Azacitidine and CAPOX in Metastatic Colorectal Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Azacitidine and CAPOX in Metastatic Colorectal Cancer (NCT01193517), a non-randomized phase 1, phase 2 study. The primary outcome is Maximum Tolerated Dose (MTD) of Azacitidine, and Capecitabine and Oxaliplatin (CAPOX).

2. Clinical Question

The study therefore addresses dose selection and safety in this treatment combination, as well as a stated interest in disease control. It does not establish a comparative treatment effect: the design has no randomized control arm, and the posted results do not include an efficacy analysis. The registry record does not report the trial’s detailed statistical analysis plan, criteria for response, or the result for the primary endpoint. Those omissions limit what can be concluded about either the selected dose or response to treatment.

3. Trial Design

This completed study was registered as a Phase I/II trial in metastatic colorectal cancer. Its allocation is described as non-randomized, its model as single-group, and its masking as none. Enrollment was 26. The registry lists the start as 2010-08 and primary completion as 2016-11. The sponsor was M.D. Anderson Cancer Center.

The Phase I portion was a dose-escalation study of azacitidine with CAPOX. Its stated purpose was to identify the highest tolerable azacitidine dose for use with capecitabine and oxaliplatin. The Phase II portion used the MTD of azacitidine with CAPOX and was intended to assess whether the combination could help control metastatic colorectal cancer; safety was also a stated interest. These are descriptions of the study’s aims, not evidence that a particular dose or degree of disease control was achieved.

The registry lists Phase I starting-dose, Phase I highest-dose-level, and Phase II groups in its adverse-event information. They are not randomized arms. Differences between their counts cannot be read as treatment effects because the groups represent study phases or dose levels, and the registry does not describe them as exchangeable comparison groups. In particular, the Phase II cohort is not a concurrent control for the Phase I cohorts.

The record does not report the dose-escalation algorithm, cohort progression rules, actual doses, eligibility criteria in detail, or the protocol’s sample-size justification. It also does not report whether any interim analyses were planned or conducted. The presence of both Phase I and Phase II labels does not by itself specify how the study governed dose selection, when Phase II began, or which data were included in each analysis.

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:

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 Maximum Tolerated Dose (MTD) of Azacitidine, and Capecitabine and Oxaliplatin (CAPOX), measured over Up to 3 weeks from the first dose.

The primary endpoint is the maximum tolerated dose of azacitidine in combination with capecitabine and oxaliplatin, expressed in mg/m^2. The registry defines it as the dose just below the dose at which at least 1/3 of subjects experience a dose limiting toxicity. The assessment timeframe is up to 3 weeks from the first dose.

In general, a dose-finding endpoint is used to guide selection of a dose for later study by balancing exposure to the investigational combination against dose-limiting toxicity within a defined observation window. The endpoint is not itself a measure of tumor response or survival. Its interpretation depends on the toxicity definition, observation period, dose levels tested, and dose-escalation rules. The registry record does not report the detailed toxicity criteria or escalation algorithm.

The posted record names the endpoint but does not state the MTD value. It therefore does not show which dose met the endpoint definition or how precisely that estimate was determined. The secondary outcome list includes response rate, but no response result or analysis is reported in the registry record. The endpoint description should be distinguished from a reported endpoint result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies a primary dose-finding endpoint, a secondary response-rate outcome, and adverse-event information. It does not report the statistical analysis plan in sufficient detail to identify the methods used for final endpoint estimation, response analysis, or safety comparisons. The concepts below are general explanations, not claims about procedures used in this trial.

What does a maximum tolerated dose mean?

In general, a maximum tolerated dose is a dose selected according to a prespecified toxicity rule. In this record, the rule is described as the dose just below the dose at which at least 1/3 of subjects experience a dose limiting toxicity. Applying such a rule requires clear definitions of dose-limiting toxicity, the observation window, dose levels, and how decisions are made when toxicity information is incomplete or uncertain. The registry gives the timeframe as up to 3 weeks from the first dose but does not report the detailed operational rules or the selected dose.

A dose-finding estimate is not automatically a precise or stable estimate of toxicity risk. Its uncertainty depends on the number of participants treated at each dose, the observed toxicities, and the design’s dose-allocation rules. In general, a model-based dose-finding approach may use information across dose levels, while a rule-based approach may use predefined escalation steps. The registry record does not identify which approach was used, so neither should be attributed to this study.

What can adverse-event counts tell a reader?

Counts of affected participants are descriptive. A count paired with the number at risk identifies how many participants in a group were reported to have an event and how many were considered at risk for that reporting category. It does not, by itself, estimate a causal effect. In general, comparisons of event proportions need to account for study design, the comparability of groups, event definitions, and uncertainty. Since this study is non-randomized and single-group, the registered Phase I and Phase II cohorts should not be treated as randomized treatment arms.

Over-dispersion is a general issue when variation in event counts exceeds what a simple model assumes. If present, it can make standard errors too small and confidence intervals too narrow under an overly simple model. Whether over-dispersion is relevant depends on the data structure and model. The registry record does not report a model for adverse events, a dispersion assessment, or an inferential safety analysis.

How should response rate be interpreted?

In general, response rate is a proportion based on a defined response criterion and a stated analysis population. Interpretation depends on how response is assessed, how missing or unevaluable assessments are handled, and whether confirmation is required. The record lists response rate as a secondary outcome but does not report its definition, numerator, denominator, estimate, confidence interval, or analysis population. No response estimate can therefore be interpreted from the posted registry information.

What do confidence intervals and p-values contribute?

In general, a confidence interval expresses statistical uncertainty around an estimate under a specified method and assumptions. A wider interval indicates less precision than a narrower interval, although the interpretation also depends on the endpoint and analysis. A p-value describes how compatible observed data are with a specified null model, under its assumptions; it is not the probability that a treatment works or that the null hypothesis is true. Neither a confidence interval nor a p-value is reported for the primary endpoint or response rate in the registry record. The adverse-event section gives counts and selected percentages, not a comparative inferential analysis.

A lack of a reported p-value does not mean that a result is statistically significant or non-significant. It means that the registry record does not provide that inferential result. Similarly, a point estimate without its uncertainty would not establish precision. Here, the MTD value itself is absent, so its uncertainty cannot be assessed from the registry record.

How do sample-size illustrations differ from trial evidence?

In general, sample-size planning calculations describe the amount of information that would be needed under stated assumptions. The accompanying planning assumptions specify a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the detectable effect is framed as a stated fraction of the observed effect. These are planning assumptions, not evidence that the study randomized participants between two arms. The registry design is non-randomized and single-group. Any sample sizes calculated under those assumptions are planning illustrations, never observed power, and do not replace analysis of trial outcomes.

Observed power is not a substitute for an effect estimate and its uncertainty. In general, once the data are observed, a confidence interval and the estimate itself provide more direct information about precision. The registry record does not report a trial-specific sample-size justification or the calculations used to plan enrollment. It would therefore be inappropriate to infer a design rationale from planning illustrations.

Overall, the statistical methods needed to reproduce or fully evaluate the trial’s endpoint analyses are not reported in the registry record. General statistical concepts can clarify what the listed endpoints and counts mean, but they cannot fill in missing analysis decisions or results.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures, including the primary maximum tolerated dose endpoint and the secondary response-rate outcome. It does not report an analysis for either outcome in the posted results. The registry record does not state whether a multiplicity strategy was prespecified, whether any outcomes were prioritized beyond the primary designation, or whether statistical testing was performed.

In general, multiplicity arises when multiple outcomes, comparisons, or analyses are examined. Without a stated plan, readers cannot determine whether any reported inferential results were adjusted to address multiple testing. No inferential results are reported, so no adjustment or unadjusted testing should be inferred.

The registry record does not report how missing outcome data were handled, how many participants were evaluable for the MTD or response-rate endpoint, or how analysis populations were defined. It also does not report whether all enrolled participants received treatment or completed the endpoint observation window. These details matter because excluding participants or treating unevaluable assessments in different ways can change an outcome estimate.

For safety, the adverse-event groups include at-risk counts and affected counts, but the registry does not describe the exact safety population definition, follow-up completeness, or handling of missing event information. The reported counts should be read as the registry’s descriptive record, not as a complete account of all possible analysis choices. Detailed missing-data and analysis-population procedures are not reported in the registry record.

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: nausea (16.7%), vomitting (9.1%), pain-pelvis (8.3%).

The serious adverse-event figures are reported separately for the Phase I starting-dose group, the Phase I highest-dose-level group, and Phase II. The registry lists 0 serious adverse events among 3 at risk in the starting-dose group, 2 among 12 at risk in the highest-dose-level group, and 1 among 11 at risk in Phase II. It lists 0 deaths in each group.

These figures are descriptive counts, not evidence from a randomized comparison. The groups correspond to different study phases or dose levels, and no inferential comparison is reported. The registry also lists other adverse events in 2 of 3, 12 of 12, and 11 of 11 at risk, respectively. A participant-level affected count does not indicate how many separate events occurred or how severe, prolonged, or recurrent they were.

The listed top serious-event terms are Nausea at 16.7%, Vomitting at 9.1%, and Pain-pelvis at 8.3%. The registry record does not report their detailed denominators or clarify whether an individual participant could be represented under more than one term. These figures should therefore be read as the registry’s reported terms and percentages, not as a complete event profile or comparative risk estimate.

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

The posted registry information establishes that a Phase I/II study of azacitidine with CAPOX was completed, with enrollment of 26, and that the primary endpoint concerned maximum tolerated dose within a defined toxicity window. It also records adverse-event counts for three study groups and identifies response rate as a secondary outcome. These registry entries describe the trial’s design and the information entered in the registry.

The record does not give the MTD result or a response-rate estimate. It therefore cannot show which azacitidine dose was selected under the stated dose-limiting toxicity rule, how much tumor response occurred, or how precise an efficacy estimate might be. The study’s stated aim of learning whether the combination can help control metastatic colorectal cancer is not itself a result. A named outcome is not equivalent to a posted finding.

The trial was non-randomized, unmasked, and single-group. Its Phase I dose levels and Phase II cohort do not create a randomized control comparison. As a general point, without an appropriate comparison group, changes observed during treatment cannot isolate the contribution of the treatment combination from other influences. The posted adverse-event counts can be described within each group, but they do not establish comparative safety effects.

The safety entries report serious adverse events and other adverse events among the stated at-risk numbers, and list zero deaths in the recorded groups. Those data do not by themselves establish overall tolerability: the registry record does not provide a full account of event timing, duration, severity grading, treatment discontinuation, or follow-up. Nor can the figures establish how risk compares with another treatment.

The interpretation is consequently bounded. The record supports statements about the study’s purpose, design, enrollment, selected safety counts, and the absence of posted endpoint results. It does not support a conclusion about efficacy, comparative benefit, or the actual dose selected. A missing result is not evidence of no effect; it is an absence of a result in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable as a dose-finding study with a later Phase II component, but its posted record illustrates why design and result availability must be considered together. The primary endpoint is a dose-selection outcome defined by a dose-limiting toxicity threshold over a specified timeframe. It is not interchangeable with an efficacy endpoint, and the registry record does not report the actual MTD value.

The study’s non-randomized, single-group structure also matters for interpretation. The Phase I starting-dose and highest-dose-level groups, along with the Phase II group, are distinct study cohorts rather than randomized controls. Their adverse-event counts can be described, but differences in those counts cannot be attributed to dose alone from the registry information. Dose level, phase, cohort composition, and other factors may differ, while detailed methods for assessing these factors are not reported.

The registry lists a response-rate outcome, which signals an efficacy question, but does not report the response result. This distinction is important in statistical reporting: an intended outcome is not a finding, and a completed study does not guarantee that results are available in the record. Without estimates and measures of uncertainty, readers cannot assess the magnitude or precision of a response or dose-selection result.

The adverse-event counts show why denominator context matters. The number affected is paired with an at-risk count for each group, while the record does not provide a comparative analysis or complete event-level detail. This supports cautious descriptive reading rather than a broad conclusion about safety.

More generally, the record demonstrates that registry descriptions can communicate trial purpose and structure while leaving key analysis details and endpoint results unreported. Statistical interpretation depends not only on the research question but also on outcome definitions, analysis populations, missing-data handling, uncertainty estimates, and design features. Where these elements are absent, conclusions must remain correspondingly limited.

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 registry record does not report the MTD value, response-rate result, or inferential analyses. Although enrollment was 26, the record does not state how many participants contributed to each primary or secondary outcome analysis, or whether all enrolled participants were included in the safety summaries. No detailed eligibility criteria or biomarker measurements are reported, despite the official title’s reference to enrichment for hypermethylation of CpG promoter islands.

The dose-escalation rules, specific doses, criteria for dose-limiting toxicity, and detailed safety grading procedures are not reported in the registry record. Without these details, readers cannot reproduce the dose-selection process or fully evaluate how the stated threshold was applied. The MTD timeframe is listed as up to 3 weeks from the first dose, but follow-up beyond that window and the handling of incomplete observation are not described.

The trial was non-randomized, single-group, and unmasked. There is no like-for-like control comparison, so the groups cannot provide a randomized estimate of treatment effect. The adverse-event counts are also limited by the absence of event-level context, including duration and recurrence, and by the lack of reported uncertainty estimates or comparative safety analyses.

The registry record does not report the statistical analysis plan, sample-size justification, interim-analysis procedures, multiplicity approach, missing-data handling, or analysis-population definitions. The record also does not report detailed response criteria or response assessment procedures. These omissions mean that neither the primary endpoint nor the secondary outcome can be fully evaluated from the posted information. The limitations concern what the registry reports; they do not establish what may or may not have been done outside that record.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Azacitidine and CAPOX in Metastatic Colorectal Cancer — NCT01193517.
• Registry record last updated 2020-04-15; status last verified 2020-04.
• Results first posted 2020-04-15.
• 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

NCT01193517 is a completed Phase I/II study of azacitidine with capecitabine and oxaliplatin (CAPOX) in metastatic colorectal cancer. The study was non-randomized, unmasked, and single-group, with enrollment of 26. Its Phase I portion used dose escalation, and its Phase II portion used the MTD of azacitidine with CAPOX.

The primary endpoint was the maximum tolerated dose, defined as the dose just below the level at which at least 1/3 of subjects experience a dose limiting toxicity, with a timeframe of up to 3 weeks from the first dose. Response rate was listed as a secondary outcome. The registry record does not report the MTD or response-rate results.

Adverse-event counts are recorded for the Phase I starting-dose, Phase I highest-dose-level, and Phase II groups. These are descriptive groups, not randomized comparators. The record does not report the detailed statistical plan, analysis populations, or missing-data procedures.