Radiation Dose Escalation in Esophageal Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Radiation Dose Escalation in Esophageal Cancer (NCT03082586), a non-randomized na study. The primary outcome is Number of Dose Limiting Toxicities (DLTs) Occurring in Participants.

2. Clinical Question

This question is about identifying a dose level that can be evaluated for tolerability within the study’s treatment setting. The primary endpoint focuses on specified toxicities associated with protocol treatment during and after the high-fractionated-radiation-therapy boost. It is not a randomized comparison of the combined regimen against another treatment.

The registry records enrollment of 31 and lists six radiation-dose groups, from 57.2 Gy through 93.2 Gy. That enrollment figure is not a report of how many participants contributed to the primary endpoint at each dose level. Group-level endpoint counts and analysis populations are not reported in the registry record.

3. Trial Design

This is a completed phase I trial with a non-randomized allocation and no masking. The model is listed as sequential. Those design features fit the stated aim of escalating radiation dose while observing dose-limiting toxicities, but the registry record does not provide further operational details about how escalation decisions were made or how the maximum tolerable dose was defined.

The six listed radiation regimens use doses of 57.2 Gy, 64.4 Gy, 71.6 Gy, 78.8 Gy, 86 Gy, and 93.2 Gy. The official title identifies concurrent weekly cisplatin and paclitaxel, while the brief summary names DDP and paclitaxel. The registry lists radiation interventions and describes each arm by its radiation dose. It does not give arm-specific chemotherapy delivery details in the record summarized for this trial.

Enrollment is recorded as 31. The registry dates the study start to 2016-11-20 and primary completion to 2018-02-20, and lists the status as completed. These dates describe the recorded study timeline; they do not establish when each participant was treated or followed.

Sequential dose escalation is not equivalent to random assignment. Participants at different dose levels may have entered at different points in the study, and the design does not establish that the groups were exchangeable. Consequently, observed differences between dose levels would not automatically estimate causal differences attributable to radiation dose alone. The registry itself cautions that the arms are not a like-for-like control comparison.

The phase is listed as NA. A detailed escalation algorithm, dose-cohort decision rules, protocol deviations, and the statistical plan are not reported in the registry record. The record also does not identify any interim analysis procedure or sample-size justification. No such details should be inferred from the sequential model alone.

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.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Dose Limiting Toxicities (DLTs) Occurring in Participants, measured over 15 months.

The primary endpoint is the number of dose-limiting toxicities occurring in participants, with a stated timeframe of 15 months. A DLT is a treatment-tolerability measure: the endpoint is intended to capture specified serious toxicities that may constrain how far a treatment dose can be increased. In a dose-escalation study, this kind of endpoint addresses the trial’s stated aim of finding a maximum tolerable radiation dose when radiation is combined with chemotherapy.

The registry defines DLTs as grade >=4 esophatitis, any other grade >=3 nonhematological toxicity except nausea and vomiting, or grade >=4 hematological toxicity lasting more than 7 days. The toxicity must be possibly, probably, or definitely associated with protocol treatment, and must occur during or after completion of the HFRT boost treatment. The registry wording is retained; it does not provide additional adjudication procedures or grading guidance.

The endpoint is a count, not a time-to-event measure or a continuous measure of symptom severity. Its interpretation depends on the DLT definition, the observation period, and which participants were assessed. The registry record does not report observed DLT counts by dose group, the number assessed for DLTs, or a maximum tolerable dose determination. It therefore does not support a quantitative conclusion about the primary endpoint.

DLTs are also not interchangeable with all adverse events or with serious adverse events. A toxicity can meet a DLT definition through its grade, type, duration, and treatment attribution criteria; the registry’s adverse-event summaries are separately reported safety information.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The available design information supports a limited statistical description. This is a sequential, non-randomized phase I study with no masking. Its stated purpose is to explore the maximum tolerable radiation dose alongside chemotherapy, and its primary endpoint is a count of DLTs. The registry does not report an inferential analysis of that endpoint, so the methods below explain how the design and endpoint would generally be interpreted without attributing an unreported analysis to this trial.

What does sequential escalation imply?

In a sequential dose-escalation design, the dose levels are studied in an ordered process rather than assigned through randomization. The data at a dose level can inform decisions about how to proceed, depending on a protocol’s escalation rules. The registry identifies the model as sequential but does not report those rules, the number of participants planned at each level, or the decisions made during the study. It is therefore not possible to reconstruct the escalation path from the registry record.

Randomization is often used to balance measured and unmeasured participant characteristics between groups. It was not used. Comparisons among the listed doses are consequently descriptive and should not be read as randomized estimates of what would have happened if otherwise comparable participants had received a different dose. The registry cautions that the arms are not a like-for-like control comparison.

Why is the endpoint a count?

The primary endpoint records the number of DLTs, with a definition based on toxicity type, grade, duration in the hematological criterion, treatment association, and timing relative to the HFRT boost. Count outcomes describe how many qualifying events were observed, but their interpretation depends on the amount of observation and the number of participants evaluated. The record gives an overall enrollment figure but does not report the number assessed for this endpoint at each dose level or the DLT counts themselves.

For count data, statistical methods may need to account for variation that is larger than a simple model assumes. In general, over-dispersion means that observations vary more than expected under a basic count model; ignoring it can make uncertainty estimates too small. This is a general consideration, not a claim about the data or the analysis used in this trial. The registry record does not report a count model, dispersion estimate, or related diagnostic.

What can uncertainty estimates tell readers?

In general, a confidence interval conveys the uncertainty around an estimate under a specified model and analysis. For an event count or rate, the interval would help readers judge how precise the estimate is, rather than treating the observed value as exact. No primary endpoint estimate or interval is reported in the registry record, so precision cannot be assessed for DLT findings.

Hypothesis tests and p-values address compatibility between data and a specified null model, conditional on the assumptions of the analysis. They do not by themselves establish that a treatment is clinically beneficial or that a dose is tolerable. The record lists zero analyses and gives no primary endpoint test, p-value, or statistical model. No significance finding should therefore be inferred.

How should planning illustrations be read?

The registry describes sample-size calculations as planning illustrations based on the observed effect, not as the trial’s own design. The stated assumptions include a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and detection of a stated fraction of the observed effect. These assumptions describe hypothetical planning scenarios. They do not mean that this sequential, non-randomized trial used equal allocation, and they do not establish its observed power.

Sample-size calculations are conditional on their assumptions about the effect, variability, significance level, allocation, and analysis. In general, planning for a smaller effect requires more information than planning for a larger effect, all else being equal. That general principle does not provide a numerical sample size for this trial. The registry record does not report the numerical outputs of the illustrations or the observed effect on which they were based. It specifies that time-to-event illustrations use the total number of events under Schoenfeld, while other illustrations use participants per arm.

What analysis details are absent?

The statistical analysis plan, analysis populations, missing-data methods, endpoint adjudication process, and handling of participants who did not complete follow-up are not reported in the registry record. The record also does not report interim analyses or a sample-size justification for the trial itself. These omissions prevent an assessment of whether the reported endpoint—if analyzed outside this record—was handled according to a prespecified plan. They do not establish that no such plan existed; its contents are simply not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists local failure as a secondary outcome and records two outcome measures. It does not report the corresponding results or explain how the primary and secondary outcomes were incorporated into a statistical testing strategy. A multiplicity adjustment, if used, would address the increased chance of false-positive findings when several hypotheses are tested. Whether an adjustment was planned or applied is not reported in the registry record.

The record lists zero analyses and provides no primary endpoint estimate, test, or interval. It does not describe an analysis population, such as all enrolled participants or only those who received treatment. The number assessed for the primary endpoint at each dose level is also not reported. It is therefore not possible to determine from this record whether endpoint data were analyzed for all enrolled participants or how participants were assigned to an analysis population.

Missing-data handling is not reported in the registry record. There is no description of how incomplete toxicity follow-up, missing assessments, or withdrawal from treatment affected endpoint counting. This matters because the chance to observe a DLT can depend on the duration and completeness of follow-up. The stated primary timeframe is 15 months, but the registry does not provide an account of participant-level follow-up completeness or missing endpoint data.

The DLT definition includes judgments about whether a toxicity was possibly, probably, or definitely associated with protocol treatment. The registry record does not report who made those judgments, whether they were reviewed independently, or how uncertain attribution was handled. These details are relevant to consistent endpoint classification, but no particular procedure should be assumed from the endpoint definition alone.

11. 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: grade 3 pneumotitis (33.3%), grade 5 esophageal fistula (33.3%).

The safety summary lists serious adverse events and other adverse events separately by arm. For serious adverse events, the affected counts are 0 in Arms 1 through 5 and 2 in Arm 6; the listed numbers at risk are 3, 3, 3, 6, 13, and 3, respectively. The registry lists 0 deaths in Arms 1 through 5 and 1 in Arm 6. For other adverse events, the affected counts are 3, 3, 3, 5, 11, and 1, with the same arm-specific numbers at risk.

The record lists grade 3 pneumotitis and grade 5 esophageal fistula among the leading serious-event terms, each at 33.3%. These percentages are registry entries; the record does not provide further event-level context in the registry record reported for this trial. The event terms and counts should not be treated as DLT counts, because the primary endpoint has its own grade, timing, duration, and treatment-association criteria.

These are descriptive arm-level safety summaries, not randomized comparisons. The arms have different listed numbers at risk and were not formed by random allocation. The figures alone do not establish that one dose caused more or fewer events than another, nor do they establish the primary endpoint result. Further details about event timing, participant characteristics, follow-up completeness, and attribution are not reported in the registry record.

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

The record establishes the study’s stated purpose, design, dose levels, and primary endpoint definition. It documents a completed sequential, non-randomized phase I trial of radiation dose escalation with concurrent weekly cisplatin and paclitaxel in patients with inoperable or medically unresectable esophageal cancer. It also records enrollment of 31 and provides descriptive adverse-event summaries by arm.

The record does not establish which radiation dose was the maximum tolerable dose. It does not report DLT counts, a dose-response analysis, or the analysis used to make an escalation decision. The existence of six listed dose groups does not itself show that each level was reached under the same conditions or that the highest listed dose was judged tolerable. Those conclusions require results and decision rules that are not reported in the registry record.

The record also does not establish the comparative effect of one radiation dose versus another. Since allocation was non-randomized and sequential, the arms are not randomized control groups. Any apparent variation across dose levels would need cautious interpretation and could not, by itself, isolate the effect of dose from other differences between groups.

The listed serious adverse events are safety observations, not proof of a treatment effect and not a substitute for the primary DLT endpoint. Conversely, the lack of reported DLT results does not show that no DLTs occurred. Likewise, the absence of reported local-failure results does not establish whether local failure occurred or whether groups differed on that outcome.

Overall, the registry record supports a description of the question and study structure, but not a conclusion about the primary result, tolerability at a particular dose, or clinical benefit. These boundaries reflect which outcomes and analyses are reported in the registry record, not a judgment about results that may exist elsewhere.

13. Why This Trial Matters Statistically

This trial is statistically notable because its central question is dose selection, not a conventional randomized comparison of treatment efficacy. The sequence of radiation doses and the DLT endpoint are aligned with the stated aim of identifying a maximum tolerable dose in combination with chemotherapy. The statistical interpretation therefore depends on dose-specific toxicity observations and the rules used to escalate or stop. Although the registry identifies a sequential model, it does not provide those rules or the observed primary endpoint data.

The record also illustrates why the design label matters when reading group summaries. The radiation-dose groups are not randomized controls, and the registry cautions that they are not like-for-like comparisons. A trend across dose groups, if reported, would not automatically be a causal dose effect. The sequence in which participants entered and the characteristics of participants at different dose levels would be relevant, but are not reported in the registry record.

A second statistical issue is the distinction between the primary DLT endpoint and adverse-event summaries. Both concern safety, but the endpoint uses specific criteria for toxicity grade, type, duration, treatment association, and timing. Serious adverse-event counts cannot be substituted for the endpoint’s DLT count without evidence that each event meets those criteria.

Finally, this record separates study design from evidence of a result. It describes a phase I investigation, lists dose levels and enrollment, and gives safety summaries, but reports zero analyses and no primary endpoint findings. Readers can identify the intended statistical question while also recognizing that the registry record does not permit an estimate of the answer or its precision.

14. 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 main limitation for interpreting this trial from the registry record is the absence of reported primary endpoint results. DLT counts by dose, the number assessed at each dose, and the maximum tolerable dose conclusion are not reported. Without those results, the central dose-selection question cannot be answered statistically from this record.

The sequential, non-randomized design limits interpretation of comparisons across dose levels. The registry explicitly cautions that the arms are not like-for-like control comparisons. The record does not describe the escalation decision rules, timing of enrollment by dose, or participant characteristics by group. Those details would be needed to understand the path through the dose levels and evaluate group differences.

The registry record does not report the trial’s statistical analysis plan, analysis populations, missing-data procedures, interim analyses, or sample-size justification. It also does not report the method for determining treatment association for DLT classification. These omissions limit evaluation of prespecification and endpoint implementation; they should not be taken as evidence that such procedures were absent from the underlying study.

Enrollment of 31 is reported overall, but the record does not provide a dose-specific accounting of participants evaluated for the primary endpoint. The adverse-event summaries provide arm-level counts but do not replace that missing endpoint information. Follow-up completeness for the stated 15-month primary timeframe is not reported in the registry record.

Local failure is listed as a secondary outcome, but its results are not reported. The record therefore cannot support interpretation of that outcome alongside toxicity. The available safety terms and percentages also lack detailed event-level context in the registry record, so they should be read descriptively rather than as comparative estimates.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Radiation Dose Escalation in Esophageal Cancer — NCT03082586.
• Registry record last updated 2020-12-22; status last verified 2020-06.
• Results first posted 2020-07-07.
• 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.

18. Record Summary

The registry record describes a completed phase I, sequential, non-randomized trial of radiation dose escalation with concurrent weekly cisplatin and paclitaxel for esophageal cancer. The stated population is patients with inoperable or medically unresectable disease. Six radiation doses are listed, from 57.2 Gy to 93.2 Gy, and enrollment is recorded as 31.

The primary endpoint is the number of DLTs occurring in participants, with a stated timeframe of 15 months and a specified toxicity definition. Local failure is listed as a secondary outcome. The record includes descriptive adverse-event counts by arm, but does not report primary endpoint results, a maximum tolerable dose conclusion, or outcome analyses. It lists zero analyses and does not report the statistical plan, analysis populations, or missing-data methods.

Accordingly, the record describes the study question and design but does not provide the result needed to quantify the primary endpoint or assess its precision.