Adjuvant Radiotherapy for Resected Hepatocellular Carcinoma With MVI: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Adjuvant Radiotherapy for Resected Hepatocellular Carcinoma With MVI (NCT04891874), a randomized na study. The primary outcome is Participants Without Recurrence (Disease-free)..

2. Clinical Question

3. Trial Design

NCT04891874 is recorded as a randomized, single-masked, parallel-group trial. Its title concerns adjuvant radiotherapy for resected hepatocellular carcinoma with microvascular invasion; the official title specifies stereotactic body radiation for hepatocellular carcinoma with microvascular invasion and a narrow resection margin. The registry lists Eastern Hepatobiliary Surgery Hospital as the sponsor and records the trial as completed.

The recorded study start was 2015-08-01, and primary completion was recorded as 2020-12-31. Enrollment was 76. The results list 38 participants in the SBRT group and 38 in the surgery-alone group. The phase is recorded as NA. The parallel structure means that participants were assigned to one of the two treatment strategies rather than receiving both strategies in sequence. Randomized allocation can support a group comparison, although the registry information does not explain the randomization procedure or allocation concealment.

The record describes masking as single, but does not identify who was masked. It does not report the treatment schedule, adherence, protocol deviations, or the analysis population. It also does not report the trial's original sample-size rationale or its planned statistical analysis in enough detail to assess those features. These omissions limit interpretation of how the posted results relate to the original protocol, even though the group counts and outcome comparisons are available.

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 Participants Without Recurrence (Disease-free)., measured over DFS rate at 5-years after randomization..

The primary endpoint is titled Participants Without Recurrence (Disease-free). It is a count of participants, with the timeframe specified as the disease-free survival rate at five years after randomization. The registry defines disease-free survival as the interval from randomization to the first recurrence and states that this interval is measured in months. The result posted for the primary measure, however, is the number of participants without recurrence rather than participant-level event times.

Recurrence is clinically relevant in this trial because the registry describes microvascular invasion as associated with early recurrence after surgery. Measuring disease-free status at a stated follow-up point offers a direct summary of how many participants in each group were recorded as recurrence-free at that point. It should not be confused with a complete time-to-event analysis: the posted counts alone do not describe when recurrences occurred, how follow-up varied, or how censoring was handled.

The endpoint is useful for comparing the observed groups, but its interpretation depends on complete and consistently defined follow-up. The registry record does not report the detailed rules for handling missing status or participants whose recurrence status at the five-year point was unknown.

7. Results Posted in the Registry

• Surgery Alone Group: 73.7% (28/38) versus 42.1% (16/38) on control; risk difference 31.6 percentage points (95% CI 10.5 to 52.6), risk ratio 1.75 (1.15–2.66); unadjusted two-proportion z-test p = 0.005.

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) Rate.

• 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:
  • Surgery Alone Group: 52.6% (20/38) versus 76.3% (29/38) on control; risk difference -23.7 percentage points (95% CI -44.5 to -2.8), risk ratio 0.69 (0.49–0.98); unadjusted two-proportion z-test p = 0.031.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Surgery Alone Group than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Overall survival is the secondary outcome reported in the registry. Its posted group counts favor the SBRT group in direction, whereas the primary recurrence-free counts favor surgery alone in direction. This contrast means the outcomes should be considered separately rather than treated as if they were different measurements of the same result. Recurrence status and overall survival capture related but distinct aspects of participants' follow-up.

The secondary comparison has a negative risk difference when expressed as surgery alone relative to SBRT, and its interval lies below zero. The unadjusted p-value is 0.031. The interpretation is limited because the registry does not report a multiplicity adjustment or a detailed analysis plan. When multiple outcomes are assessed, the chance of obtaining a small p-value somewhere among them can increase, depending on how outcomes and analyses were selected and evaluated. The record does not describe how that issue was addressed.

The primary endpoint remains the stated main outcome, while the secondary outcome can add context but does not replace the primary question. The differing observed directions do not prove that one outcome is more important or that the results are inconsistent; they show that the posted summaries are outcome-specific. No participant-level analyses are reported that would clarify the relationship between recurrence and survival for individuals.

The registry reports three outcome measures, but the listed results describe the primary recurrence-free measure and one secondary overall-survival measure. Details for any additional outcome measure are not reported in the registry record. This incomplete accounting, combined with the absence of a reported multiplicity strategy, makes it difficult to assess whether the displayed secondary result was one of several exploratory comparisons. Its estimate and interval are informative as posted, but should carry less evidential weight than a prespecified primary result analyzed under a documented plan.

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

• Overall Survival (OS) Rate.

9. Statistical Interpretation

10. Statistical Methods Explained

These results use binary outcomes: each participant contributes to a count of those meeting the stated status at the follow-up point. The primary measure counts participants without recurrence, and the secondary measure counts participants for the overall-survival rate. A binary count is not the same as a time-to-event analysis. It indicates how many participants were classified in a category at the specified time, but does not retain the timing of events across follow-up.

What does the risk difference describe?

A risk difference compares the proportions with the outcome in two groups by subtracting one group's proportion from the other's. Its zero value represents equal proportions under this measure. The comparison is labeled with surgery alone as the comparison arm and SBRT as the control, so the reported direction is surgery alone relative to SBRT. A positive estimate means the recorded outcome proportion is higher in surgery alone; a negative estimate means it is lower there. This direction depends on the outcome being counted: a higher count without recurrence and a higher count for overall survival are not interchangeable clinical outcomes.

The risk difference is expressed on the proportion scale and is often easier to relate to absolute group differences than a relative measure. It is still a summary of the posted data, not a guarantee that the same difference would occur in other populations. The record also contains relative-risk and odds-ratio summaries, but the risk difference provides the direct absolute comparison used in the posted primary and secondary comparisons.

How should the confidence interval be read?

A confidence interval gives a range of values that are reasonably compatible with the observed data under the method used to construct it. In repeated use of the same procedure under its assumptions, intervals at a stated confidence level would cover the underlying value at the procedure's stated rate. A confidence interval is not a range containing a specified proportion of individual participant outcomes, and it is not a probability distribution for the true effect after seeing the data.

For the primary risk difference, the reported interval is entirely on the positive side of zero. For the secondary risk difference, the reported interval is entirely on the negative side of zero. This indicates that the intervals do not include a zero difference under the reported calculations. The registry does not state all details of the interval method in the posted record, so the interval's precision should be considered alongside that limited methodological information.

What does the p-value add?

The unadjusted p-value summarizes the compatibility of the observed data with a null model of no group difference, given the method's assumptions. Smaller values indicate that the observed data would be less usual under that model. A p-value does not quantify the magnitude of the group difference and cannot account for bias, missing outcome data, or unreported analysis choices. The reported values are unadjusted; no multiplicity correction is reported in the registry record.

In this trial, each comparison is based on the posted group counts. The primary comparison has an unadjusted p-value of 0.005, and the secondary overall-survival comparison has an unadjusted p-value of 0.031. These should be interpreted with the confidence intervals and the status of each outcome in mind, rather than as stand-alone proof of a treatment effect.

Why does the endpoint definition matter?

The registry describes disease-free survival as the interval from randomization to first recurrence, measured in months, but reports the primary outcome as a count of participants without recurrence at five years. A time-to-event analysis can use the timing of events and account for differing follow-up through methods that incorporate censoring. A simple count at a fixed time does not show whether participants had equal observation time or how incomplete follow-up was treated. The registry record does not report a time-to-event estimate, censoring details, or participant-level data for this endpoint.

The distinction also matters for the overall-survival measure. The record gives a five-year rate and counts, but does not specify how deaths, loss to follow-up, or incomplete status were handled in calculating that measure. Thus the reported comparisons are best understood as the registry's posted binary summaries. The record does not provide reported analyses beyond these comparisons, and does not give enough detail to verify the analytic choices independently.

11. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimated group difference. The primary risk-difference interval is entirely above zero, while the secondary overall-survival interval is entirely below zero. The intervals therefore point in the same respective directions as their estimates and exclude zero under the reported calculations. They also span a range of possible group differences, so the point estimates should not be treated as exact descriptions of an underlying effect.

The groups each contain 38 participants, and the observed outcome counts are 16 and 28 for the primary endpoint, and 29 and 20 for overall survival. These group sizes and counts provide the basis for the reported comparisons, but do not eliminate uncertainty. The registry does not report an adjusted analysis, an interval method, or enough detail about missing observations to assess how those factors might affect precision.

Confidence intervals are more informative than a significance label alone because they communicate the range of values compatible with the analysis, including the direction and breadth of uncertainty. However, they do not account for every possible source of error. In particular, a confidence interval calculated from the posted group summaries cannot by itself address selection issues, deviations from the intended analysis, measurement problems, or multiplicity. Those elements are not sufficiently described in the registry record.

The primary and secondary intervals should be interpreted separately because they concern different outcomes. Their directions differ, and the record does not report a joint analysis that combines them. The secondary interval and p-value also arise in the context of an endpoint designated as secondary, so their apparent precision does not remove the additional caution associated with multiple outcome assessment.

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, not total trial enrollment and not total events. They illustrate sample sizes for future planning based on the observed effect, rather than documenting the trial's original design rationale. The registry's 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 set as the stated fraction of the observed effect. The original trial's own sample-size assumptions are not reported in the registry record.

The 80% power column gives the planned number of participants per arm under which the calculation would have an 80% chance of detecting an effect of the specified size, if that effect and the calculation assumptions were correct. The 90% power column expresses the corresponding planning target for a higher chance of detection. These are design calculations, not guarantees about what a trial will find. The planning illustrations use the observed effect as their starting point, so they should not be treated as evidence that the trial had a particular probability of finding its result.

The rows for 100%, 75%, and 50% of the observed effect ask how the planning requirement changes if the effect targeted is the full observed effect or a smaller fraction of it. Smaller effects are harder to distinguish from random variation under the same assumptions and therefore require more participants per arm. The planning curve reflects that relationship: requirements rise as the target effect becomes smaller. This is not observed power; it is a planning illustration based on the stated assumptions, not a retrospective measure of the trial's success.

These calculations do not account for unreported factors such as loss to follow-up, outcome ascertainment, or the original protocol's analysis choices. They also do not establish that the observed effect is a stable estimate for future research. Their purpose is to show how sample-size needs depend on the effect size selected for planning.

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 = 0.421; treatment proportion p₁ = 0.737; average p̄ = 0.579.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6982 = 1.3685.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6616 = 0.5568.
4. n = (1.3685 + 0.5568)² / (0.3158)² = 37.2, rounded up to 38 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 <- 0.421052631578947                          # control proportion
p1 <- 0.736842105263158                          # 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(4891874)
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 = 38 (80% power), 50 (90% power)
 75% of observed effect: n per arm = 69 (80% power), 91 (90% power)
 50% of observed effect: n per arm = 157 (80% power), 209 (90% power)
power.prop.test: 37.2 
Simulated power at n = 38 per arm: 0.809 (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

Multiplicity arises when a study assesses more than one endpoint or conducts multiple analyses. Each additional test creates another opportunity for a result to appear unusual under a no-effect model by chance. Researchers may address this by prespecifying a hierarchy, adjusting significance thresholds, or otherwise defining how the results will be interpreted together. The registry record does not report whether any such strategy was planned or used.

The record identifies a primary recurrence-free outcome and a secondary overall-survival outcome. It reports unadjusted p-values for both comparisons, but no adjustment is reported. The registry also records three outcome measures, while the listed results provide details for the primary outcome and one secondary outcome. Details of any further measure are not reported in the registry record. Without a full outcome and analysis plan, it is not possible to determine how many analyses were considered or whether the posted outcomes reflect all planned assessments.

Missing data are another important concern for fixed-time status outcomes. The record does not report the number of participants with unknown outcome status, the completeness of follow-up, or how missing observations were handled. It also does not report censoring methods or an analysis of participants who did not have status recorded at the five-year point. The posted group counts can therefore be interpreted as the recorded results, but the record does not establish how representative those counts are of all randomized participants at that time.

The analysis population is not reported in the registry record. It is not stated whether participants were analyzed according to their randomized group, whether any participants were excluded, or how protocol deviations were treated. The record reports zero posted analyses in its analysis listing, despite providing outcome comparisons; details of a formal analysis set and statistical procedure are not posted by the registry entry. These omissions limit assessment of robustness and make it difficult to reproduce the comparisons or evaluate potential bias.

15. 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 non-serious adverse event terms

The registry separates serious adverse events from other adverse events and deaths. It reports zero serious adverse events in the SBRT group and zero in the surgery-alone group. It also reports zero deaths in each group in the adverse-event listing. For other adverse events, the posted counts are 12 in the SBRT group and zero in the surgery-alone group; each group lists 38 participants at risk.

These figures should be read as registry counts, not as a complete description of the nature or timing of adverse events. The record does not list serious adverse-event terms, and it does not provide descriptions of the other adverse events, their severity, duration, or relationship to treatment. A zero count in a posted category does not establish that no events could have occurred outside the recorded scope or observation period.

The safety figures also do not replace a full comparison of adverse-event risk. The registry does not report the collection procedures, follow-up completeness, or detailed safety analysis. The counts can describe what was entered in the record, but they do not support broader conclusions about safety beyond those reported observations.

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

The posted results show different observed patterns for the two outcomes. The count without recurrence at five years is higher in the surgery-alone group, while the five-year overall-survival count is higher in the SBRT group. The reported risk-difference intervals exclude zero in the respective directions, and the unadjusted p-values are 0.005 for the primary comparison and 0.031 for the secondary comparison. These findings describe the registry's group comparisons; they do not establish that the treatment caused the differences.

Randomization is an important feature because it is intended to make groups comparable on average, but interpretation also depends on how randomization was implemented, whether participants remained in their assigned groups for analysis, and whether follow-up and outcome ascertainment were complete. Those details are not reported in the registry record. The primary outcome is presented as a count at a fixed time, even though the description defines disease-free survival as time to first recurrence. The absence of reported time-to-event analyses and missing-data handling makes the endpoint's follow-up structure harder to assess.

The secondary overall-survival finding should be interpreted with additional caution because it is a secondary outcome and no multiplicity adjustment or full analysis plan is reported. The safety counts are also limited: serious events and deaths are recorded as zero in both groups, while the other adverse-event counts differ, but the event descriptions and ascertainment details are absent.

These results do not show how the findings apply beyond the trial's stated population, and the registry record does not provide enough participant characteristics to judge that question. They do not establish a treatment recommendation, prove that one strategy is superior, or show that the observed outcome patterns will recur in other studies. They are best understood as unadjusted, registry-posted summaries that require attention to the endpoint definitions, uncertainty, missing methodological details, and the difference between primary and secondary evidence.

17. Why This Trial Matters Statistically

This trial is statistically interesting because its primary and secondary outcomes point in different directions. The recurrence-free count is higher in the surgery-alone group, while the overall-survival count is higher in the SBRT group. This makes clear why outcome definitions matter: a conclusion drawn from one endpoint cannot automatically stand in for a conclusion about another.

The primary measure also highlights a distinction between a time-to-event concept and a fixed-time count. Disease-free survival is described as the time from randomization to first recurrence, but the posted result gives the number without recurrence at five years. Time-to-event methods and fixed-time counts answer related but not identical questions. The registry record does not provide the additional follow-up and censoring details needed to reconcile those approaches.

Precision and analysis context matter as well. The intervals exclude zero for both comparisons, but the secondary result is unadjusted and no multiplicity approach is reported. The registry does not describe the original sample-size rationale, analysis population, missing-data handling, or formal statistical plan. These gaps illustrate why a p-value alone is insufficient to assess evidence: the endpoint, estimate, uncertainty, analysis choices, and completeness of reporting all shape interpretation.

Finally, the sample-size illustrations show how planning depends on the effect size chosen. Smaller target effects require more participants per arm under the stated planning assumptions. Those calculations are not the trial's original rationale and do not measure observed power. Keeping planning calculations separate from trial results helps prevent a retrospective calculation from being mistaken for evidence that the study was adequately designed.

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 trial-specific limitations affect interpretation. The registry does not report the eligibility criteria in detail or describe the enrolled participants beyond the disease and treatment context. It also does not identify who was masked, explain the randomization procedure, or provide treatment-delivery details. These omissions make it harder to judge how closely the trial's implementation can be evaluated from the registry record.

The primary endpoint is described as disease-free survival measured from randomization to first recurrence, but the posted result is a count of participants without recurrence at five years. No time-to-event estimate, censoring information, or follow-up completeness is reported. The record also does not explain how missing five-year status was handled. These details matter because a fixed-time count can be affected by incomplete or unequal observation.

The registry does not report the analysis population, whether participants were analyzed in their randomized groups, or how protocol deviations were addressed. It gives no detailed statistical plan, and does not report whether the secondary outcome or other measures were adjusted for multiplicity. Although three outcome measures are recorded, only the primary measure and one secondary outcome have descriptions in the posted results; details for any additional measure are not reported in the registry record.

Safety reporting is limited to counts. Serious adverse-event terms are not listed, the other adverse events are not described, and the record does not explain the safety observation period or collection method. These gaps do not invalidate the posted counts, but they restrict what can be inferred from them. The registry data therefore allow a focused description of the reported group summaries, while leaving important questions about analysis, follow-up, and reporting unresolved.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Adjuvant Radiotherapy for Resected Hepatocellular Carcinoma With MVI — NCT04891874.
• Registry record last updated 2021-09-10; status last verified 2021-08.
• Results first posted 2021-09-10.
• 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

NCT04891874 is a completed randomized, single-masked, parallel-group trial sponsored by Eastern Hepatobiliary Surgery Hospital. It enrolled 76 participants and compared adjuvant stereotactic radiotherapy with surgery alone after resection of hepatocellular carcinoma with microvascular invasion and a narrow resection margin. The recorded start date was 2015-08-01, and primary completion was 2020-12-31.

The primary outcome is the count without recurrence at five years after randomization. The secondary outcome described in the record is overall survival at five years. Posted counts show different directions across these outcomes, and the reported unadjusted comparisons should be interpreted with their uncertainty and the registry's missing methodological detail in mind. Information about analysis populations, missing data, multiplicity, and the original sample-size assumptions is not reported in the registry record.