Pelvic Radiation Therapy or Vaginal Implant Radiation Therapy, Paclitaxel, and Carboplatin in Treating Patients With High-Risk Stage I or Stage II Endometrial Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pelvic Radiation Therapy or Vaginal Implant Radiation Therapy, Paclitaxel, and Carboplatin in Treating Patients With High-Risk Stage I or Stage II Endometrial Cancer (NCT00807768), a randomized phase 3 study. The primary outcome is Number of Participants With Recurrence or Death Events at Primary Analysis.

2. Clinical Question

The central question is whether the strategies differ in the observed occurrence of recurrence or death among participants with high-risk early-stage endometrial cancer. One strategy uses pelvic radiation therapy; the other combines vaginal cuff brachytherapy with paclitaxel and carboplatin. Because these are treatment strategies with different components, the comparison is between the assigned regimens as a whole. The posted figures do not isolate the effect of radiation technique, brachytherapy, or either chemotherapy drug.

The trial is described as randomized, phase III, parallel-group, and unmasked. The registry record reports 601 enrolled participants, divided between Arm I (Pelvic Radiation Therapy) and Arm II (Brachytherapy, Paclitaxel, Carboplatin). The primary outcome is a participant count, not a reported median time to recurrence, hazard ratio, or survival curve. Interpretation should therefore stay anchored to the event counts and the posted comparison estimates rather than treating them as a full account of the timing of events.

The registry's stated observation schedule includes assessment within 4 weeks after completing or discontinuing study therapy, then every 6 months for 2 years, then annually for 3 years, and then as clinically indicated, assessed up to 10 years. The primary outcome description defines recurrence-free survival from study entry through recurrence, death, or last contact. Details on how the primary analysis handled follow-up time, censoring, and incomplete outcome ascertainment are not reported in the registry record.

3. Trial Design

This is a randomized phase III trial with parallel groups and no masking. The registry names the Gynecologic Oncology Group as sponsor. Enrollment was 601. The start date is 2009-03-23, and the primary completion date is 2014-12-29. The current status is listed as UNKNOWN, so the registry record does not establish the present administrative status of the study.

Arm I received pelvic radiation therapy. The description specifies conventional or intensity-modulated pelvic radiation once daily, 5 days a week, for 5-6 weeks, totaling 25-28 fractions in the absence of disease progression or unacceptable toxicity. It also refers to treatment considerations for some participants with stage II disease or stage I disease with a confirmed diagnosis of clear cell, but the rest of that sentence is not reported in the registry record.

Arm II received vaginal cuff brachytherapy followed by paclitaxel and carboplatin. The registry describes 3-5 high-dose-rate brachytherapy treatments over approximately 2 weeks, or 1 or 2 low-dose-rate treatments over 1-2 days. It states that chemotherapy began within 3 weeks after initiating brachytherapy. The arm description is incomplete after mentioning paclitaxel administered intravenously over 3 hours and carboplatin; further regimen details are not reported in the registry record.

The primary result includes 301.0 participants in Arm I and 300.0 in Arm II. The outcome is the number of participants with recurrence or death events at primary analysis, and the outcome population is identified as all enrolled patients. The registry also lists seven outcome measures and five posted secondary outcomes. No analysis records are listed for the primary comparison, and no reported analyses are listed for the posted secondary result. The protocol's original sample-size justification, allocation procedure details, and any planned interim analyses are not reported in the registry record.

The trial's comparison is open label: masking is listed as none. Random assignment can support a group comparison by helping balance participant characteristics at the point of assignment, but the registry record does not report details of sequence generation, allocation concealment, or any stratification factors. The posted information describes the assigned treatment strategies but does not explain whether adherence, crossovers, or treatment discontinuation were incorporated into the primary analysis.

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

Analysis population for the primary outcome, as defined in the registry: All enrolled patients

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Enrolled patients

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Recurrence or Death Events at Primary Analysis, measured over Within 4 weeks after completing or discontinuing study therapy, then every 6 months for 2 years, then annually for 3 years, then as clinically indicated, assessed up to 10 years.

The primary outcome is the number of participants with recurrence or death events at primary analysis. The registry describes recurrence-free survival as the period from study entry until disease recurrence, death, or the date of last contact. This definition makes recurrence and death events relevant to the endpoint, while last contact identifies the end of recorded follow-up for participants without a recorded event.

Recurrence-free survival is used in cancer trials to capture disease recurrence and death within a single outcome. As a general statistical concept, combining these events can summarize clinically important failures of remaining alive and recurrence-free, rather than considering recurrence alone. It does not mean the two event types have the same clinical meaning, and the registry separately lists death events among secondary outcomes.

The posted primary results are counts of participants with events: 44.0 in the pelvic-radiation group and 43.0 in the brachytherapy-and-chemotherapy group. The comparison reports a risk difference, with confidence limits and an unadjusted p-value. These are binary-outcome summaries of whether a participant had an event, rather than reported estimates of time until an event. The registry outcome name includes “recurrence-free survival,” but a time-to-event estimate such as a hazard ratio, a survival curve, or a median event-free time is not reported in the registry record.

The stated follow-up schedule extends assessment up to 10 years, but the posted event counts alone do not convey when events occurred or how much follow-up each participant contributed. Details of censoring and the exact primary analysis method are not reported in the registry record. Those omissions matter when interpreting a survival-labelled outcome from a result presented as a participant count.

7. Results Posted in the Registry

• Arm II (Brachytherapy, Paclitaxel, Carboplatin): 14.3% (43/300) versus 14.6% (44/301) on control; risk difference -0.3 percentage points (95% CI -5.9 to 5.3), risk ratio 0.98 (0.66–1.45); unadjusted two-proportion z-test p = 0.921.

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.

Number of Participants With Death Events

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Arm II (Brachytherapy, Paclitaxel, Carboplatin): 12.3% (37/300) versus 13.0% (39/301) on control; risk difference -0.6 percentage points (95% CI -5.9 to 4.7), risk ratio 0.95 (0.63–1.45); unadjusted two-proportion z-test p = 0.818.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm II (Brachytherapy, Paclitaxel, Carboplatin) and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes extend the study's questions beyond the primary recurrence-or-death count. They include death events, sites of recurrence, patient-reported fatigue, patient-reported neurotoxicity, and patient-reported quality of life. Together, these outcomes cover survival, disease recurrence, symptoms, and quality of life. They may add context to the primary result, but they address distinct aspects of participants' outcomes and do not replace the primary endpoint.

The posted death-event comparison has a risk-difference interval that includes zero, in keeping with the broad uncertainty seen in the primary result. The registry lists the other secondary outcome categories, but no reported analyses are listed for the posted secondary outcomes. The record therefore does not establish a consistent pattern across all secondary domains. Without reported analyses for each outcome, it is not possible to determine whether apparent differences, if any, reflect reproducible effects or sampling variation.

Secondary outcomes can be valuable because a treatment strategy may influence outcomes beyond the primary measure. However, each additional outcome creates another opportunity for a chance finding. When many outcomes are tested, the chance of at least one apparently unusual result can increase unless the analysis plan addresses multiplicity. The registry record does not report a multiplicity adjustment or a hierarchy specifying which secondary questions were confirmatory.

Patient-reported outcomes also require attention to missing responses, assessment schedules, and the interpretation of scores. The registry identifies fatigue, neurotoxicity, and quality of life as secondary outcome areas but does not report detailed instruments, scoring rules, missing-response procedures, or analyses. These measures should not be interpreted as demonstrating benefit or harm without those details and the corresponding results.

The secondary evidence should therefore be read as supplementary to the primary comparison. It may help describe the range of outcomes the trial intended to monitor, but it does not overturn or settle the uncertainty in the primary result. The registry record does not report a combined secondary-outcome conclusion or a prespecified procedure for weighting secondary evidence.

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

• Number of Participants With Death Events
• Number of Participants With Sites of Recurrence
• Patient Reported Fatigue
• Patient-reported Neurotoxicity
• Patient-reported Quality of Life

9. Statistical Interpretation

10. Statistical Methods Explained

What does a risk difference describe?

For a binary outcome, the risk difference compares the proportion of participants with an event in one group with the proportion in the other. The posted estimate is -0.003 for Arm II compared with Arm I. A negative value means the observed event proportion in Arm II was lower in this particular sample, but the estimate by itself does not show whether the difference is distinguishable from random variation. The confidence interval and p-value are needed to understand the uncertainty around it.

The risk difference is an absolute comparison: it expresses the gap between event proportions, not the relative size of one group's event proportion compared with the other's. The posted comparison also gives a risk ratio and an odds ratio, which express relative comparisons in different ways. A risk ratio compares event proportions as a ratio. An odds ratio compares event odds. Both relative measures have a no-difference value of 1, while the risk difference has a no-difference value of 0. These are general definitions; the registry record does not report the specific computational method used for the confidence intervals.

What does the confidence interval add?

A confidence interval is a range of estimates that reflects sampling uncertainty under a statistical procedure and its assumptions. In repeated use of a procedure, intervals constructed in this way would cover the underlying value at the stated confidence level at the stated rate, if the model assumptions hold. It is not a range that assigns a probability to the true value after the data have been observed. The primary risk-difference interval runs from -0.059 to 0.053. Since it includes zero, it does not rule out no difference at the interval's confidence level. Its width also signals uncertainty about the size and direction of the difference.

How should the p-value be read?

The primary comparison reports an unadjusted p-value of 0.921. In general, a p-value describes how often data at least as incompatible with a specified null model would occur if that model and the test assumptions were true. It is not a direct measure of effect size or clinical importance. It is also not the probability that chance alone caused the results. The registry does not report whether the primary test was prespecified in a statistical analysis plan, or the complete details of the test procedure.

The word “unadjusted” indicates that the posted p-value is not reported as corrected for multiple comparisons. That is relevant because the record includes several outcomes. Without a reported multiplicity strategy, the probability of finding at least one apparently notable result across many tests cannot be assessed from the individual unadjusted p-values alone. No secondary analyses are listed as reported analyses in the record.

Why distinguish counts from time-to-event analysis?

In general, a time-to-event analysis uses information about both whether an event happened and when it happened. It can also account for participants whose event has not been observed by the end of their available follow-up through censoring methods. A simple count of participants with events does not, by itself, describe event timing or account for differing follow-up durations. The primary outcome description refers to recurrence-free survival and gives an assessment schedule extending up to 10 years, but the posted primary comparison is a count-based binary outcome. A time-to-event estimate and its analysis method are not reported in the registry record.

This distinction is especially important when the length of follow-up can vary. Two groups might have the same count of participants with events but different event timing, or participants might contribute different amounts of observation. The registry record does not state how censoring, last contact, or unequal follow-up was handled in the posted primary analysis. It is therefore not appropriate to infer a time-to-event effect from the counts alone.

What is over-dispersion, and does it apply?

Over-dispersion is a general issue for count outcomes when observed variation is greater than expected under a chosen count model, often because events are more heterogeneous or clustered than the model allows. It can make standard errors too small if the model fails to account for that extra variation. The primary endpoint is reported as participants with an event, not as a rate of repeated events, and the registry lists no dispersion estimate or count-model analysis. Over-dispersion is therefore not identified as a feature of the posted primary analysis; it is a general modeling consideration rather than a reported finding of this trial.

What does the randomization tell us?

Randomization is intended, in general, to make treatment groups comparable at assignment by balancing both measured and unmeasured prognostic factors on average. It does not guarantee balance in a particular trial, and it cannot by itself correct for missing outcomes, deviations from assigned treatment, or selective reporting. The record describes the trial as randomized but does not report the random-sequence method, allocation concealment, stratification, or detailed analysis population rules. The primary outcome population is identified as all enrolled patients; additional population definitions are not reported in the registry record.

The trial is unmasked. In general, lack of masking can influence reporting, care, or outcome assessment, especially for subjective outcomes such as fatigue, neurotoxicity, and quality of life. Recurrence and death are more concrete outcomes, but the registry does not report how outcome assessment was adjudicated or whether assessors were blinded. These possibilities are methodological considerations, not evidence that bias occurred.

What are the posted relative measures?

The risk ratio of 0.981 and odds ratio of 0.977 offer relative summaries alongside the absolute risk difference. Their respective confidence limits are 0.665 to 1.446 and 0.620 to 1.54. For both relative measures, 1 represents no relative difference. The intervals include 1, which is consistent with the uncertainty reflected in the risk-difference interval. Since all estimates use the same participants and events, agreement among these summaries is not independent evidence.

11. Confidence Intervals and Precision

The primary risk-difference estimate is -0.003, with a 95% confidence interval from -0.059 to 0.053. The interval crosses zero, the value corresponding to no absolute difference. The data therefore do not precisely determine whether the observed event proportion is higher or lower in Arm II, or how large a difference might be compatible with the estimate.

Precision concerns the range of values that remain compatible with the data under the statistical procedure. It is not the same as whether a result is statistically significant, nor does it state which values would matter clinically. The registry record does not give a threshold for a clinically important difference, so the interval cannot be classified against a stated clinical margin. It does show that the estimate is uncertain over values on both sides of no difference.

The relative estimates convey similar uncertainty. The risk ratio is 0.981, with confidence limits from 0.665 to 1.446, and the odds ratio is 0.977, with confidence limits from 0.620 to 1.54. Each interval includes its no-difference value of 1. These estimates should be read as alternate descriptions of the same comparison rather than as separate replications.

Event counts alone do not reveal the precision of a time-to-event comparison, and the registry does not report a time-to-event estimate for the primary outcome. The listed assessment schedule extends up to 10 years, but the count-based result does not show the timing of recurrences or deaths, or the follow-up contribution of each participant. A survival curve or hazard ratio, along with its uncertainty, is not reported in the registry record.

Thus, the primary estimate is compatible with little difference as well as with differences in either direction. The posted data do not establish equivalence, and they do not resolve whether any compatible difference would be clinically important. The confidence limits express statistical uncertainty only under the analysis procedure; the full procedure and its assumptions are not reported in the registry record.

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 figures are participants per arm because the primary outcome is binary and the sample-size assumptions specify equal allocation between the two arms. The assumptions listed for these illustrations are a two-sided significance level of 0.05, equal allocation, and a normal approximation. The effect to detect is set to a stated fraction of the observed effect. The original trial's own design assumptions and sample-size justification are not reported in the registry record.

The 80% power column describes the number of participants per arm that the planning calculation estimates would be needed to detect the selected effect at the stated significance level with 80% power, under its assumptions. The 90% power column gives the corresponding planning estimate for 90% power. In general, power is the chance of detecting an effect of a specified size if that effect is truly present and the model assumptions hold. These figures are not a forecast of the trial's outcome.

The displayed scenarios plan for 100%, 75%, and 50% of the observed effect. For the primary comparison, the listed participants per arm for 80% power are 239,924, 427,402, and 963,612, respectively; the corresponding figures for 90% power are 321,189, 572,169, and 1,290,002. Smaller effects require more participants to distinguish them from random variation, which is why the planning figures increase as the assumed fraction of the observed effect decreases.

These are planning illustrations computed from the observed effect, not the trial's original design calculation. They are not observed power, and they should not be used to judge the trial's design without the original rationale, which is not reported in the registry record. For the primary outcome, the figures are participants per arm, not total events. The registry notes that time-to-event planning displays, if present, use total events, but the primary planning figures concern participants per arm for a binary outcome.

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

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 = 0.146; treatment proportion p₁ = 0.143; average p̄ = 0.145.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4976 = 0.9753.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4976 = 0.4188.
4. n = (0.9753 + 0.4188)² / (-0.0028)² = 239,923.2, rounded up to 239,924 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 <- 0.146179401993355                          # control proportion
p1 <- 0.143333333333333                          # 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(807768)
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 = 239,924 (80% power), 321,189 (90% power)
 75% of observed effect: n per arm = 427,402 (80% power), 572,169 (90% power)
 50% of observed effect: n per arm = 963,612 (80% power), 1,290,002 (90% power)
power.prop.test: 239923.2 
Simulated power at n = 239,924 per arm: 0.800 (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.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary recurrence-or-death outcome and multiple secondary outcomes, including death events, recurrence sites, fatigue, neurotoxicity, and quality of life. Testing several outcomes creates a multiplicity issue: even if no real differences exist, one or more tests can appear notable by chance. As a general statistical principle, interpretation depends on whether outcomes and analyses were prespecified and whether any adjustment or hierarchy controlled the chance of false-positive conclusions across tests.

The registry record does not report a multiplicity strategy, such as an adjustment method, ordered testing plan, or distinction between confirmatory and exploratory analyses. It reports an unadjusted p-value for the primary comparison and lists no reported secondary analyses. Without the full statistical analysis plan, the intended handling of multiple outcomes and comparisons is not reported in the registry record. Individual secondary findings, if encountered elsewhere, would need to be interpreted in that context.

Missing data and incomplete follow-up can also affect interpretation. The primary endpoint description includes last contact as an endpoint of follow-up, and the registry gives an assessment schedule extending up to 10 years. However, details on censoring, loss to follow-up, incomplete outcome ascertainment, or imputation are not reported in the registry record. For the patient-reported outcomes, the instruments and handling of unanswered assessments are also not reported.

The primary outcome population is identified as all enrolled patients. The registry does not state whether every randomized participant was included in the posted comparison, how participants with missing outcome status were handled, or how treatment deviations and discontinuations were addressed. The details of the analysis population beyond the stated all-enrolled population are not reported in the registry record. It would be inappropriate to label the analysis intention-to-treat or per-protocol without that information.

The registry describes the trial as randomized and unmasked, but does not report allocation concealment, randomization procedures, blinded endpoint assessment, or interim analyses. It also does not report the original sample-size justification or a full prespecified statistical plan. These omissions do not show that the trial lacked such procedures; they mean the procedures are not reported in the registry record, and so their role in interpreting the results cannot be verified from that record.

15. Safety Results

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

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: neutrophils (3.8%), febrile neutropenia (1.0%), diarrhea (0.7%), hyperglycemia (0.7%), syncope (0.7%).

The serious adverse event figures are reported among 283 participants at risk in Arm I and 290 participants at risk in Arm II. The record lists 9 serious adverse events in Arm I and 32 in Arm II. These counts describe the posted safety figures, but they do not by themselves establish that the treatment strategy caused an event or explain how events were attributed, coded, or collected.

The registry lists neutrophils as the most frequent serious-event term at 3.8%. Other listed terms are febrile neutropenia at 1.0%, and diarrhea, hyperglycemia, syncope, and thrombosis/thrombus/embolism, each at 0.7%. The record does not indicate in this list whether each percentage is presented across both groups or by group. These terms should not be interpreted as exhaustive of all serious adverse events.

Other adverse events are also reported: 245 participants at risk in Arm I and 284 in Arm II are listed as affected, among 283 and 290 at risk, respectively. The registry record does not provide further detail on severity categories or event counts per affected participant. Deaths attributed within the adverse-event reporting fields are not reported in the registry record. Safety summaries should be read alongside the event definitions, follow-up duration, and attribution procedures, which are not reported in the registry record.

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

The posted primary result shows similar observed counts of recurrence or death events in the two randomized groups: 44.0 among 301.0 participants receiving pelvic radiation therapy and 43.0 among 300.0 receiving brachytherapy followed by paclitaxel and carboplatin. The estimated risk difference is -0.003, and its confidence interval includes zero. The unadjusted p-value is 0.921. Taken together, these figures do not provide clear statistical evidence of a difference in the posted event-count comparison.

That is not the same as evidence that the strategies are equivalent. Equivalence requires a prespecified margin defining how large a difference can be considered acceptably small and an analysis designed to evaluate that margin. The registry record does not report such a margin or an equivalence analysis. The confidence interval spans values on both sides of zero, so the data do not identify a single direction or a precise size of difference.

The result also does not show that either strategy works or fails in an absolute sense. The posted comparison is between two active treatment strategies, not against no treatment. It cannot isolate the effect of a single component in Arm II, because brachytherapy, paclitaxel, and carboplatin are delivered as a combined strategy. Nor does the event-count comparison tell when events occurred or provide a time-to-event estimate.

The registry gives a follow-up schedule extending up to 10 years and defines recurrence-free survival by recurrence, death, or last contact. Yet the primary result is posted as a count of participants with events. Time-to-event curves, hazard ratios, handling of censoring, and detailed follow-up completeness are not reported in the registry record. Those omissions limit conclusions about the timing and durability of outcomes.

The secondary outcomes and safety counts add further domains of interest, but they do not change the meaning of the primary result. The registry does not list reported analyses for most secondary outcomes and does not state how multiple testing was addressed. These data describe the contents of the registry record; they should not be extended to claims about clinical effectiveness, equivalence, or treatment choice.

17. Why This Trial Matters Statistically

Statistically, this trial illustrates the difference between a small observed gap and a precisely established effect. The primary event counts are close, but the relevant inference depends not just on how close the counts appear; it also depends on uncertainty, follow-up, outcome definitions, and the prespecified analysis. The posted risk-difference interval includes zero and permits differences in both directions.

The outcome is called recurrence-free survival, but the posted primary comparison is a binary count of participants with events. That distinction is important for clinical researchers. A participant count can summarize whether an event was recorded, while a survival analysis also uses event timing and can account for censoring. The registry record does not report a time-to-event estimate for the primary outcome, so readers should not infer one from the title or outcome definition.

The trial also compares treatment strategies containing different elements. A randomized comparison can address the contrast between assigned strategies, but it cannot disentangle each component's contribution without additional design features and analyses. The unmasked design and patient-reported secondary outcomes further make reporting and assessment methods relevant, though the record does not establish that bias occurred.

Finally, the planning illustrations demonstrate how sample-size requirements depend on the effect size selected for detection. Smaller effects require more participants under the stated assumptions. Because these illustrations use the observed effect and are not the original trial's design justification, they should not be confused with a retrospective judgment about design adequacy or observed power. The original justification is not reported in the registry record.

18. Limitations

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

Several limitations arise from the way the results are reported. The primary outcome is described as recurrence-free survival, but the posted analysis gives event counts and a binary comparison. A time-to-event estimate, event timing, censoring details, and follow-up completeness are not reported in the registry record. As a result, the count comparison cannot fully convey the survival experience over the stated follow-up period.

The registry identifies the primary outcome population as all enrolled patients, but it does not clarify how missing outcomes, withdrawal, treatment discontinuation, crossovers, or deviations from assigned treatment were handled. The record does not provide a detailed statistical analysis plan, original sample-size rationale, interim-analysis information, multiplicity approach, or clinically important difference threshold. These omissions restrict assessment of whether the posted analysis matches the intended confirmatory plan.

The trial was unmasked. The registry does not report whether recurrence assessment was independently adjudicated or whether outcome assessors were blinded. This may matter particularly for patient-reported outcomes, for which instruments, scoring rules, and missing-response methods are not reported. It is not possible to conclude from these omissions that outcome assessment was biased.

Some regimen details are incomplete in the registry descriptions. The pelvic-radiation arm description ends during a statement about stage II disease or stage I clear cell disease, and the brachytherapy-and-chemotherapy description is incomplete after mentioning intravenous paclitaxel and carboplatin. Additional treatment details are therefore not reported in the registry record.

The secondary outcomes have limited posted analytical detail, and the record lists no reported analyses for them. Safety event counts are reported, but the definitions, attribution, and full arm-specific breakdown of the listed serious-event terms are not stated. These gaps mean the registry record supports a focused reading of the posted estimates, not a comprehensive evaluation of every clinical or safety question raised by the study.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Pelvic Radiation Therapy or Vaginal Implant Radiation Therapy, Paclitaxel, and Carboplatin in Treating Patients With High-Risk Stage I or Stage II Endometrial Cancer — NCT00807768.
• Registry record last updated 2019-03-20; status last verified 2019-03.
• Results first posted 2018-04-27.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

NCT00807768 is a randomized, unmasked, parallel-group phase III trial sponsored by the Gynecologic Oncology Group. It enrolled 601 participants with high-risk stage I or stage II endometrial cancer. The comparison is pelvic radiation therapy versus vaginal cuff brachytherapy followed by paclitaxel and carboplatin.

The primary outcome is the number of participants with recurrence or death events at primary analysis. The posted counts are 44.0 of 301.0 in Arm I and 43.0 of 300.0 in Arm II. The risk-difference estimate for Arm II compared with Arm I is -0.003, with a 95% confidence interval from -0.059 to 0.053 and an unadjusted p-value of 0.921. The interval includes no difference and allows uncertainty in either direction.

The registry lists secondary outcomes relating to death, recurrence sites, fatigue, neurotoxicity, and quality of life, along with adverse-event information. The original sample-size justification, detailed analysis plan, missing-data procedures, and time-to-event analysis for the primary result are not reported in the registry record.