Chemoprevention of Gastric Carcinogenesis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Chemoprevention of Gastric Carcinogenesis (NCT02794428), a randomized phase 2 study. The primary outcome is Gastric Epithelial Cell DNA Damage, at 6 Months Versus Baseline.

2. Clinical Question

The study addresses a biomarker question in people described as having premalignant gastric lesions. Its primary measure is not the occurrence of gastric cancer or another clinical event; it is a laboratory assessment of DNA damage in gastric epithelial cells. The registry describes comparison of the groups' mean changes at 6 months. It does not report a threshold for a clinically important biomarker difference or establish how a change in this measure relates to later clinical outcomes.

The question is therefore about the contrast between the two assigned groups on this specified measure and timeframe. The posted primary analysis includes only participants with evaluable gamma H2AX staining at both baseline and the 6-month assessment. That restriction is part of how the reported comparison should be understood: it concerns participants with measurements at both time points, rather than necessarily everyone enrolled.

3. Trial Design

The registry identifies this as a phase 2 study with randomized allocation and double masking. It lists eflornithine and eflornithine placebo as interventions and identifies placebo as the control. The registry also labels the study model as single group, while describing randomization and listing two intervention groups. That inconsistency in the design descriptors is not explained in the registry record.

The enrollment is 91. The study status is completed; the recorded start date is 2016-09-19 and the primary completion date is 2022-12-20. The sponsor is Vanderbilt-Ingram Cancer Center. The brief description concerns oral eflornithine in male and female participants ages 30-60 with gastric premalignant lesions in two high-risk regions of Latin America. Specific countries, the detailed eligibility criteria, and the allocation ratio are not reported in the registry record.

The primary outcome is the change in gastric epithelial cell DNA damage from baseline up to 6 months, measured as the percentage of positive cells. The stated analysis population includes only participants with evaluable gamma H2AX staining at both time points. The registry says that a central pathology review of pre- and post-intervention biopsy specimens was performed at Vanderbilt University for the DNA-damage endpoints.

The record lists five secondary outcomes: DNA damage at 18 and 24 months, and changes in gastric histopathology or histology scores at 6, 18, and 24 months. Detailed numerical results are posted for only a subset of these secondary outcomes. The original sample-size justification, prespecified statistical analysis plan, interim-analysis procedures, and detailed handling of missing measurements are not reported in the registry record.

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.

6 months (primary end point)

18 months (secondary endpoint)

24 months (end of study)

Analysis population for the primary outcome, as defined in the registry: Only patients with evaluable IHC for gamma H2AX staining for DNA damage at both time points are included in this analysis. A central pathology review of pre- and post- intervention biopsy specimens was performed at Vanderbilt University for all DNA damage endpoints.

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 Gastric Epithelial Cell DNA Damage, at 6 Months Versus Baseline, measured over Baseline up to 6 months.

The primary endpoint is the mean change from baseline to 6 months in gastric epithelial cell DNA damage, compared between the eflornithine and placebo groups. DNA damage is assessed using immunohistochemistry for gamma H2AX, and the measure is the percentage of positive gastric epithelial cells. The registry describes the reported quantity as a mean and the arm-level spread as a standard deviation.

This is a tissue-based biomarker endpoint. In general, a biomarker can be used to assess a biological process that may be relevant to a disease, while remaining distinct from a clinical outcome such as diagnosis, symptoms, or survival. The registry record does not establish that this measure predicts later disease or that a particular value represents a clinically meaningful change.

The endpoint is assessed over a defined interval, from baseline up to 6 months, and the comparison is between randomized groups. For the primary analysis, only participants with evaluable gamma H2AX staining at both measurement points are included. The registry also reports central pathology review of the biopsy specimens for DNA-damage endpoints. Those details matter because the comparison describes measured biomarker changes among the evaluable participants, not necessarily outcomes for every enrolled participant.

7. Results Posted in the Registry

• Eflornithine: mean 4.93 versus 4.31 on control; difference 0.62 (95% CI -9.36 to 10.60), standardized effect d = 0.03; z-test p = 0.903.

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.

Gastric Epithelial Cell DNA Damage, at 18 Months Versus Baseline

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Eflornithine Follow up 18 Months: mean 6.30 versus -0.20 on control; difference 6.50 (95% CI -4.18 to 17.18), standardized effect d = 0.31; z-test p = 0.233.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Eflornithine Follow up 18 Months and the control. The interval gives the range of effects that remain compatible with the data.

Gastric Epithelial Cell DNA Damage, at 24 Months Versus Baseline

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Eflornithine: mean -5.78 versus -0.13 on control; difference -5.65 (95% CI -22.13 to 10.83), standardized effect d = -0.19; z-test p = 0.502.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Eflornithine and the control. The interval gives the range of effects that remain compatible with the data.

Change in Gastric Histopathology Score, at 6 Months Versus Baseline

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Eflornithine Follow up 6 Months: mean -0.14 versus -0.18 on control; difference 0.04 (95% CI -0.22 to 0.30), standardized effect d = 0.07; z-test p = 0.762.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Eflornithine Follow up 6 Months and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes extend the DNA-damage assessment to later time points and include gastric histopathology or histology scores. The numerical comparisons posted for a subset of these outcomes do not show a consistent direction across time and measure. Their recalculated confidence intervals include zero, so the estimates remain compatible with no between-group difference as well as differences in either direction. This broad pattern does not provide a clear, consistent signal that changes the reading of the primary comparison.

The registry lists five secondary outcomes, but numerical results for the 18- and 24-month histology outcomes are not reported in the registry record. The secondary numerical results that are posted should not be treated as a complete account of every listed endpoint. As with the primary endpoint, the later DNA-damage analyses concern participants with observed measurements; specific analysis-population details for each secondary endpoint are not reported in the registry record.

In general, testing multiple outcomes or time points creates more opportunities for chance findings. An unadjusted p-value for one secondary comparison does not account for the number of other comparisons examined. The registry does not state whether a multiplicity adjustment was planned or applied. Secondary findings therefore carry less confirmatory weight than a prespecified primary result unless the analysis plan and interpretation support otherwise. They are useful as additional context, but should not be treated as independent proof of a treatment effect or as a replacement for the primary comparison.

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

• Gastric Epithelial Cell DNA Damage, at 18 Months Versus Baseline
• Gastric Epithelial Cell DNA Damage, at 24 Months Versus Baseline
• Change in Gastric Histopathology Score, at 6 Months Versus Baseline
• Change in Gastric Histology Score, at 18 Months Versus Baseline
• Change in Gastric Histology Score, at 24 Months Versus Baseline

9. Statistical Interpretation

10. Statistical Methods Explained

What is being compared?

The primary analysis compares the average change in a continuous measurement between two groups. The endpoint is the percentage of gamma H2AX-positive gastric epithelial cells, assessed at baseline and 6 months. The reported arm values are 4.93 for eflornithine and 4.31 for placebo; the comparison is expressed as the eflornithine mean minus the placebo mean. The recalculated estimate is 0.62.

In general, a mean summarizes the center of a set of measurements, while a standard deviation describes how dispersed individual measurements are around that mean. The reported standard deviations are 21.38 for eflornithine and 19.5 for placebo. A mean difference describes the contrast in group averages on the original measurement scale. It does not show the full range of individual responses.

How should the p-value be read?

The registry reports a two-sided t-test for the primary comparison, with p = 0.904. A two-sided test considers departures in either direction rather than only a difference favoring one group. In general, a p-value describes how compatible the observed data are with a specified no-difference model under the assumptions of the test. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect.

The recalculation from the posted summary statistics gives p = 0.903. The recalculated and registry-reported values are close but not identical; the reason for the difference is not reported in the registry record. The registry's reported analysis is identified as a two-sided t-test, but details beyond that label, such as the handling of any deviations from assumptions, are not reported.

What does the confidence interval add?

The recalculated 95% confidence interval for the mean difference is -9.36 to 10.6. In general, a confidence interval gives a range of effect estimates compatible with the observed data and the method used to calculate the interval. Its width indicates precision: a wider interval leaves more uncertainty about the size and direction of the contrast. Including zero means the interval includes no difference as one possibility; it does not prove that there is no difference.

The interval is recalculated from posted summary statistics and is not identified as the registry sponsor's confidence interval. No confidence interval for the sponsor's own analysis is reported in the registry record. The recalculation is an unadjusted comparison, not a substitute for a fuller analysis using participant-level data or any prespecified model that may have been used.

Why do the analysis population and measurement process matter?

The registry specifies that the primary analysis includes only participants with evaluable gamma H2AX staining at both baseline and the follow-up time point. This is a complete-measurement requirement for the endpoint. In general, excluding people without both measurements can affect interpretation if missing measurements are related to participant characteristics, treatment assignment, or outcomes. The registry record does not report the reasons for missing measurements, a comparison of included and excluded participants, or an imputation method.

The registry reports central pathology review of the pre- and post-intervention biopsy specimens for DNA-damage endpoints. Central review describes a feature of measurement assessment, but the record does not give additional detail about reader agreement or quality-control procedures. The statistical analysis should therefore be interpreted alongside the stated endpoint definition and analysis population, without assuming unreported procedures.

What is and is not known about the analysis plan?

The posted primary analysis uses a two-sided t-test. The record does not report the original sample-size justification, the planned effect size, or whether the test was the sole primary analysis in the original protocol. It also does not report an adjustment for multiple comparisons, interim analyses, covariate adjustment, or a detailed missing-data strategy. Those elements cannot be inferred from the posted results.

In general, a statistical analysis plan specifies how outcomes, analysis populations, missing data, and multiple testing will be handled before results are examined. Such detail helps readers distinguish planned confirmatory analyses from exploratory ones. Since the relevant plan is not reported in the registry record, interpretation should remain close to the results and methods that are explicitly posted.

11. Confidence Intervals and Precision

The recalculated primary mean difference is 0.62, with a 95% confidence interval from -9.36 to 10.6. The point estimate is the single best estimate from this calculation, while the interval conveys how uncertain that estimate is under the calculation method. Its range includes zero and also allows for contrasts in either direction. The result therefore does not pin down the size or direction of the underlying difference with high precision.

In general, a confidence interval is not a guarantee that the true value lies within the displayed range. It is a measure of uncertainty based on the data and statistical method. A wide interval can reflect limited information, substantial variability, or both. The primary endpoint standard deviations, 21.38 and 19.5, indicate substantial spread in the measured values relative to the estimated mean difference. The primary analysis also includes only participants with evaluable staining at both time points.

The registry reports p = 0.904 for its own two-sided t-test, while the unadjusted recalculation reports p = 0.903. The confidence interval is recalculated from posted summary statistics; an interval from the sponsor's analysis is not reported in the registry record. The interval should be considered together with the outcome definition and analysis population, not as a clinical threshold. The registry does not report what magnitude of biomarker change would be clinically important, so the interval alone cannot determine the practical importance of the estimated difference.

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 sample-size illustrations describe participants per arm, not total events. They use the observed effect as a starting point and show planning for 100%, 75%, and 50% of that observed effect. For the primary comparison, the illustrations are 17,023 and 22,789 participants per arm for 80% and 90% power when planning for 100% of the observed effect; 30,263 and 40,514 when planning for 75%; and 68,092 and 91,155 when planning for 50%.

The listed 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. These are planning illustrations computed from the observed effect, not the trial's original design assumptions; those original assumptions are not reported in the registry record. They should not be read as the number of participants enrolled or analyzed.

In general, a power target describes the chance of detecting an effect of a specified size under the planning assumptions if that effect is present. The 80% column gives the planned number per arm for an 80% target, and the 90% column gives the planned number per arm for a 90% target. A higher power target generally requires more participants under otherwise identical assumptions. Likewise, planning for a smaller effect requires more participants, because a smaller group contrast is harder to distinguish from random variation. These values are not observed power and do not measure the probability that the trial's result is correct.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.620; pooled standard deviation σ = 20.417.
2. n = 2 × (1.96 + 0.842)² × 20.417² / 0.620² = 17,022.8, rounded up to 17,023 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
delta <- 0.62                      # observed difference in means
sigma <- 20.4168394550512                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(2794428)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 17,023 (80% power), 22,789 (90% power)
 75% of observed effect: n per arm = 30,263 (80% power), 40,514 (90% power)
 50% of observed effect: n per arm = 68,092 (80% power), 91,155 (90% power)
power.t.test: 17023.8 
Simulated power at n = 17,023 per arm: 0.799 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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 registry's own primary analysis is a two-sided t-test and reports p = 0.904. The unadjusted recalculation from the posted group summaries gives a mean difference of 0.62, a 95% confidence interval from -9.36 to 10.6, and p = 0.903. The two p-values are close but not identical. The reason for the difference is not reported in the registry record, so it should not be attributed to a particular rounding rule or analysis choice.

The recalculated comparison is a simple analysis of posted summary statistics. It does not reproduce an analysis using participant-level data, and it cannot establish whether the sponsor's analysis used any additional procedures beyond the two-sided t-test named in the record. The registry does not report the sponsor's confidence interval or further details of the primary test. The recalculated interval is therefore useful for describing uncertainty from the posted summaries, while the registry p-value remains the reported result of the sponsor's analysis.

Neither value gives a clinical interpretation by itself. The estimate and interval describe the measured biomarker contrast, and the registry does not report a clinically important difference threshold for this endpoint.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies one primary outcome and five secondary outcomes. The primary outcome is DNA damage at 6 months; secondary outcomes cover DNA damage at later follow-up and histology changes at several time points. The record reports six analyses overall, but a detailed statistical analysis plan is not reported in the registry record. In particular, it does not state whether the secondary comparisons were prespecified in a hierarchy, whether a multiplicity adjustment was used, or how any family of tests was defined.

In general, when multiple hypotheses are tested, the chance of at least one apparently notable result due to random variation can increase. Adjustments or a prespecified testing strategy can address that concern, but no such approach should be assumed. The posted secondary p-values are described as sponsor-reported two-sided t-tests, and the recalculated comparisons are unadjusted. Secondary results should consequently be interpreted as exploratory unless additional prespecified procedures are documented elsewhere in the registry record.

The analysis population for the primary outcome is restricted to participants with evaluable gamma H2AX staining at both baseline and 6 months. The endpoint-specific numbers are 31 in the eflornithine group and 34 in the placebo group. The record does not explain why other enrolled participants are not represented in this comparison, describe missingness patterns, or report whether missing data were imputed. It also does not report a sensitivity analysis testing how different missing-data assumptions might affect the result.

Those omissions limit assessment of how representative the analyzed participants are of everyone enrolled and how robust the estimate is to missing measurements. Detailed analysis populations and missing-data procedures for the secondary outcomes are not reported in the registry record. Interim analyses, stopping rules, and the original sample-size justification are also not reported, so their role in the study's design and interpretation cannot be assessed from the registry record.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: covid-19 (2.2%), diverticulitis (2.2%), hypertensive urgency (2.2%), appendicitis (2.2%).

For serious adverse events, the registry lists 45 participants at risk in the eflornithine group and 46 in the placebo group. Serious adverse events affected 2 participants in each group. Deaths are listed as 0 in both groups. The registry lists COVID-19, diverticulitis, hypertensive urgency, and appendicitis among the serious adverse event terms, each with a recorded value of 2.2%. The record does not assign these listed terms to a treatment group in the details.

For other adverse events, the record lists 23 affected participants among 45 at risk in the eflornithine group and 32 among 46 at risk in the placebo group. These counts describe the posted safety summaries; they do not establish that an event was caused by an intervention. The registry record does not provide further detail on event severity, duration, relationship to study treatment, or how events were collected. The counts should therefore be read as descriptive safety reporting rather than as a causal comparison.

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

The primary result is an estimate of the difference between groups in a tissue-based measure of DNA damage at 6 months. The recalculated mean difference is 0.62, with a 95% confidence interval from -9.36 to 10.6. The interval includes zero and extends in both directions. The registry's two-sided test reports p = 0.904. Taken together, these results do not establish a clear between-group difference in the primary biomarker measure, but they also do not show that the groups are equivalent.

The endpoint is not a direct measure of gastric cancer occurrence, and the primary comparison does not establish whether the biomarker predicts later clinical outcomes. The registry record does not report a clinically meaningful threshold for a change in the percentage of positive cells. It is therefore not possible to translate the estimated difference into a clinical benefit or harm from the posted primary result alone.

The analysis is limited to participants with evaluable measurements at both time points. The record does not explain the absence of other enrolled participants from this analysis or provide enough detail to assess how missing measurements might affect the estimate. The study's completed status and overall enrollment do not remove this uncertainty about the endpoint-specific analysis.

General interpretation of a clinical trial result depends on the prespecified question, the analysis population, the reliability of measurement, and the precision of the estimate. Here, the registry identifies randomization and double masking, and reports central pathology review for DNA-damage endpoints, but it does not report the original statistical analysis plan or a clinically important effect threshold. The result should be understood as an uncertain estimate for this biomarker and timeframe, not as a definitive conclusion about clinical outcomes or the effect of eflornithine in other populations.

18. Why This Trial Matters Statistically

This trial illustrates the difference between estimating a biomarker contrast and demonstrating a clinical effect. The primary endpoint measures the percentage of gamma H2AX-positive gastric epithelial cells, not cancer incidence or another direct clinical outcome. That distinction matters when interpreting a statistically analyzed result: a precise biomarker estimate would still need an established clinical interpretation before it could be treated as evidence about patient outcomes.

The posted primary mean difference is small relative to the reported spread of measurements, and its confidence interval is broad enough to include contrasts in both directions. This makes precision central to interpretation. A p-value alone cannot convey the range of effects compatible with the data, while an interval alone cannot define clinical importance without a relevant threshold.

The analysis also highlights the importance of the analysis population. Enrollment is 91, while the primary endpoint comparison includes participants with evaluable staining at both time points. Transparent reporting of reasons for missing measurements and methods for handling them would help readers assess the reach of the findings. The registry's incomplete detail on multiplicity, sample-size justification, and the full analysis plan likewise limits how confidently secondary comparisons can be interpreted.

Finally, randomized allocation and double masking are important design features, but they do not remove uncertainty created by measurement variability, missing data, or limited reporting. The record offers a useful example of how design labels, endpoint definitions, observed estimates, and uncertainty must be considered together.

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.

The primary endpoint comparison includes only participants with evaluable gamma H2AX staining at baseline and 6 months. The registry does not report the reasons that participants are not represented in that analysis, a comparison of included and excluded participants, or sensitivity analyses for missing data. This limits assessment of whether the endpoint analysis reflects the enrolled population.

The registry labels the study randomized and double-masked, but also lists the study model as single group while identifying eflornithine and placebo groups. The record does not explain this difference in design descriptors. Specific countries within the two high-risk Latin American regions and detailed eligibility criteria are also not reported in the registry record, limiting assessment of how broadly the population may represent other groups.

The primary endpoint is a tissue biomarker, and no clinically meaningful difference threshold or link to later clinical outcomes is reported. The posted primary confidence interval is wide and includes effects in either direction. The record does not report the original sample-size rationale, a complete statistical analysis plan, multiplicity procedures, or interim-analysis details. These omissions make it difficult to judge whether the analyses and planning were aligned with a prespecified confirmatory strategy.

Numerical results are not reported for all listed secondary outcomes. Safety summaries are descriptive, with limited detail about event severity, treatment relationship, and collection procedures. These limitations do not negate the posted results, but they narrow the conclusions that can be drawn from them.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Chemoprevention of Gastric Carcinogenesis — NCT02794428.
• Registry record last updated 2026-09-02; status last verified 2026-08.
• Results first posted 2024-04-05.
• 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

The registry record describes a completed phase 2 study of oral eflornithine versus placebo in people with gastric premalignant lesions. It lists randomized allocation, double masking, enrollment of 91, and a single-group study model, although it also identifies two intervention groups. The primary endpoint is mean change from baseline to 6 months in the percentage of gamma H2AX-positive gastric epithelial cells.

The primary comparison has a recalculated mean difference of 0.62 and a 95% confidence interval from -9.36 to 10.6; the registry reports p = 0.904 for its two-sided t-test. The interval is broad and includes zero. The analysis population requires evaluable measurements at both time points. The original sample-size rationale, detailed missing-data procedures, multiplicity approach, and full analysis plan are not reported in the registry record. The result is an uncertain biomarker comparison, not a direct estimate of a clinical outcome.