PET Scan Imaging in Assessing Response in Patients With Esophageal Cancer Receiving Combination Chemotherapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for PET Scan Imaging in Assessing Response in Patients With Esophageal Cancer Receiving Combination Chemotherapy (NCT01333033), a randomized phase 2 study. The primary outcome is Complete Pathological Response (pCR) of PET/CT Non-responders.

2. Clinical Question

The clinical question is conditional on the PET/CT response assessment: it concerns the subgroup classified as non-responders, rather than all randomized participants. The endpoint therefore addresses pathological findings in that subgroup after treatment, not the probability of PET/CT response itself. PET/CT response between treatment arms and pathological response among PET/CT responders were listed as secondary outcomes. Those outcomes provide related but distinct descriptions of the treatment pathway.

The registry record does not report a more detailed eligibility definition or the full clinical decision rules used after PET/CT assessment. The arm descriptions are incomplete in the record, so the full sequence of subsequent therapy cannot be reconstructed from it.

3. Trial Design

This was a randomized phase II trial with an open-label design: the registry lists random allocation and no masking. The registry model is crossover, and the total enrollment is 257. The sponsor is the Alliance for Clinical Trials in Oncology. The record lists the study as completed, with a start date of 2011-07 and primary completion in 2015-11.

The randomized induction approaches were a modified FOLFOX-6 regimen and carboplatin plus paclitaxel with radiation. Participants underwent PET/CT assessment after induction treatment. The trial’s clinical structure therefore linked the initial treatment assignment with imaging response and later pathological assessment. The registry lists PET and computed tomography procedures, radiation therapy, and the named chemotherapy drugs as interventions. Details of all treatment steps after response assessment are not fully available in the arm descriptions.

The primary analysis population was not simply everyone enrolled. It consisted of all eligible patients registered to treatment who reported no PET/CT response to induction treatment. The endpoint was the percentage in each induction group with a complete pathological response. The registry record gives 39 participants for the FOLFOX non-responder group and 50 for the CP non-responder group.

Randomization supports a comparison of the assigned induction approaches, but the primary endpoint conditions on a post-induction response category. Such subgroups can differ from the full randomized population and should be interpreted as the specified subgroup outcome, not as an overall randomized comparison. The original sample-size rationale, detailed randomization procedures, and trial-specific statistical plan 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All eligible patients that registered to treatment and reported no PET/CT response to induction treatment were included in this endpoint.

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: All participants who enrolled on this trial are included in this analysis.

6. Primary Endpoint

The registry lists the primary outcome as Complete Pathological Response (pCR) of PET/CT Non-responders, measured over Up to 5 years.

The primary endpoint is the percentage of PET/CT non-responders in each induction treatment group who had a complete pathological response, or pCR, at surgery. The registry defines pCR as no tumor found on pathology review in all resected lymph nodes and tissue, with no viable tumor in any sampled tissue. The timeframe is up to 5 years.

This endpoint uses pathology findings to characterize whether viable tumor was present in the tissue examined at surgery. It is distinct from the imaging response assessment: PET/CT determines which patients enter this primary endpoint population, while pathology determines whether the endpoint event occurred. The registry population includes eligible patients registered to treatment who reported no PET/CT response to induction treatment.

For this endpoint, the posted pCR percentage is 17.95% in the FOLFOX non-responder group, with a 95% confidence interval from 7.54% to 33.53%. In the CP non-responder group, it is 20.0%, with a 95% confidence interval from 10.03% to 33.72%. These estimates describe the observed proportions and their uncertainty in the defined subgroups. The endpoint does not by itself establish how the groups would compare in a larger population, nor does it describe outcomes for patients who had a PET/CT response.

In general, a pathological endpoint can provide a direct assessment of tumor in sampled surgical tissue. Its interpretation depends on the stated tissue and population definitions and on the availability and completeness of pathological assessment; further details on missing assessments are not reported in the registry record.

7. Results Posted in the Registry

• CP Non-Responder: 20.0% (10/50) versus 17.9% (7/39) on control; risk difference 2.1 percentage points (95% CI -14.3 to 18.4), risk ratio 1.11 (0.47–2.66); unadjusted two-proportion z-test p = 0.807.

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.

PET/CT Response Between Treatment Arms

• 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:
  • CP (Carboplatin + Paclitaxel + Radiation): 56.1% (64/114) versus 64.9% (72/111) on control; risk difference -8.7 percentage points (95% CI -21.4 to 4.0), risk ratio 0.87 (0.70–1.07); unadjusted two-proportion z-test p = 0.181.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between CP (Carboplatin + Paclitaxel + Radiation) and the control. The interval gives the range of effects that remain compatible with the data.

pCR Compared Between Induction Treatment Arms Among PET/CT Responders

• 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:
  • CP (Carboplatin + Paclitaxel + Radiation) Responder: 14.1% (9/64) versus 40.3% (29/72) on control; risk difference -26.2 percentage points (95% CI -40.4 to -12.0), risk ratio 0.35 (0.18–0.68); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for CP (Carboplatin + Paclitaxel + Radiation) Responder than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

pCR Compared Among Non-responders Between Induction Treatment Arms if Treatment Regimens Are Found to be Efficacious

• 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:
  • CP (Carboplatin + Paclitaxel + Radiation) Non-Responders: 20.0% (10/50) versus 17.9% (7/39) on control; risk difference 2.1 percentage points (95% CI -14.3 to 18.4), risk ratio 1.11 (0.47–2.66); unadjusted two-proportion z-test p = 0.807.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between CP (Carboplatin + Paclitaxel + Radiation) Non-Responders and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover PET/CT response between induction treatment groups, pCR among PET/CT responders, pCR among non-responders, and progression-free survival among non-responders within each induction group. Together, they describe different stages or subgroups in the PET-directed treatment pathway. They are not interchangeable endpoints: imaging response is a different measure from pathological response, and the primary result concerns only non-responders.

Across these outcomes, the broad pattern is that the observed PET/CT response percentages were higher in the FOLFOX group, and the observed pCR percentage among PET/CT responders was also higher in that group. In the primary non-responder population, the observed pCR percentages were close. This pattern is descriptive and should not be collapsed into a single overall treatment effect, because the endpoints concern different populations and outcomes.

The registry’s secondary-outcome listing includes progression-free survival among non-responders, but no corresponding result values are included in the posted secondary results described in the record. Further details about that outcome are therefore not reported in the registry record.

The recalculated secondary comparisons are unadjusted. General explanation: when several outcomes are examined, the chance of finding at least one apparently unusual result can increase, even if no underlying effects exist. Multiplicity methods can address this issue, but adjustment procedures and a hierarchy of confirmatory versus exploratory outcomes are not reported in the registry record. As a result, unadjusted secondary p-values carry less weight than a prespecified, multiplicity-controlled analysis. They are best treated as supportive descriptions that warrant cautious interpretation rather than definitive evidence.

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

• PET/CT Response Between Treatment Arms
• pCR Compared Between Induction Treatment Arms Among PET/CT Responders
• pCR Compared Among Non-responders Between Induction Treatment Arms if Treatment Regimens Are Found to be Efficacious
• Progression Free Survival (PFS) Among PET/CT Non-responders Within Each Induction Treatment Group

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary at the participant level: a person either had a pCR as defined by pathology review or did not. The displayed percentages are proportions with that event, with the denominator restricted to eligible registered patients who were PET/CT non-responders. General explanation: a binary endpoint is often summarized as an event proportion in each group, then as a difference or ratio between groups. The proportion describes frequency in the analyzed group; a comparison estimate describes the contrast between groups.

What does the risk difference express?

The recalculated risk difference is 0.021 for CP compared with FOLFOX. It is the CP event proportion minus the FOLFOX event proportion, expressed on the probability scale. A value of zero would correspond to equal observed proportions; positive values indicate a higher observed proportion in CP, and negative values indicate a higher observed proportion in FOLFOX. The 95% confidence interval is -0.143 to 0.184. Since that interval includes zero, the results do not clearly distinguish the observed contrast from no difference at the conventional zero point.

General explanation: a risk difference is directly interpretable as an absolute contrast in proportions. It does not describe a relative change. The estimate should be read together with its confidence interval, because the interval conveys how much uncertainty remains around the observed contrast.

What does a confidence interval convey?

A 95% confidence interval gives a range of values compatible with the data and the method used to calculate it, under repeated use of that procedure. It is not a guarantee that the true value lies in the particular interval, nor does it assign a probability to a treatment effect. Here, the interval extends from a difference favoring FOLFOX to a difference favoring CP. Its breadth is important: the observed near-similarity between percentages does not rule out meaningful differences in either direction.

The arm-specific 95% confidence intervals also show uncertainty around each percentage: 7.54% to 33.53% for FOLFOX and 10.03% to 33.72% for CP. Comparing overlap between separate arm intervals is not a substitute for reading the interval calculated for the between-group difference. The latter directly addresses uncertainty in the contrast.

How should the p-values be read?

The recalculated unadjusted p-value is 0.807. General explanation: a p-value measures how unusual the observed data, or data more extreme, would be under a specified null model and the chosen test. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. The registry’s reported Fisher Exact test gives p=1.0. The record does not provide further details explaining the distinction between the posted test result and the recalculated unadjusted result.

Fisher’s exact test is a method for comparing binary event counts between groups. It is based on the observed group counts and is commonly used for categorical data. The registry identifies Fisher Exact as the reported method for the primary analysis. The recalculated estimate and interval give additional scale and precision context, but are not the registry’s reported test result.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when observed variation is greater than a model assumes. If a model underestimates variation, its standard errors and intervals may be too small, making results look more precise than they are. This primary endpoint is a binary proportion, and the registry does not report an over-dispersion assessment or a model-based adjustment for it. The issue is relevant generally to model-based analyses, but the record does not support claims that over-dispersion was detected or affected this comparison.

What is the role of the subgroup definition?

The primary analysis is restricted to PET/CT non-responders, a status assessed after induction treatment. This is different from comparing all randomized participants according to their assigned group. General explanation: selecting a subgroup based on a post-randomization outcome can change the balance created by random assignment, because subgroup membership may relate to treatment and patient characteristics. The subgroup estimate remains meaningful as a description of the defined non-responder populations, but it should not automatically be generalized to all randomized patients.

The registry reports the endpoint population definition and arm-specific group sizes but does not report details of covariate adjustment, handling of missing pathology, or a sensitivity analysis for the subgroup. Those aspects of the statistical plan are not reported in the registry record.

11. Confidence Intervals and Precision

The primary arm estimates have 95% confidence intervals of 7.54% to 33.53% for FOLFOX non-responders and 10.03% to 33.72% for CP non-responders. Both ranges are broad relative to their point estimates. They indicate that the observed percentages are not highly precise and that a range of underlying event frequencies remains compatible with the data.

For the direct comparison, the recalculated risk difference is 0.021, with a 95% confidence interval from -0.143 to 0.184. This interval includes no difference and also allows for differences in either direction. The point estimate alone should therefore not be treated as a precise estimate of the between-group contrast. The uncertainty is particularly relevant because the primary population comprises 39 FOLFOX non-responders and 50 CP non-responders.

General explanation: precision is reflected by the spread of an estimate’s confidence interval. Greater precision means a narrower interval, while a wider interval leaves more plausible values unresolved. Precision is not the same as validity: a narrow interval can still be affected by bias or analysis choices, and a wide interval does not prove that groups are equivalent.

The registry’s Fisher Exact result is p=1.0 but has no accompanying confidence interval or effect estimate in the record. The recalculated interval helps describe uncertainty in the posted proportions, while remaining a separate unadjusted analysis. No equivalence margin or non-inferiority threshold is reported, so the interval cannot be interpreted as establishing equivalence between the treatment groups.

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 planning figures are participants per arm, not total events. They are illustrations based on the observed effect, rather than the trial’s original design calculations. The stated 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 to the stated fraction of the observed effect. The original trial’s sample-size assumptions and justification are not reported in the registry record.

The 80% power column gives the participants per arm indicated by those assumptions for an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column gives the corresponding planning figure for a 90% chance. These are design concepts for a hypothetical study, not descriptions of what happened in this trial. They are not observed power.

For the primary non-responder comparison, planning for 100% of the observed effect gives 5,735 participants per arm for 80% power and 7,677 for 90% power. Planning for 75% of the observed effect gives 10,090 participants per arm for 80% power and 13,507 for 90% power. Planning for 50% of the observed effect gives 22,463 participants per arm for 80% power and 30,071 for 90% power.

As the effect targeted for detection becomes smaller, more participants are needed under the same assumptions. These figures are sensitive to the observed effect used as their starting point, so they should not be read as a recommended design or as evidence that the original study was adequately or inadequately sized.

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 = 0.179; treatment proportion p₁ = 0.200; average p̄ = 0.190.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5545 = 1.0868.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5543 = 0.4665.
4. n = (1.0868 + 0.4665)² / (0.0205)² = 5,734.4, rounded up to 5,735 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 <- 0.179487179487179                          # control proportion
p1 <- 0.2                          # 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(1333033)
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 = 5,735 (80% power), 7,677 (90% power)
 75% of observed effect: n per arm = 10,090 (80% power), 13,507 (90% power)
 50% of observed effect: n per arm = 22,463 (80% power), 30,071 (90% power)
power.prop.test: 5734.4 
Simulated power at n = 5,735 per arm: 0.797 (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 registry reports a Fisher Exact test for the primary comparison between FOLFOX and CP non-responders, with p=1.0. It does not report an effect estimate or confidence interval for that test. The recalculated unadjusted comparison gives a risk difference of 0.021, a 95% confidence interval from -0.143 to 0.184, and p=0.807.

These are distinct presentations of the primary comparison: one is the test result reported in the registry, and the other is an unadjusted recalculation from the posted arm summaries. The registry record does not give details that explain the difference between the two p-values, so no explanation should be inferred from the record alone. The discrepancy flag is false. The recalculated result should therefore be read as supplementary context, not as a correction to the registry’s reported analysis.

Neither p-value conveys the size of the effect or its precision by itself. The recalculated risk difference and interval make the scale and uncertainty of the contrast explicit, while the registry analysis records the sponsor-reported test result. The statistical plan underlying the sponsor’s analysis is not reported in the registry record.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple secondary outcomes in addition to the primary endpoint. It does not report whether the analyses were arranged in a prespecified hierarchy, whether multiplicity was controlled, or whether any secondary outcomes were designated exploratory. Adjustments for multiple comparisons are not reported in the registry record. Accordingly, the secondary p-values should be interpreted as unadjusted results, not as conclusions protected against the additional uncertainty associated with examining several outcomes.

The primary population is defined as eligible patients registered to treatment who reported no PET/CT response to induction treatment. The registry also gives separate populations for PET/CT responders and reports group-specific counts for those outcomes. However, it does not provide a complete account of analysis populations for every outcome, nor does it describe how missing PET/CT assessments, pathology results, or other outcome data were handled. Missing-data methods are not reported in the registry record.

Details of the original statistical analysis plan, including model specifications and any planned sensitivity analyses, are not reported in the registry record. Interim analyses and stopping rules are also not reported. These absences do not establish that such procedures were or were not used; they mean that the registry record does not document them.

General explanation: a clear analysis plan helps readers distinguish prespecified tests from analyses selected after results are known. A transparent account of missing data also helps establish who contributed to each estimate and whether exclusions could affect interpretation. Without those details, the posted group estimates can be described as recorded, but the effect of analysis choices and incomplete outcome information cannot be fully assessed.

Finally, the primary endpoint conditions on PET/CT non-response after randomized induction treatment. That definition is explicit, but the record does not report further analyses of how post-randomization selection may influence between-group comparisons. The primary finding should therefore be interpreted as a comparison within the stated subgroup, rather than as a direct estimate for all randomized participants.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cardiac arrest (2.6%), esophageal perforation (2.0%), dyspnea (2.0%), respiratory failure (2.0%), esophageal anastomotic leak (1.6%).

The registry lists serious adverse event counts by response and treatment group. Among FOLFOX responders, 3 of 72 were reported with a serious adverse event; among FOLFOX non-responders, 1 of 39. Among CP responders, 2 of 64 were reported; among CP non-responders, 4 of 49. The no-crossover groups are listed separately: 0 of 9 in CP No Cross-over and 0 of 5 in FOLFOX No Cross-over. These are group-specific counts and denominators, not a randomized comparison of safety.

The registry’s listed serious-event terms include cardiac arrest at 2.6%, esophageal perforation at 2.0%, dyspnea at 2.0%, respiratory failure at 2.0%, esophageal anastomotic leak at 1.6%, and thromboembolic event at 1.6%. The record does not provide further context for interpreting those percentages, such as event-specific group breakdowns in the information.

General explanation: serious adverse event counts describe events recorded under the registry’s reporting definitions. They do not by themselves establish causation, comparative risk, or the full tolerability profile. Differences in group size and subgroup formation also matter when reading raw counts. Detailed safety-analysis methods and attribution information are not reported in the registry record.

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

The primary result describes pCR percentages among patients classified as PET/CT non-responders after induction treatment. The observed percentages are 17.95% for FOLFOX and 20.0% for CP. The recalculated difference is small in the observed data, but its confidence interval is broad and includes differences in either direction. The record therefore leaves substantial uncertainty about the size and direction of the contrast in this subgroup.

This result does not establish that the treatments are equivalent. Equivalence requires a prespecified margin and an analysis designed to assess whether the interval lies within that margin; no such margin is reported in the registry record. Nor does the result establish that either approach is superior for the primary endpoint. The reported Fisher Exact p-value is p=1.0, while the unadjusted recalculated p-value is 0.807; neither is a measure of clinical importance or a probability that a treatment effect is real.

The primary endpoint is conditional on PET/CT non-response. It cannot be used alone to describe pCR for all randomized participants or to infer what would happen among PET/CT responders. The secondary outcomes address some of those other groups, but they have separate denominators and are unadjusted. Their results should not be merged with the primary result into a single conclusion.

General explanation: a trial result is most informative when the endpoint, population, estimate, uncertainty, and analysis are read together. Here, the endpoint is a pathology-based response measure, the population is a post-induction subgroup, and the interval indicates limited precision. Further conclusions about long-term outcomes, survival, treatment choice, or individual care are not established by this endpoint. Progression-free survival results are not included in the posted secondary results described in the registry record.

18. Why This Trial Matters Statistically

This trial illustrates how a biomarker or imaging assessment can define the population for a later clinical endpoint. PET/CT response is not the primary outcome: it determines which participants enter the primary non-responder analysis, while pathology at surgery determines whether pCR occurred. Keeping those stages separate is essential for interpreting both the estimand and the denominator.

It also demonstrates why a point estimate and its uncertainty should be considered together. The two primary subgroup percentages are close, but the confidence interval for their difference is wide. A small observed contrast is not the same as precise evidence of no difference. The interval retains uncertainty in both directions, and the registry does not report an equivalence margin that would support an equivalence conclusion.

The trial also provides an example of distinct inferential roles for sponsor-reported and recalculated statistics. The registry reports a Fisher Exact p-value, while the unadjusted recalculation supplies an effect estimate and interval from the posted summaries. Reporting both transparently helps readers distinguish the official analysis from an additional descriptive calculation without treating either p-value as a measure of clinical importance.

Finally, the secondary results span imaging and pathology outcomes in response-defined subgroups. Their interpretation depends on population definitions and multiplicity, not simply on whether an individual p-value appears small. This is a useful reminder that a trial with several related endpoints can generate a pattern of results that requires careful attention to prespecification, precision, and the limits of the registry record.

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 primary analysis is restricted to PET/CT non-responders, a subgroup identified after induction treatment. This makes it different from an analysis of all randomized participants and limits how directly its comparison can be generalized to the full trial population. The record does not report additional analyses addressing the effects of this subgroup selection.

The non-responder group sizes are 39 for FOLFOX and 50 for CP, and the confidence interval for the recalculated risk difference is broad. The resulting estimate is imprecise. The registry does not report a clinically important difference threshold or an equivalence margin, so the near similarity of the observed percentages cannot be interpreted as proof of equivalence.

The arm descriptions are truncated, and the registry model is listed as crossover without a complete account of the treatment sequence after PET/CT assessment. The details of treatment received after response classification are therefore not fully reported in the registry record. Eligibility criteria beyond the endpoint population description, missing-data procedures, covariate adjustments, sensitivity analyses, and full statistical methods are also not reported.

Several secondary outcomes are listed, but multiplicity handling and outcome hierarchy are not reported. The progression-free survival outcome is listed without result values in the posted secondary results described in the record. Safety counts are given by response groups, and the registry listing does not establish a causal or comparative interpretation of those counts. These limitations constrain what can be concluded from the posted results, especially beyond the defined subgroup and endpoint.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: PET Scan Imaging in Assessing Response in Patients With Esophageal Cancer Receiving Combination Chemotherapy — NCT01333033.
• Registry record last updated 2023-04-13; status last verified 2023-04.
• Results first posted 2021-11-03.
• 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

NCT01333033 is a completed randomized phase II trial of PET scan-directed combined modality therapy in esophageal cancer. It enrolled 257 participants and compared induction FOLFOX with carboplatin, paclitaxel, and radiation, using PET/CT response as part of the treatment pathway.

The primary endpoint was pCR among PET/CT non-responders. The posted percentages were 17.95% for FOLFOX and 20.0% for CP. The registry reports a Fisher Exact p-value of 1.0; the unadjusted recalculation gives a risk difference of 0.021, with a 95% confidence interval from -0.143 to 0.184 and p=0.807. The interval indicates considerable uncertainty about the subgroup contrast.

Secondary results address PET/CT response and pCR among responders and non-responders. Their interpretation is limited by subgroup definitions, unadjusted analyses, and missing detail on multiplicity and analysis procedures. The registry record does not report the original sample-size rationale or full statistical plan.