This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
M14-032 was a phase 2, open-label, non-randomized, single-group study of venetoclax in participants with chronic lymphocytic leukemia who had relapse or refractory disease following B-cell receptor signaling pathway inhibitor therapy. The registry reports 127 participants and four arms.
| Feature | M14-032 |
|---|---|
| Phase | Phase 2 |
| Condition | Chronic Lymphocytic Leukemia |
| Intervention | Venetoclax |
| Allocation | Non-randomized |
| Design model | Single-group |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 127 |
| Number of arms | 4 |
| Status | Completed |
| Start date | 2014-09-10 |
| Primary completion date | 2021-12-22 |
| Lead sponsor | AbbVie |
| Sponsor type | Industry |
| Results posted | Yes |
2. Clinical Question
The clinical question was whether venetoclax could be evaluated for efficacy and safety in participants with chronic lymphocytic leukemia whose disease had relapsed or was refractory following therapy with a B-cell receptor signaling pathway inhibitor.
Population
Participants with chronic lymphocytic leukemia with relapse or refractory disease following B-cell receptor signaling pathway inhibitor therapy.
Intervention
Venetoclax, identified in the registry as ABT-199 and GDC-0199.
Comparator
The registry describes a non-randomized, single-group design; no randomized comparator is reported.
Primary question
What overall response rate was observed, according to investigator assessment using 2008 Modified IWCLL NCI-WG criteria?
3. Trial Design
M14-032 was an open-label phase 2 study with non-randomized allocation and a single-group design. Although the registry identifies four arms, the design was not randomized and therefore does not provide the treatment-versus-control structure required for a conventional randomized comparative efficacy analysis.
Why the single-group design matters statistically
In a single-group study, the observed response rate describes the experience of the enrolled study population. It is not automatically a randomized estimate of the difference between venetoclax and another treatment. Without random allocation to a concurrent comparator, differences in prognosis, prior treatment, disease characteristics, follow-up, and other factors cannot be balanced through randomization.
This distinction is fundamental when interpreting a phase 2 response endpoint. A high observed response rate can provide important evidence about antitumor activity, but it answers a different statistical question from a randomized trial comparing two treatment strategies.
4. Study Arms and Analysis Structure
The registry identifies four arms associated with treatment following prior ibrutinib therapy, treatment following prior idelalisib therapy, and an arm described as treatment following either ibrutinib or idelalisib therapy. The serious-adverse-event records further distinguish main and expansion groups.
ABT-199 After Ibrutinib Therapy
- Main cohort: 43 at risk for the reported serious-adverse-event measure.
- Expansion cohort: 48 at risk for the reported serious-adverse-event measure.
- Main and expansion records are reported separately in the safety data.
ABT-199 After Idelalisib Therapy
- Main cohort: 21 at risk for the reported serious-adverse-event measure.
- Expansion cohort: 15 at risk for the reported serious-adverse-event measure.
- Main and expansion records are reported separately in the safety data.
Ibrutinib or idelalisib therapy
The registry also reports a group described as ABT-199 after ibrutinib or idelalisib therapy, with 64 participants at risk in the main record and 63 participants at risk in the expansion record for the reported serious-adverse-event measure.
Overall enrollment
The overall registry enrollment is 127 participants. The arm-level safety records should therefore be read using their specific registry labels rather than treating the four arm records as four independent randomized groups.
5. Primary Endpoint
| Endpoint | Registry definition | Assessment schedule |
|---|---|---|
| Overall Response Rate (ORR) | Overall response rate is defined as the percentage of participants with an overall response, per the investigator assessment, using 2008 Modified International Workshop for Chronic Lymphocytic Leukemia (IWCLL) National Cancer Institute-Working Group (NCI-WG) criteria. | At Wk 5, Day 1; Wk 8, Day 1; Wk 12, Day 1; Wk 16, Day 1; Wk 20, Day 1; Wk 24, Day 1; Wk 36, Day 1; every 12 wks after Wk 36; Final Visit. |
What ORR measures
ORR is a binary-response endpoint at the participant level: each participant is classified according to whether an overall response was achieved under the specified response criteria. The resulting proportion summarizes the frequency of response in the analyzed population.
Because the endpoint is based on investigator assessment, the classification depends on the clinical and disease-assessment process used to determine whether a participant meets the 2008 Modified IWCLL NCI-WG criteria. The registry does not provide a randomized comparator against which the observed proportion can be directly contrasted.
6. Statistical Methodology
Analysis of a single-group response proportion
For a single-group ORR endpoint, the basic statistical quantity is the observed proportion of participants meeting the definition of overall response. If \(x\) participants respond among \(n\) evaluable participants, the observed response proportion is
The principal inferential task is then to quantify uncertainty around that proportion, typically with an appropriate confidence interval for a binomial proportion.
Binomial framework
At the simplest level, each participant contributes a response indicator: response or no response. Under a binomial framework, the observed number of responses provides information about the underlying response probability in the population represented by the study.
The choice of confidence-interval method matters, particularly when the sample size is modest or the response proportion is near 0% or 100%. A confidence interval is preferable to reporting a response proportion alone because it communicates the statistical precision of the observed estimate.
Why there is no randomized treatment-effect estimate
Randomized comparative analyses commonly estimate quantities such as a risk ratio, odds ratio, risk difference, or hazard ratio between treatment groups. M14-032 is registered as non-randomized and single-group, so the primary endpoint is naturally interpreted as an observed response proportion rather than a randomized treatment effect.
The second framework depends on randomization to support a direct causal comparison. The first primarily describes the response experience within the enrolled study population.
Assessment over repeated time points
The registry specifies repeated response assessments beginning at Wk 5, Day 1 and continuing through the Final Visit, with assessments every 12 weeks after Wk 36. Repeated assessments allow response status to be evaluated throughout follow-up, but the primary endpoint should still be interpreted according to its prespecified endpoint definition and assessment framework.
Investigator assessment
Because ORR is defined using investigator assessment, response classification is not simply an automated calculation from a continuous laboratory measurement. The endpoint incorporates the specified disease-response criteria and the investigator's assessment of whether those criteria are satisfied.
7. Planned Analysis
Results are posted for the M14-032 record, but no formal statistical analyses are posted to ClinicalTrials.gov. The registry identifies ORR as the primary endpoint and provides its definition and assessment schedule, while the posted record does not provide a formal statistical estimate, confidence interval, or p-value for that endpoint.
The statistical interpretation therefore centers on the structure of the endpoint rather than on a reported treatment-effect estimate. In particular, an ORR from a single-group study should not be converted into a relative treatment effect against an external population unless the comparison population and statistical framework have been explicitly defined.
8. Follow-Up and Endpoint Timing
The primary ORR endpoint was assessed at multiple scheduled visits: Wk 5, Day 1; Wk 8, Day 1; Wk 12, Day 1; Wk 16, Day 1; Wk 20, Day 1; Wk 24, Day 1; Wk 36, Day 1; every 12 wks after Wk 36; and the Final Visit.
Study start
The registry lists September 10, 2014 as the study start date.
Scheduled response assessments
ORR assessment begins at Wk 5, Day 1 and continues through the prespecified assessment schedule.
12-week assessment intervals
The registry specifies assessment every 12 wks after Wk 36, continuing to the Final Visit.
Primary completion
The registry lists December 22, 2021 as the primary completion date.
The estimated median time on follow-up was 1694 d for the ibrutinib failure cohort and 1942 d for the idelalisib failure cohort. These durations indicate that response assessment occurred over a substantial follow-up period, but they do not themselves constitute an efficacy estimate.
9. Safety Results
The registry reports serious adverse events by arm as affected participants divided by participants at risk. Across all participants, 84 of 127 were affected by a serious adverse event in the reported safety measure.
| Registry arm label | Affected / at risk |
|---|---|
| ABT-199 After Ibrutinib Therapy: Main Co | 35/43 |
| ABT-199 After Idelalisib Therapy: Main C | 14/21 |
| ABT-199 After Ibrutinib or Idelalisib Th | 49/64 |
| ABT-199 After Ibrutinib Therapy: Expansi | 26/48 |
| ABT-199 After Idelalisib Therapy: Expans | 9/15 |
| ABT-199 After Ibrutinib or Idelalisib Th: Expans | 35/63 |
| ABT-199 After Ibrutinib Therapy: Main an | 61/91 |
| ABT-199 After Idelalisib Therapy: Main a | 23/36 |
| All Participants | 84/127 |
How to interpret the safety counts
The numerator is the number of affected participants and the denominator is the number at risk for the particular registry safety record. These records are not randomized treatment comparisons. In particular, the arm labels distinguish prior-therapy and main/expansion categories, so the records should be interpreted according to their registry definitions rather than treated as four independent randomized treatment groups.
The overall serious-adverse-event record is 84/127. This means 84 of 127 participants were recorded as affected in the reported serious-adverse-event measure. It does not identify the severity, timing, causality, or individual adverse-event terms beyond what is contained in the registry record.
10. Statistical Methods Explained
Why is ORR analyzed as a proportion?
ORR classifies participants according to whether they meet the specified definition of overall response. The natural summary is therefore the number responding divided by the number evaluated. This produces an estimated response probability for the study population.
Why is a confidence interval important for ORR?
A response proportion is an estimate rather than a perfectly known population quantity. A confidence interval expresses uncertainty around that estimate. Two studies can have the same observed ORR but very different precision if their numbers of evaluable participants differ.
Why is this not a hazard-ratio analysis?
A hazard ratio is designed for a time-to-event endpoint such as overall survival or progression-free survival. The registered primary endpoint in M14-032 is ORR, a response proportion assessed at scheduled visits. A hazard ratio is therefore not the natural effect measure for the primary endpoint as defined in the registry.
Why does the lack of randomization matter?
Randomization balances measured and unmeasured prognostic factors in expectation across treatment groups. Because M14-032 is registered as non-randomized and single-group, there is no randomized allocation mechanism creating a concurrent control group. Consequently, an observed response proportion cannot by itself establish how venetoclax would compare causally with another treatment.
What does repeated response assessment accomplish?
Repeated assessments provide multiple opportunities to determine whether a participant meets the response criteria during follow-up. The schedule extends from Wk 5, Day 1 through the Final Visit, with assessments every 12 wks after Wk 36. The timing is therefore part of the endpoint definition and should be retained when interpreting the response measure.
Why should main and expansion cohorts be distinguished?
The registry reports separate main and expansion safety records. Combining such records without understanding their definitions could obscure differences in the populations represented by the records and could make denominators difficult to interpret. The registry labels should therefore be preserved when presenting the safety information.
11. Understanding the Primary Endpoint Statistically
The most important statistical feature of M14-032 is the distinction between describing response and estimating a randomized treatment effect.
Descriptive question
What percentage of participants experienced an overall response according to the investigator assessment and the specified 2008 Modified IWCLL NCI-WG criteria?
Comparative question
How does the response probability under venetoclax differ from that under another treatment assigned through randomization?
M14-032 design
The registered design addresses the first type of question through a non-randomized, single-group phase 2 study.
Interpretive consequence
The observed response proportion should be interpreted within the enrolled study population rather than as a randomized comparative treatment effect.
This distinction also explains why the choice of statistical language matters. Terms such as "response rate" describe an observed or estimated proportion. Terms such as "risk reduction," "hazard ratio," and "treatment effect" generally imply a comparative framework that is not created by the registered single-group design.
12. Confidence Intervals, Precision, and Uncertainty
For a binary endpoint such as ORR, the observed proportion alone does not reveal how precisely the underlying response probability has been estimated. Confidence intervals address this by placing an interval around the estimated proportion under a specified statistical framework.
A confidence interval then describes statistical uncertainty around the underlying response probability. It is not a prediction interval for the response of a future individual participant.
Precision also depends on the number of participants contributing to the endpoint. The overall enrollment of 127 provides the broad study size, but the registry information in the ClinicalTrials.gov record does not provide a formal ORR numerator, ORR denominator, confidence interval, or p-value.
13. Analysis Populations and Denominators
Denominators are particularly important in single-group response studies because ORR is calculated from the participants who contribute to the response analysis. The registry clearly reports an overall enrollment of 127, but the primary-endpoint record provided here does not supply a separate ORR analysis denominator.
| Quantity | Registry information | Statistical role |
|---|---|---|
| Total enrollment | 127 | Defines the enrolled study population. |
| Number of arms | 4 | Describes the registered study structure; allocation was non-randomized. |
| ORR | Primary endpoint | Response proportion defined using investigator assessment and 2008 Modified IWCLL NCI-WG criteria. |
| Serious adverse events | 84/127 overall | Safety measure reported as affected participants / at risk. |
A denominator should always accompany a proportion when possible. Without the exact endpoint-specific analysis population, it is not appropriate to assume that every enrolled participant necessarily contributes identically to every efficacy assessment.
14. Safety and Efficacy Are Different Statistical Questions
M14-032 illustrates why efficacy and safety should be analyzed separately. ORR asks whether participants achieved an overall response under the specified disease-response criteria. Serious adverse events count participants affected by the reported safety outcome.
Efficacy
Primary endpoint: Overall Response Rate, assessed repeatedly according to investigator assessment and 2008 Modified IWCLL NCI-WG criteria.
Safety
Serious adverse events are reported by registry-defined arm and cohort categories as affected participants divided by participants at risk.
Different denominators
The denominator appropriate for an efficacy endpoint can differ from a safety denominator when analysis populations or exposure definitions differ.
Different interpretations
An efficacy response proportion should not be treated as evidence about adverse-event probability, and a safety count should not be interpreted as an efficacy estimate.
15. Limitations
- Non-randomized design: the study does not use random allocation to create a concurrent comparator, limiting causal comparison with another treatment.
- Single-group structure: the primary efficacy question is centered on the observed response experience within the study population rather than a randomized between-group treatment effect.
- Open-label assessment: masking is reported as none, and the primary endpoint is based on investigator assessment. Awareness of treatment can therefore be relevant to endpoint assessment.
- Endpoint denominator: the ClinicalTrials.gov record identifies enrollment of 127 but does not provide a formal ORR numerator and analysis denominator.
- No posted formal primary analysis: the registry identifies ORR as the primary endpoint and reports that results are posted, but no formal statistical analysis is posted for the endpoint in the record.
- Multiple assessment times: response was assessed repeatedly from Wk 5 through the Final Visit. Interpretation should remain tied to the registered endpoint definition and assessment schedule.
- Cohort structure: safety records distinguish main and expansion cohorts and prior-therapy categories. These should not automatically be treated as independent randomized groups.
- External comparison: an observed response proportion should not be interpreted as a randomized treatment effect against a historical or external population without a prespecified and appropriate comparative framework.
16. Why This Trial Matters Statistically
M14-032 is a useful teaching example because it demonstrates a common but important distinction in clinical-trial statistics: a phase 2 study can generate a clinically meaningful response estimate without having the randomized design required for a direct comparative treatment-effect estimate.
| Statistical concept | How it appears in M14-032 |
|---|---|
| Phase 2 development | The study is registered as phase 2. |
| Single-group analysis | The design model is single-group. |
| Non-randomized allocation | Allocation is registered as non-randomized. |
| Binary efficacy endpoint | ORR classifies participants according to whether an overall response occurred. |
| Binomial proportion | ORR is naturally summarized as responders divided by participants contributing to the endpoint. |
| Repeated assessment | ORR is assessed at prespecified visits from Wk 5 through the Final Visit. |
| Investigator assessment | The registry defines ORR according to investigator assessment. |
| Response criteria | The endpoint uses 2008 Modified IWCLL NCI-WG criteria. |
| Safety denominators | Serious adverse events are reported as affected / at risk for registry-defined groups. |
| Interpretive limitations | The absence of randomization limits direct causal comparison with another treatment. |
17. A Statistical Reading of the Trial Design
The design can be understood as a sequence of statistical questions rather than as a single number.
The first statistical task is defining the population. The second is defining the intervention. The third specifies what constitutes success. The fourth defines how that success is classified. The fifth determines when the endpoint is assessed.
Only after these elements are established does the numerical calculation become meaningful. For ORR, the fundamental calculation is a proportion. The interpretation then depends on the study design, the response criteria, the analysis population, and the uncertainty around the estimate.
18. What a Formal ORR Analysis Would Contain
For a single-group response endpoint, a complete statistical presentation would ordinarily identify the number of participants included in the response analysis, the number who achieved an overall response, the resulting ORR, and an appropriate confidence interval. If a formal hypothesis test were prespecified, it would also identify the benchmark being tested and the corresponding p-value.
| Component | Purpose |
|---|---|
| Response numerator | Counts participants meeting the overall-response definition. |
| Analysis denominator | Defines the population from which the response proportion is calculated. |
| ORR estimate | Summarizes the proportion achieving overall response. |
| Confidence interval | Quantifies statistical uncertainty around the response estimate. |
| Historical benchmark, if prespecified | Provides a reference value for a one-sample hypothesis test. |
| P-value, if prespecified | Addresses the specified null hypothesis; it does not measure the size or clinical importance of the response proportion. |
The registry record identifies the endpoint but does not provide these formal statistical analysis outputs. Consequently, the key statistical lesson is how the endpoint should be analyzed and interpreted, rather than a reported treatment-effect estimate.
19. Why This Is Not a Randomized Comparative Trial
The registry explicitly classifies allocation as non-randomized and the design model as single-group. These two characteristics determine much of the statistical interpretation.
Randomization creates treatment groups whose baseline differences arise from chance rather than from treatment-selection mechanisms, supporting causal comparisons under the trial design.
A single-group study instead follows one treatment strategy and estimates outcomes within that population. The observed outcome can be described precisely, but a direct treatment-versus-control contrast is not created by the design.
An ORR observed in M14-032 should therefore be read as evidence about response under the study's eligibility and treatment context, not as a randomized estimate of the difference between venetoclax and another therapy.
20. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The primary endpoint is a single-group response proportion assessed according to specified criteria at repeated time points. The appropriate inferential framework centers on a binomial proportion and its uncertainty.
Clinical interpretation
The study addresses venetoclax in participants with chronic lymphocytic leukemia after relapse or refractory disease following B-cell receptor signaling pathway inhibitor therapy.
What the design supports
The design supports characterization of response and safety within the enrolled population.
What requires caution
Because allocation was non-randomized and the design was single-group, direct causal comparisons with another treatment are not generated by the trial design itself.
21. Registry Record and Source Material
The ClinicalTrials.gov record identifies M14-032 as NCT02141282 and reports the study as completed, with results posted. The record supplies the registered primary endpoint, endpoint definition, assessment schedule, study design, enrollment, follow-up information, and serious-adverse-event counts summarized on this page.
Additional PubMed records
22. Sources
- ClinicalTrials.gov: NCT02141282 — M14-032.
- PubMed: PubMed record.
- PubMed: PubMed record.
- PubMed: PubMed record.
Continue through the Clinical Biostats knowledge graph
Explore additional clinical-trial methodology, statistical concepts, and analysis workflows across the Clinical Biostats site.
23. Record Summary
M14-032 is a phase 2, open-label, non-randomized, single-group study of venetoclax in chronic lymphocytic leukemia following relapse or refractory disease after B-cell receptor signaling pathway inhibitor therapy. The registry reports 127 enrolled participants, four arms, a primary endpoint of Overall Response Rate, repeated response assessments through the Final Visit, estimated median follow-up of 1694 d for the ibrutinib failure cohort and 1942 d for the idelalisib failure cohort, and serious-adverse-event counts by registry-defined cohort. The central statistical lesson is that ORR in this design is naturally understood as a response proportion with uncertainty, while the non-randomized single-group structure does not create a randomized treatment-effect comparison.