Hormone Therapy and Radiation Therapy in Treating Patients With Prostate Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Hormone Therapy and Radiation Therapy in Treating Patients With Prostate Cancer (NCT00005044), a randomized phase 3 study. The primary outcome is Disease-specific Survival (DSS) (10-year Rates Reported).

2. Clinical Question

The trial was registered as a phase III randomized comparison and is marked completed. Its stated rationale was that the relative effectiveness of different hormone-therapy and radiation-therapy regimens was not yet known. The primary endpoint concerns survival attributed specifically to prostate cancer, rather than survival from any cause. The registry describes the outcome as a 10-year rate, with follow-up from randomization until death or study termination, whichever occurred first.

This clinical question is narrower than asking whether hormone therapy or radiation therapy is effective in general. Both groups received TAS and radiation therapy; the randomized contrast was the duration of TAS before radiation, with concurrent TAS in both regimen descriptions. The registry record does not report an additional treatment comparison against radiation alone or against no treatment.

3. Trial Design

This was a phase III, randomized, parallel-group trial with no masking reported. The registry lists enrollment of 1579 and marks the study completed. The record identifies the Radiation Therapy Oncology Group as sponsor. It describes two treatment groups: TAS for 8 weeks followed by radiation therapy with concurrent TAS, and TAS for 28 weeks followed by radiation therapy with concurrent TAS. TAS was described as an LHRH agonist with Casodex or Eulexin.

The primary endpoint was disease-specific survival (DSS), reported as a 10-year rate. The outcome population is described as eligible patients with follow-up data. The record states that patients were followed from randomization until death or study termination, whichever occurred first. It lists overall survival, disease-free survival, recurrence-related outcomes, biochemical failure outcomes, and treatment-induced morbidity among the secondary outcomes. Only some of those secondary outcomes have results detailed in the posted results; the other posted endpoint results are not reported in the registry record.

Randomization supports a comparison between the assigned regimens, but the registry record does not report the allocation ratio, sequence-generation method, concealment procedures, or trial-specific sample-size justification. It also does not report a protocol-level statistical analysis plan, an interim-analysis plan, or the handling of treatment changes and adherence. The fact that masking is listed as none means the record does not describe a masked trial; further details about who knew assignments are not reported in the registry record.

The posted primary results give 10-year DSS estimates of 95.0% for TAS x 8 Weeks and 96.0% for TAS x 28 Weeks, each with a 95% confidence interval. A separate sponsor analysis is reported using a log-rank method and a hazard ratio. Those results address the randomized comparison through a time-to-event analysis, whereas the percentages express the estimated 10-year outcome. The distinction matters when interpreting the effect estimate and its precision.

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: Eligible patients with follow-up data

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: Eligible patients with follow-up data

6. Primary Endpoint

The registry lists the primary outcome as Disease-specific Survival (DSS) (10-year Rates Reported), measured over From randomization to 10 years. (Patients are followed until death or study termination, whichever occurs first.).

Disease-specific survival measures time from randomization to death due to prostate cancer, using the study chair's review to classify cause of death. The registry defines prostate-cancer death through specified categories: a death certified primarily as due to prostate cancer, a complication of therapy irrespective of disease status, disease progression without anti-tumor therapy, or a qualifying rise in serum prostate-specific antigen during or after salvage androgen suppression therapy. The posted description specifies a rise exceeding 1.0 ng/ml on at least two consecutive occasions for that last category.

Deaths from other causes are treated as competing risks, while other participants are censored. DSS is estimated with the cumulative incidence method. In general, a competing risk is an event that prevents the event of interest from occurring later; accounting for competing risks can make an estimate of event probability more appropriate than treating every other-cause death as ordinary loss to follow-up. The endpoint is therefore not interchangeable with overall survival, which counts death from any cause.

A 10-year DSS rate gives a time-specific summary of the probability of avoiding prostate-cancer death under the endpoint's definitions. It is useful for this trial because the comparison concerns long-term cancer-specific outcomes after the two treatment regimens. The registry record does not report additional details on cause-of-death adjudication beyond the stated definition, nor does it describe how uncertain cause classifications were resolved.

7. Results Posted in the Registry

• TAS x 28 Weeks: 96.1% (708/737) versus 94.9% (714/752) on control; risk difference 1.1 percentage points (95% CI -1.0 to 3.2), risk ratio 1.01 (0.99–1.03); unadjusted two-proportion z-test p = 0.298.

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.

Overall Survival (OS) (10-year Rates Reported)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • TAS x 28 Weeks: 67.0% (494/737) versus 66.0% (496/752) on control; risk difference 1.1 percentage points (95% CI -3.7 to 5.9), risk ratio 1.02 (0.95–1.09); unadjusted two-proportion z-test p = 0.662.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between TAS x 28 Weeks and the control. The interval gives the range of effects that remain compatible with the data.

Disease-free Survival (DFS) (10-year Rates Reported)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • TAS x 28 Weeks: 23.1% (170/737) versus 23.9% (180/752) on control; risk difference -0.9 percentage points (95% CI -5.2 to 3.4), risk ratio 0.96 (0.80–1.16); unadjusted two-proportion z-test p = 0.692.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between TAS x 28 Weeks and the control. The interval gives the range of effects that remain compatible with the data.

Time to Second Biochemical Failure (SBF) (10-year Rates Reported)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • TAS x 28 Weeks: 9.0% (66/737) versus 10.0% (75/752) on control; risk difference -1.0 percentage points (95% CI -4.0 to 2.0), risk ratio 0.90 (0.66–1.23); unadjusted two-proportion z-test p = 0.502.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between TAS x 28 Weeks and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary results span overall survival, disease-free survival, and second biochemical failure. The estimates are broadly close between the treatment groups, and the sponsor-reported log-rank confidence intervals for these outcomes include the hazard-ratio null value of 1. Across this set of outcomes, the results do not offer a clearly different overall pattern from the primary comparison. They provide context, but they do not turn the primary finding into proof of equal effects.

Secondary endpoints address related but distinct questions. Overall survival includes death from any cause, unlike disease-specific survival. Disease-free survival and biochemical failure capture other aspects of disease status or recurrence as defined for the trial. Their interpretation depends on each endpoint's operational definition and analysis. The posted descriptions do not include a full definition for every secondary outcome, and some outcomes listed in the registry do not have results reported in the registry record.

Unadjusted secondary results carry less weight when many outcomes are examined because the chance of at least one apparently unusual result can increase with the number of comparisons. In general, multiplicity procedures can address that concern, but the registry record does not report whether any adjustment was prespecified or performed. Outcomes may also be correlated, so they should not be treated as independent replications of the primary endpoint. These secondary results are best viewed as supporting descriptive evidence, with the primary endpoint and its reported analysis remaining central.

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

• Overall Survival (OS) (10-year Rates Reported)
• Disease-free Survival (DFS) (10-year Rates Reported)
• Clinical Patterns of Tumor Recurrence: Time to Locoregional Progression (LRP) and Time to Distant Metastasis (DM) (10 Year Rates Reported)
• Time to First Biochemical Failure (BF) (10-year Rates Reported)
• Time to Second Biochemical Failure (SBF) (10-year Rates Reported)
• Treatment-induced Morbidity (Highest Grade Toxicity Reported Per Patient)

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports two forms of comparison for the primary outcome. The sponsor's analysis is a log-rank comparison summarized by a hazard ratio. The unadjusted re-calculation is a risk difference derived from the posted summary statistics. These statistics answer related but not identical questions, so their numerical values should not be expected to match.

What does the hazard ratio describe?

In general, a hazard ratio compares the instantaneous event rates over follow-up between groups, among participants who have not yet had the event. It is a time-to-event measure and uses information about when events occur, rather than reducing the entire follow-up to a single event proportion. A hazard ratio of 1 is the null value for equal hazards. The sponsor reports a hazard ratio of 0.81 for the primary comparison, with a 95% confidence interval of 0.48 to 1.39 and p = 0.45. Because the interval includes 1, the reported analysis does not establish a clear difference in hazards.

The log-rank test compares the observed timing of events between groups against what would be expected under the null hypothesis of no difference in event experience. In general, it is a common method for comparing time-to-event outcomes. Its interpretation depends on the trial's analysis population, event definitions, and follow-up. Further details of the trial-specific analysis are not reported in the registry record.

How does a risk difference differ?

A risk difference compares the estimated event probabilities between groups at a specified time or in a specified summary. The recalculated primary risk difference is 0.011, with a 95% confidence interval from -0.010 to 0.032. Zero is its null value: a confidence interval spanning zero is compatible with no difference and with effects in either direction. The displayed DSS rates are estimates of survival, while the reported calculation fields use the posted summary statistics to produce a risk difference. The record's unadjusted recalculation is not a substitute for the sponsor's time-to-event analysis.

In general, a time-specific rate and a hazard ratio may emphasize different aspects of follow-up. The rate describes estimated outcome status at a particular time, whereas the hazard ratio summarizes relative event rates across follow-up under the analysis model. They can therefore yield different-looking numerical summaries without necessarily contradicting one another. The record does not report enough detail to reconstruct the sponsor's calculation independently, and the sponsor's analysis takes precedence as the trial's reported analysis.

What do the p-values say?

A p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The primary sponsor p-value is 0.45; the unadjusted re-calculation gives p = 0.298. Neither is evidence of a statistically clear difference under its respective method. The registry record does not state a trial-wide significance threshold or provide a complete prespecified analysis plan, so interpretation should remain tied to the reported analyses rather than assume unreported decision rules.

Why does the endpoint use competing risks?

General explanation: when a participant dies from a cause other than prostate cancer, that death prevents a later prostate-cancer death from being observed. The registry therefore treats other-cause death as a competing risk and estimates DSS with the cumulative incidence method. Censoring is used for other participants as described in the endpoint definition. This differs from simply counting every death as the event of interest, and it differs from overall survival, which concerns death from any cause.

Confidence intervals describe statistical uncertainty in an estimate, conditional on the method and its assumptions. They are not guarantees that a particular range contains the true value, and they do not capture every possible source of bias or uncertainty. The methods used to calculate the displayed rate intervals are not reported in the registry record. The methods for handling missing follow-up, deviations from assigned treatment, and any covariate adjustment are also not reported.

How should the planning illustrations be understood?

The planning illustrations use the observed effect as an input and show hypothetical sample-size requirements under stated assumptions. They are not the original trial's design justification, which is not reported in the registry record. The calculations assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. Time-to-event planning figures give total events, while other planning figures give participants per arm. In general, detecting a smaller effect requires more information, which can mean more participants or more events depending on the outcome and design.

11. Confidence Intervals and Precision

Precision is reflected in the width of a confidence interval: narrower intervals indicate less statistical uncertainty about the estimate under the method used, while wider intervals indicate more. For the primary DSS estimates, the 95% confidence intervals are 93.0% to 97.0% for TAS x 8 Weeks and 95.0% to 98.0% for TAS x 28 Weeks. These intervals describe uncertainty in each group's 10-year rate; they are not direct intervals for the difference between groups.

The recalculated risk-difference interval runs from -0.010 to 0.032. Because it crosses zero, the data are compatible with no difference as well as with values on either side of zero. The sponsor's hazard-ratio interval, 0.48 to 1.39, crosses 1, the null value for that relative measure. Both intervals are important: the point estimates alone give only one best estimate, while the intervals show how uncertain that estimate is.

In general, confidence intervals reflect sampling variation under the analysis method; they do not automatically account for bias, missing data, outcome misclassification, or unreported design choices. The interval calculation method for the displayed group rates is not reported in the registry record. Nor does the record provide a clinically acceptable difference or a precision target against which to judge whether the intervals are sufficiently narrow. Thus, the posted intervals support caution: they do not establish a clear difference, but they also do not demonstrate that only negligible differences remain plausible.

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 hypothetical illustrations based on the observed effect, not the original trial's design justification. The original sample-size assumptions are not reported in the registry record. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect.

For ordinary outcomes in these planning illustrations, the figures mean participants per arm. For time-to-event outcomes, the figures mean the total number of events, not participants per arm or a total enrollment target. This distinction matters because a time-to-event design's information depends substantially on the number of events observed. The displayed planning figures should be read according to the outcome type rather than treated as interchangeable counts.

The 80% power column describes a hypothetical design that would have an 80% probability of detecting the stated effect under the planning assumptions if that effect were the true one. The 90% power column describes the corresponding design with a 90% probability. In general, a higher target probability requires more information. Likewise, planning for 75% or 50% of the observed effect means targeting a smaller effect, and smaller effects need more participants or events to detect with the same power target. These figures are not observed power; they are planning illustrations based on assumptions and the observed effect, and do not report the probability that the completed trial detected an effect.

They should not be interpreted as proof that the completed trial was adequately or inadequately sized. That judgment would require the trial's original target, assumptions, design details, and intended analysis, which are not reported in the registry record. The planning calculations also do not establish what difference would be clinically important.

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 = 0.949; treatment proportion p₁ = 0.961; average p̄ = 0.955.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.2930 = 0.5742.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.2929 = 0.2465.
4. n = (0.5742 + 0.2465)² / (0.0112)² = 5,386.1, rounded up to 5,387 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 <- 0.949468085106383                          # control proportion
p1 <- 0.960651289009498                          # 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(5044)
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 = 5,387 (80% power), 7,210 (90% power)
 75% of observed effect: n per arm = 9,860 (80% power), 13,199 (90% power)
 50% of observed effect: n per arm = 22,822 (80% power), 30,552 (90% power)
power.prop.test: 5386.1 
Simulated power at n = 5,387 per arm: 0.804 (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. Analyses Reported by the Sponsor

The sponsor's own primary analysis is reported as a log-rank analysis with a hazard ratio of 0.81, 95% CI 0.48 to 1.39, and p = 0.45. This is the trial's reported inferential result for the primary comparison. The confidence interval includes the null hazard ratio of 1, and the p-value does not indicate a statistically clear difference under that analysis.

The unadjusted re-calculation uses posted summary statistics and reports a risk difference of 0.011, 95% CI -0.010 to 0.032, and p = 0.298. A risk difference and a hazard ratio summarize treatment contrast differently, and the calculation methods also differ. Their estimates and p-values need not coincide. The registry record does not flag a discrepancy between the sponsor result and the recalculation. Accordingly, the sponsor's analysis should be read as the primary reported analysis, with the recalculation serving as a secondary, unadjusted view rather than a replacement.

The registry does not report additional information that would allow an independent reconstruction of the sponsor's analysis, including detailed modeling choices or the complete analysis population. Neither result demonstrates equivalence, and neither by itself establishes clinical importance.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple outcomes, including DSS, overall survival, disease-free survival, recurrence and metastasis outcomes, biochemical failure outcomes, and treatment-induced morbidity. Eight analyses are reported in the record, while only a subset of the listed secondary endpoints has detailed results in the posted results. The registry record does not report a multiplicity strategy, such as a hierarchy of endpoints, an adjustment to control family-wise error, or a designated confirmatory status for secondary analyses. Without that information, secondary p-values should be treated as exploratory or descriptive rather than as independent confirmatory tests.

Missing-data handling is also not reported in the registry record. The primary outcome population is described as eligible patients with follow-up data, but the record does not state how participants without follow-up were handled, how censoring was assessed, or whether sensitivity analyses addressed missing outcomes. Although the endpoint description specifies that other-cause deaths are competing risks and all others are censored, it does not explain how missing or incomplete cause-of-death information was handled.

The record does not describe whether the analyses followed intention-to-treat, per-protocol, or another analysis-population definition. It also does not report how treatment discontinuations, crossovers, or deviations from assigned regimen were addressed. Randomization establishes the assigned comparison, but interpretation of the results can depend on how participants are included and how follow-up and intercurrent events are managed.

These omissions do not show that the analyses were inadequate; they mean the registry record does not contain enough information to assess these methodological choices. A protocol or detailed analysis report would be needed to determine the prespecified endpoint hierarchy, missing-data methods, analysis populations, and any multiplicity control.

16. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: late rt toxicity: bladder: nos (4.4%), late rt toxicity: other: nos (3.5%), late rt toxicity: other gu: nos (3.4%), diarrhea nos (2.4%), alanine aminotransferase increased (2.0%).

The registry reports serious adverse events for groups at risk of 747 in TAS x 8 Weeks and 735 in TAS x 28 Weeks. Serious adverse events affected 121 and 109 participants, respectively. Other adverse events are reported for the same group-specific at-risk counts, with 695 and 707 participants affected, respectively. These are counts of participants with events, not event totals; a participant may potentially experience more than one event, but the record does not clarify the counting rules for the aggregate figures.

The listed common serious-event terms include Late RT Toxicity: Bladder: NOS at 4.4%, Late RT Toxicity: Other: NOS at 3.5%, Late RT Toxicity: Other GU: NOS at 3.4%, Diarrhea NOS at 2.4%, Alanine aminotransferase increased at 2.0%, and Late RT Toxicity: Other GI: NOS at 1.3%. The registry record does not provide enough detail in these summaries to attribute each listed term to one group or to compare group-specific rates for the individual terms. Death counts are not reported.

These safety figures are descriptive and should be interpreted alongside the trial's event definitions, follow-up, and ascertainment procedures, which are not fully reported in the registry record. They do not by themselves establish a causal difference in harm between regimens.

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

The posted evidence answers a focused question: how the 10-year disease-specific survival estimates compare between two randomized TAS-duration regimens, each followed by radiation therapy with concurrent TAS. The estimates are 95.0% and 96.0%, and the sponsor's log-rank analysis reports a hazard ratio of 0.81 with a 95% confidence interval of 0.48 to 1.39. The primary analysis does not show a statistically clear difference between the assigned regimens.

That result does not mean the regimens have been proven equivalent. Demonstrating equivalence or non-inferiority requires an appropriate prespecified margin and analysis; neither is reported in the registry record. The confidence interval includes the null, but also permits a range of relative effects. The re-calculated risk-difference interval likewise includes zero while allowing values on either side. Absence of clear statistical evidence for a difference is not proof that no difference exists.

The endpoint is disease-specific survival, not overall survival, and its classification depends on the registry's stated cause-of-death definitions and competing-risk method. The result should not be generalized beyond the trial's intermediate-risk prostate cancer setting or its particular treatment regimens without additional evidence. The registry record does not provide detailed eligibility criteria, treatment adherence, subgroup analyses, or the information needed to assess how representative the trial population is of other settings.

The finding is statistical evidence from a randomized comparison, not medical advice or a treatment recommendation. Decisions about treatment require clinical context and evidence beyond these registry results. The posted information supports a restrained conclusion: the reported analyses did not establish a clear difference in the primary endpoint, and the uncertainty does not justify claiming that the treatment durations are identical or that one is preferable.

18. Why This Trial Matters Statistically

This trial illustrates why endpoint definitions matter. Disease-specific survival is not the same as overall survival: it focuses on deaths attributed to prostate cancer, treats other-cause deaths as competing risks, and uses a cumulative incidence method. A reader interpreting the primary result therefore needs to know how deaths were classified and how competing events were handled, not just the reported survival percentages.

It also demonstrates the distinction between a time-specific rate and a time-to-event comparison. The 10-year rates describe estimated status at a chosen follow-up horizon, while the sponsor's log-rank analysis summarizes event timing through a hazard ratio. These measures can be informative together, but they answer different statistical questions. The unadjusted risk-difference recalculation offers another summary, not a replacement for the sponsor's analysis.

Finally, the trial is a useful example of interpreting uncertainty rather than treating a p-value as a verdict. The reported confidence intervals include their respective null values, but their ranges also leave room for effects in either direction. This supports neither a claim of a demonstrated difference nor a claim of equivalence. The registry's incomplete reporting of design assumptions, multiplicity, missing-data handling, and analysis populations also shows why registry results need to be read alongside a detailed protocol or analysis report when those documents are available.

19. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The registry record does not report the original sample-size justification, target effect, or design assumptions. It also does not provide a complete statistical analysis plan, including the prespecified primary analysis details, analysis population, or any planned interim analyses. These gaps limit assessment of whether the reported methods match the intended design.

Details of randomization procedures and allocation concealment are not reported. Masking is listed as none, but further information about knowledge of assignment among participants, clinicians, outcome assessors, and analysts is not reported. The record does not describe treatment adherence, crossovers, or how deviations from assigned therapy were handled.

The outcome population is eligible patients with follow-up data, but missing-data methods and censoring assumptions are not described. The primary endpoint depends on cause-of-death classification, yet details beyond the stated definition and study chair review are not reported. The cumulative incidence approach addresses competing risks as described, but the record does not provide enough information to assess sensitivity to classification or follow-up assumptions.

The registry lists several outcomes, but a multiplicity plan is not reported. Some listed secondary endpoints have no results reported in the registry record, limiting a complete view of the intended outcome set. The safety summaries provide group-level counts and selected serious-event terms, but do not report death counts or enough detail to evaluate group-specific rates for each term. These are limitations of the information available in the registry record, not evidence that the trial did not collect or analyze the information.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Hormone Therapy and Radiation Therapy in Treating Patients With Prostate Cancer — NCT00005044.
• Registry record last updated 2022-06-15; status last verified 2022-05.
• Results first posted 2017-10-31.
• 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.

23. Record Summary

NCT00005044 is a completed phase III randomized parallel-group trial comparing TAS for 8 weeks with TAS for 28 weeks, with radiation therapy and concurrent TAS in both regimens. The record lists enrollment of 1579 and identifies the Radiation Therapy Oncology Group as sponsor.

The primary outcome is 10-year disease-specific survival, estimated with the cumulative incidence method and treating other-cause deaths as competing risks. The posted rates are 95.0% for the 8-week regimen and 96.0% for the 28-week regimen. The sponsor's log-rank analysis reports a hazard ratio of 0.81, 95% CI 0.48 to 1.39, and p = 0.45; it does not establish a statistically clear difference.

Secondary results cover selected survival and biochemical failure outcomes, with a broadly similar pattern of no clearly established group difference. The registry record does not report key details including the original sample-size justification, multiplicity procedures, missing-data methods, or analysis-population definitions.