in the Registry

• GI-6301 Vaccine: 9.1% (1/11) versus 7.7% (1/13) on control; risk difference 1.4 percentage points (95% CI -20.9 to 23.7), risk ratio 1.18 (0.08–16.78); unadjusted two-proportion z-test p = 0.902.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the risk difference represent?

The risk difference compares the proportion of participants with the endpoint in one group with the proportion in the other. Here it is reported as 0.014 for GI-6301 Vaccine relative to Placebo. A value of zero represents equal observed response proportions. Positive values indicate a higher proportion in the vaccine arm under that orientation of the comparison; negative values indicate a higher proportion in the placebo arm. The confidence interval communicates uncertainty around the estimated difference. It includes zero and extends in both directions, so the posted estimate does not precisely distinguish a difference favoring one arm from a difference favoring the other.

How do relative measures differ?

The risk ratio compares the response proportions as a ratio, while the odds ratio compares the odds of response. A ratio of one represents no relative difference. The reported risk ratio is 1.182 and the odds ratio is 1.2, but their confidence intervals are very wide: 0.083 to 16.78 for the risk ratio and 0.066 to 21.72 for the odds ratio. These ranges show that the relative effects are poorly determined by the observed counts. With few outcomes, a change in a small number of participants can substantially affect a relative measure. The risk difference offers an absolute scale, while the ratio measures describe relative association; neither scale removes the uncertainty in these data.

What does a p-value say?

The reported unadjusted p-value is 0.902. In general, a p-value describes how unusual the observed data, or data more extreme, would be under a specified null model. It is not the probability that the null hypothesis is true, nor is it a measure of the size or importance of an effect. A large p-value does not prove that the groups are equivalent. In this setting, the sparse responses and wide confidence intervals mean that the lack of a small p-value should be interpreted together with the uncertainty in the effect estimates.

Why does sparse binary data matter?

General statistical context: a binary endpoint records whether each participant did or did not have an event, here a complete or partial response. When events are sparse, estimates of proportions and comparative effects can be unstable. There is little outcome information to distinguish possible underlying response rates, and relative measures can become particularly uncertain. The counts of 1 response in each arm provide the essential context for interpreting the point estimates and intervals.

What is over-dispersion?

General statistical context: over-dispersion occurs when outcome variability is greater than a model assumes. It can make standard errors too small if the model does not account for the extra variability. The registry record does not report a model for the primary binary endpoint or any assessment of over-dispersion. Because the posted result is a sparse binary comparison, it is more important not to imply a model-based adjustment that is not documented than to assume a particular variance structure.

What analysis details are recorded?

The trial record gives arm-level counts, a risk difference, relative effect estimates, confidence intervals, and an unadjusted p-value. It lists no analysis entries. The specific test or estimation procedure, any covariate adjustment, any stratification, and any treatment of the crossover model in the primary analysis are not reported in the registry record. The numbers should therefore be read as the posted unadjusted comparison rather than attributed to an undocumented statistical model.

General statistical context: an analysis plan ordinarily specifies how the primary endpoint is calculated and compared, which participants are included, how missing outcomes are handled, and whether any adjustments are made. Such choices can affect the interpretation of a result. The available record does not establish those choices for this trial, so no further claims about the analysis method are warranted.

Why distinguish statistical significance from precision?

A p-value and a confidence interval answer related but different questions. A p-value measures compatibility with a null model under specified assumptions; an interval conveys a range of effect values compatible with the data and method at a stated confidence level. Here, the p-value is 0.902 and the risk-difference interval ranges from -0.209 to 0.237. The interval makes the limited precision visible: it allows effects in either direction, rather than locating the result tightly around the point estimate. Neither statistic alone determines whether an effect would matter clinically, and the registry record does not report a threshold for a clinically important response-rate difference.

10. Confidence Intervals and Precision

The risk-difference estimate is 0.014, with a 95% confidence interval from -0.209 to 0.237. The interval crosses zero and extends in both directions, so the data do not precisely determine whether the response proportion was higher with GI-6301 or with placebo. Its width indicates substantial uncertainty around the estimate. The interval should be read as a range of effect values compatible with the data and the method used, rather than as a guarantee that the true effect lies within it.

The relative estimates are also imprecise. The risk ratio is 1.182, with a confidence interval from 0.083 to 16.78; the odds ratio is 1.2, with a confidence interval from 0.066 to 21.72. These broad ranges reflect how little information is available when only 1 response was recorded in each arm. Relative estimates are particularly difficult to pin down with sparse events, and the intervals allow widely differing relative associations.

General statistical context: confidence intervals describe uncertainty in an estimate under the assumptions of the calculation. They do not show the probability that a particular effect is true, and a wide interval is not evidence that every value inside it is equally plausible. The combination of a small point estimate, broad intervals, and few responses means that the trial result is not precise enough to support a narrow conclusion about the magnitude of the response-rate difference. Precision is distinct from whether a result crosses a conventional statistical threshold; both the interval and the underlying counts matter to interpretation.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size display gives planning illustrations in participants per arm, not total events. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect used for each illustration is a stated share of the observed effect. The trial's own original sample-size assumptions and justification are not reported in the registry record; these planning calculations are not a report of the trial's design basis.

The 80% power column gives the illustrated number of participants per arm for a design with an 80% chance of detecting the chosen effect if that effect is the true one, under the listed assumptions. The 90% power column gives the corresponding planning number for a 90% chance. These are prospective planning interpretations, not certainty that a study will detect an effect or that an effect exists. The listed values use the observed effect as their starting point, so they depend on an estimate that is itself highly uncertain in this trial.

The entries for planning for 100%, 75%, and 50% of the observed effect illustrate that smaller effects need more participants to detect under otherwise stated assumptions. This is a general feature of power calculations: weaker signals are harder to distinguish from sampling variation. The sample-size display is not observed power. It should not be read as a measure of the trial's success, nor as evidence that a future study with a particular enrollment would reproduce the observed result. No sample-size curve-specific assumptions beyond the listed planning assumptions are reported.

12. How the Sample Size Is Derived

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

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.077; treatment proportion p₁ = 0.091; average p̄ = 0.084.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3921 = 0.7685.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.3920 = 0.3299.
4. n = (0.7685 + 0.3299)² / (0.0140)² = 6,168.0, rounded up to 6,169 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.0769230769230769                          # control proportion
p1 <- 0.0909090909090909                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(2383498)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 6,169 (80% power), 8,257 (90% power)
 75% of observed effect: n per arm = 10,759 (80% power), 14,402 (90% power)
 50% of observed effect: n per arm = 23,738 (80% power), 31,778 (90% power)
power.prop.test: 6168 
Simulated power at n = 6,169 per arm: 0.803 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and no secondary outcomes. It also records no analysis entries. This limited outcome listing reduces what can be assessed about the range of planned comparisons, but it does not establish how many analyses were conducted or whether any multiplicity adjustment was planned or used. The plan for handling multiple looks, endpoints, or comparisons is not reported in the registry record. Interim analyses are not reported in the registry record.

Multiplicity matters when several outcomes or comparisons are tested: the chance of at least one false-positive result can rise as the number of tests increases. General statistical context: prespecified testing strategies or adjustment procedures can address that issue, but their use cannot be assumed. The registry record does not describe any such strategy. The reported primary comparison is identified as an unadjusted result, and its p-value should be interpreted as such rather than as evidence about an undocumented multiplicity plan.

The record gives arm-level response counts but does not report a participant flow, the analysis population definition, or whether all randomized participants were included in the primary comparison. The published arm sizes for the outcome are 11 and 13, matching the displayed arm counts, but that alone does not clarify all analysis-population rules. Missing outcome handling, imputation, exclusion criteria after randomization, and sensitivity analyses are not reported in the registry record. It is therefore not possible to determine from these entries whether there were incomplete assessments or how any were treated.

The crossover model is recorded, but the crossover sequence, timing, and its role in the primary endpoint analysis are not described. The response assessment is tied to the irradiated tumor site over up to 24 months; the precise assessment schedule and handling of observations after a crossover are not reported. These omissions limit the ability to reconstruct the analysis or judge whether the estimate follows a specific crossover estimand. The absence of a description should not be taken to mean that no such procedures occurred; it means they are not documented in the registry record.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: vomiting (9.1%), abdominal pain (9.1%), diarrhea (9.1%), lower gastrointestinal hemorrhage (9.1%), nausea (9.1%).

The registry reports serious adverse events for 5 of 11 participants in the GI-6301 Vaccine arm and 1 of 13 in the Placebo arm. It reports other adverse events for 11 of 11 participants in the vaccine arm and 13 of 13 in the placebo arm. Deaths are recorded as 0 in each arm. These are participant counts, not a comparison adjusted for differences between groups.

The listed serious adverse-event terms are Vomiting, Abdominal pain, Diarrhea, Lower gastrointestinal hemorrhage, Nausea, and Urinary tract infection, each reported as 9.1%. The registry extract does not identify an arm for each listed term or provide further event-level detail in these entries. The percentages should therefore be read as the registry lists them, without assuming a denominator beyond what is stated for the adverse-event group counts.

General statistical context: adverse-event summaries are descriptive, and small group sizes can make comparisons unstable. Counts alone do not establish whether events were caused by a study treatment, how severe or prolonged they were, or whether participants experienced more than one event. Further information about event timing, severity grading, relatedness, and statistical comparisons is not reported in the registry record. These figures are not a basis for an individual medical decision.

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

The posted primary result is a comparison of overall response in two small randomized groups receiving radiotherapy, with GI-6301 in one group and placebo in the other. There was 1 response among 11 participants in the vaccine arm and 1 among 13 participants in the placebo arm. The estimated risk difference was 0.014, but its 95% confidence interval, -0.209 to 0.237, is broad and includes effects in either direction. The unadjusted p-value was 0.902.

These data do not precisely establish the size or direction of a difference in response. They also do not show that the groups are equivalent: an absence of strong evidence for a difference is not proof that no meaningful difference exists. The broad interval and sparse outcome counts mean that the result is compatible with a wide range of possible underlying effects. The risk-ratio and odds-ratio intervals are also very wide, reinforcing the uncertainty.

The endpoint captures complete or partial response at the irradiated tumor site, as defined by RECIST 1.1, over up to 24 months. It does not by itself describe duration of response, later tumor behavior, survival, quality of life, or other outcomes. Secondary outcome measures are not listed in the registry record, so those questions cannot be answered from the posted primary results.

The comparison applies to the enrolled trial population and the studied combination of radiotherapy with vaccine or placebo. It should not be treated as a result for other chordoma populations, treatment settings, or interventions. The registry's crossover designation, without details of crossover timing or analysis, also limits reconstruction of the design's role in the primary comparison. Overall, the result is best characterized as an imprecise estimate that leaves substantial uncertainty, not a definitive conclusion about the presence or absence of a response-rate difference.

16. Why This Trial Matters Statistically

This trial illustrates how a randomized comparison can still yield an uncertain estimate when the outcome is rare in the enrolled sample. The response counts were 1 in each arm, leaving little information to distinguish the groups. Randomization supports a structured comparison, but it does not guarantee precision; precision also depends on how many participants and outcomes contribute information.

The risk difference gives an absolute comparison, while risk ratio and odds ratio express relative comparisons. In this trial, the point estimates on all three scales are accompanied by broad confidence intervals. Seeing those intervals alongside the event counts is statistically important: a point estimate can appear specific while the underlying data leave substantial uncertainty about plausible effects. The unadjusted p-value alone cannot convey that range or establish equivalence.

The design also emphasizes the importance of transparent reporting. The registry records a crossover model and quadruple masking, but does not describe the crossover procedures or which roles were masked. It does not report the analysis population, missing-data approach, detailed statistical plan, or original sample-size justification. These omissions limit the ability to reproduce the analysis and assess how design features connect to the reported estimate.

Finally, planning calculations based on the observed effect are illustrations, not evidence of what the trial was powered to detect and not a substitute for a prespecified design justification. Because the observed effect is uncertain, planning around it inherits that uncertainty. The statistical lesson is to read counts, effect estimates, intervals, design details, and reporting gaps together. None alone provides a complete account of the evidence.

17. 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 principal limitation of the primary comparison is the small number of participants and responses: 11 participants were reported in the GI-6301 Vaccine arm and 13 in the Placebo arm, with 1 response in each. The resulting confidence intervals are broad, limiting precision. The registry record does not report a clinically important difference threshold, so the statistical estimates cannot be judged against a stated clinical threshold from these entries.

The registry records a crossover design but does not report the crossover schedule, sequence, eligibility for crossover, or how any crossover was incorporated into the primary comparison. It is therefore not possible to determine the precise estimand behind the posted arm-level result. Masking is recorded as quadruple, while the official title says double-blind; the roles masked and the relationship between these descriptions are not reported.

The registry does not report the original sample-size justification, the detailed statistical analysis plan, interim monitoring, the analysis population, missing-data procedures, or whether endpoint assessments were independently reviewed. No analysis entries or secondary outcome measures are listed. These omissions limit evaluation of prespecification, analytic choices, and the handling of incomplete or post-randomization data.

For safety, the adverse-event summaries are counts with limited detail in the record. The listed serious-event terms are not linked to a particular arm in the available summary, and event timing, severity, relatedness, and participant-level overlap are not described. Overall, the results are informative as a record of the posted counts and estimates, but incomplete reporting and sparse outcomes constrain deeper assessment.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: QUILT-3.011 Phase 2 Yeast-Brachyury Vaccine Chordoma — NCT02383498.
• Registry record last updated 2024-05-20; status last verified 2019-10.
• Results first posted 2024-05-20.
• 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.

21. Record Summary

NCT02383498 is a completed randomized phase 2 trial of GI-6301, a yeast-brachyury vaccine, plus definitive radiotherapy versus placebo plus radiotherapy for chordoma. The registry lists a crossover model, double-blind design, and quadruple masking. Total enrollment was 24, with primary outcome counts presented for 11 participants in the vaccine arm and 13 in the placebo arm.

The primary endpoint was overall response rate at the irradiated tumor site over up to 24 months, defined as complete or partial response by RECIST 1.1. One participant in each arm had a response. The reported risk difference was 0.014, with a 95% confidence interval from -0.209 to 0.237 and an unadjusted p-value of 0.902. These estimates are imprecise and do not resolve the direction or size of a difference.

The record lists serious adverse events among 5 of 11 vaccine-arm participants and 1 of 13 placebo-arm participants, and reports no deaths in either arm. Detailed analysis-plan, crossover, missing-data, and sample-size-justification information is not reported in the registry record.