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
ENESTop was a phase 2, single-group trial in chronic myeloid leukemia evaluating treatment-free remission after cessation of nilotinib in participants who had achieved sustained MR4.5. The registry reports one binary primary endpoint measured during the first 48 weeks following nilotinib cessation.
| Feature | ENESTop |
|---|---|
| Trial name | ENESTop |
| NCT identifier | NCT01698905 |
| Brief title | Treatment-free Remission After Achieving Sustained MR4.5 on Nilotinib (ENESTop) |
| Phase | Phase 2 |
| Condition | Chronic Myeloid Leukemia |
| Allocation | NA |
| Design model | Single group |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 163 |
| Intervention | Nilotinib |
| Lead sponsor | Novartis Pharmaceuticals |
| Sponsor type | Industry |
| Status | Completed |
| Start | 2012-12-20 |
| Primary completion | 2015-11-26 |
2. Clinical Question
The central statistical question is whether participants who had achieved sustained MR4.5 could remain in treatment-free remission after nilotinib cessation. Because the registry describes a single-group design rather than a randomized comparison, the principal estimand is a remission proportion within the defined follow-up period rather than a between-arm treatment effect.
Population
Participants with chronic myeloid leukemia enrolled in the ENESTop study after achieving sustained MR4.5 on nilotinib.
Intervention
Nilotinib, followed by cessation of nilotinib for evaluation of treatment-free remission.
Comparator
The registry identifies a single-group design, so there is no randomized comparator arm in the registered design.
Primary question
What percentage of participants remains in treatment-free remission within the first 48 weeks following nilotinib cessation?
3. Trial Design
Statistically, the single-group structure is important. There is no randomized treatment-versus-control contrast from which to estimate a hazard ratio, risk ratio, or treatment difference. Instead, the primary outcome is naturally expressed as a proportion of participants satisfying the registry's definition of treatment-free remission during the prespecified period.
4. Endpoint Definition
| Endpoint | Time frame | Registry definition |
|---|---|---|
| Percentage of Participants in TFR Within 48-weeks Following Nilotinib Cessation | first 48 weeks following nilotinib cessation | TFR is defined as no confirmed loss of MR4 (Molecular response 4.0 log reduction from baseline) or loss of MMR (major molecular response) and no re-starting of nilotinib therapy within 12 months following cessation of nilotinib. Confirmed loss of MR4 is two consecutive BCR-ABL > 0.01% IS. Loss of MMR does not require confirmation. |
The registry therefore defines treatment-free remission using a composite clinical rule: the participant must avoid the specified molecular-response losses and must not restart nilotinib within the relevant period. This is more specific than simply asking whether a participant remains off therapy.
5. Statistical Methodology
Primary endpoint: a binary proportion
The registry classifies the primary endpoint as binary. Each participant can therefore be classified according to whether the registry-defined treatment-free remission criterion is met within the specified assessment framework.
The resulting quantity is a proportion or percentage. The appropriate uncertainty interval depends on the prespecified statistical method and the analysis population used for the endpoint.
What would normally be analysed
For a single-group binary endpoint, the principal statistical task is estimation of the observed proportion together with an appropriate confidence interval. A one-sample binomial framework is a natural approach because the endpoint is binary at the participant level.
If a protocol specifies a target remission proportion for formal hypothesis testing, a one-sample binomial test can additionally compare the observed proportion with that prespecified reference value. However, the ClinicalTrials.gov record does not provide a null proportion, an alternative hypothesis, an alpha level, or a formal statistical analysis record.
Why a hazard ratio is not the primary statistic here
A hazard ratio is designed for a comparison of time-to-event hazards, usually between treatment groups or between covariate-defined groups. ENESTop is registered as a single-group study with a binary primary endpoint. On the ClinicalTrials.gov record, there is no randomized comparator against which a treatment hazard ratio could be estimated.
Analysis population matters
For a binary endpoint, the denominator is crucial. The percentage can change depending on whether the analysis uses all enrolled participants, all participants who reach nilotinib cessation, or another prespecified analysis population. The ClinicalTrials.gov record does not provide a formal analysis-population definition for the primary endpoint, so no denominator beyond the overall enrollment of 163 should be assumed.
6. Reported Results
The registry indicates that results are posted and that the primary endpoint has results posted for it. However, the ClinicalTrials.gov record contains no formal statistical analyses. Consequently, the available trial data supports identification of the endpoint and its definition, but it does not support reporting a numerical TFR estimate, confidence interval, or p-value.
| Primary endpoint | Registry status | Formal statistical result reported |
|---|---|---|
| Percentage of Participants in TFR Within 48-weeks Following Nilotinib Cessation | Results posted for it: Yes | No statistical analysis reported |
7. Planned Analysis
Because no formal statistical-analysis record was reported, the most defensible statistical reconstruction is the following: classify participants according to the registered TFR definition and estimate the percentage meeting that definition within the registered time frame.
Step 1 · Define the outcome
Determine whether each participant satisfies the registry definition of treatment-free remission, including the molecular-response criteria and the absence of nilotinib restart.
Step 2 · Count participants
Count participants meeting the TFR definition and identify the prespecified analysis denominator.
Step 3 · Estimate the proportion
Divide the number meeting the endpoint by the relevant analysis denominator and express the result as a percentage.
Step 4 · Quantify uncertainty
Use a one-sample binomial confidence interval or another method prespecified in the statistical analysis plan.
For this endpoint, a confidence interval is particularly useful because it distinguishes the observed percentage from the precision with which that percentage is estimated. The width of the interval depends on both the observed proportion and the amount of information contributing to the estimate.
8. Statistical Methods Explained
What makes ENESTop a single-group analysis?
The registry specifies SINGLE_GROUP as the design model and lists one arm. This means the primary analysis is centered on the outcome observed after nilotinib cessation rather than on a randomized comparison between two concurrently assigned treatment strategies.
Why is the primary endpoint binary?
The endpoint asks whether each participant satisfies a defined treatment-free remission rule. That creates a participant-level yes/no classification, making a proportion a natural summary measure.
Why is the confidence interval important for a percentage?
An observed percentage is an estimate rather than a perfectly known population parameter. A confidence interval communicates sampling uncertainty around the estimated proportion. A narrow interval indicates greater statistical precision than a wide interval, all else being equal.
Why would a binomial method be appropriate?
When each participant contributes a binary outcome, the binomial distribution provides a direct probability model for the number of participants meeting the endpoint. Exact or other binomial confidence intervals can then be used to quantify uncertainty without requiring a continuous-outcome approximation.
Why can treatment-free remission not simply be equated with being off treatment?
The registry definition is more restrictive than treatment cessation alone. TFR requires no confirmed loss of MR4 or loss of MMR and no restarting of nilotinib within the stated definition. Thus, remaining off therapy is not by itself sufficient to establish the endpoint.
Why does the molecular-response definition matter statistically?
The rule determines exactly which observations count as events. Confirmed loss of MR4 requires two consecutive BCR-ABL measurements above the specified threshold, whereas loss of MMR does not require confirmation. These rules directly affect endpoint classification and therefore the estimated proportion.
Can this trial establish that nilotinib causes treatment-free remission?
The registry design is single-group rather than randomized. A single-group remission proportion can describe outcomes after the treatment pathway studied, but without a randomized comparator it does not by itself isolate the causal effect of nilotinib cessation from the underlying clinical course or other factors.
9. Safety Results
The ClinicalTrials.gov record reports serious adverse events by named study groups. These groups include treatment-consolidation, treatment-free-remission, treatment-reinitiation, and treatment-continuation categories. Because the ClinicalTrials.gov record does not provide a complete denominator framework or a unified safety-analysis definition for these categories, the figures are reported exactly as provided rather than converted into rates or compared statistically.
| Reported group | Serious adverse events affected / at risk |
|---|---|
| Nilotinib Treatment Consolidation (NTCS) | 21/163 |
| Nilotinib Treatment-free Remission (TFR) | 34/126 |
| Nilotinib Treatment Re-initiation (NTRI) | 27/61 |
| Nilotinib Treatment Continuation (NTCT) | 3/26 |
| Nilotinib Treatment-free Remission - 2 P | 3/7 |
| Nilotinib Treatment Re-initiation-2 Phas | 2/2 |
| Nilotinib | Registry data registry-reported without a complete affected/at-risk value |
These figures should not be interpreted as randomized safety comparisons. The groups represent stages or categories within the study pathway, and the ClinicalTrials.gov record does not establish that participants were independently randomized to each of these categories.
10. Treatment-Free Remission as a Statistical Endpoint
Treatment-free remission is a useful example of why endpoint definitions matter. The endpoint combines a clinically meaningful state—remaining without treatment—with objective molecular-response criteria and a treatment-restart rule.
| Component | Role in endpoint classification |
|---|---|
| No confirmed loss of MR4 | Participants must not meet the registry's criterion for confirmed loss of MR4. |
| No loss of MMR | Participants must not lose major molecular response under the registry definition. |
| No nilotinib restart | Restarting nilotinib within the stated period prevents classification as TFR. |
| 48-week assessment frame | The primary endpoint is identified as occurring within the first 48 weeks following nilotinib cessation. |
This structure illustrates an important principle in clinical-trial statistics: the endpoint definition determines the data-generating rule before the statistical method is applied. Statistical analysis cannot repair an ambiguous outcome definition after the fact. Once the rules are fixed, the analysis should apply them consistently to every participant in the relevant population.
11. Molecular Response and Event Classification
The registry describes MR4 as a molecular response representing a 4.0 log reduction from baseline. It defines confirmed loss of MR4 using two consecutive BCR-ABL measurements greater than 0.01% IS. Loss of MMR does not require confirmation.
This is a composite classification rule rather than a simple measurement of one laboratory value at one time point.
The distinction between confirmed and unconfirmed loss is statistically consequential. Requiring two consecutive measurements for loss of MR4 reduces the possibility that a single measurement triggers an endpoint classification, whereas the registry specifies no such confirmation requirement for loss of MMR.
12. Censoring, Missing Data, and Follow-Up
The ClinicalTrials.gov record does not specify a formal missing-data or imputation strategy for the primary endpoint. It also does not provide enough participant-level information to reconstruct censoring patterns or determine how every participant was handled at the 48-week assessment.
For a binary endpoint, missing outcome status can be especially important because the numerator and denominator depend on endpoint ascertainment. A prespecified analysis plan might define how participants without sufficient molecular-response information are handled, but that rule cannot be inferred from the registry summary.
13. Non-Inferiority, Factorial Design, and Bayesian Methods
The registry-reported ENESTop data does not identify a non-inferiority hypothesis, a non-inferiority margin, a factorial design, or a Bayesian statistical method.
| Design topic | Supported by the ClinicalTrials.gov record? | Interpretation |
|---|---|---|
| Non-inferiority margin | No | No margin is provided. |
| Crossover | No | No randomized comparator or crossover scheme is reported in the ClinicalTrials.gov record. |
| Factorial design | No | The registered design is single-group. |
| Multiplicity | Not reported | No multiplicity strategy is reported. |
| Interim analysis | Not reported | No interim-analysis strategy is reported. |
| Missing-data / imputation method | Not reported | No formal imputation strategy is reported. |
| Stratification | Not reported | No randomization stratification factors are reported. |
| Bayesian analysis | No | No Bayesian method is reported. |
This distinction is important for an independent statistical review. The absence of a method in the ClinicalTrials.gov record does not prove that no such method existed in a protocol or statistical analysis plan; it means that the method cannot be responsibly attributed to the ClinicalTrials.gov record.
14. Interpretation of the Primary Endpoint
The primary endpoint would quantify the percentage of participants satisfying the predefined treatment-free remission rule during the specified post-cessation assessment period.
A single-group TFR percentage would not, by itself, provide a randomized estimate of the causal effect of stopping nilotinib versus continuing nilotinib. It would describe the proportion meeting the defined remission criterion in the studied treatment pathway.
The point estimate alone does not indicate how precisely the underlying remission proportion has been estimated. A confidence interval is therefore an essential companion to the percentage in a complete statistical report.
If a one-sample hypothesis test were prespecified, its p-value would quantify evidence against a specified null proportion under the test's assumptions. It would not measure the magnitude of treatment-free remission and would not substitute for the estimated percentage and its confidence interval.
15. Limitations and Interpretation Issues
- Single-group design: the registered design does not provide a randomized comparator for estimating a between-treatment causal effect.
- No formal statistical analyses reported: no formal statistical analyses were posted, so no numerical primary-endpoint estimate, confidence interval, or p-value is reported here.
- Denominator uncertainty: the overall enrollment is 163, but the ClinicalTrials.gov record does not establish that 163 is the appropriate denominator for the primary TFR percentage.
- Endpoint complexity: TFR incorporates molecular-response rules and treatment-restart criteria rather than a single measurement.
- Different confirmation rules: confirmed loss of MR4 requires two consecutive BCR-ABL measurements above the specified threshold, whereas loss of MMR does not require confirmation.
- Time-frame wording: the endpoint time frame and the 12-month criterion in the endpoint definition should be reported as registered rather than harmonized by assumption.
- Missing-data strategy: no formal imputation or missing-outcome rule is provided in the ClinicalTrials.gov record.
- Safety interpretation: serious-adverse-event counts are posted on ClinicalTrials.gov for several non-randomized treatment-status groups with different denominators, limiting direct comparison.
- Generalizability: the study population consists of participants meeting the trial's eligibility and sustained-response pathway; the registry summary does not establish how broadly the observed TFR experience applies to other populations.
16. Why This Trial Matters Statistically
ENESTop is a useful teaching case because it demonstrates how clinical-trial statistics change when the design is single-group rather than randomized. The key statistical problem is not estimation of a treatment hazard ratio between arms. It is careful definition and estimation of a clinically meaningful binary outcome in a population that has reached a specified molecular-response state.
| Concept | How it appears in ENESTop |
|---|---|
| Single-group design | The registry identifies a SINGLE_GROUP design with one intervention. |
| Binary endpoint | The primary endpoint is the percentage of participants in TFR. |
| Endpoint definition | TFR combines molecular-response criteria with a nilotinib-restart rule. |
| Binomial estimation | A natural framework for estimating a participant-level binary remission proportion. |
| Confidence interval | Needed to describe statistical precision around the estimated proportion. |
| Analysis population | The choice of denominator is critical to interpreting the percentage. |
| Missing data | Incomplete molecular-response information can affect binary endpoint classification. |
| Safety denominators | Reported serious-adverse-event groups have different affected/at-risk counts. |
| Causal interpretation | Without randomized comparison, a TFR percentage does not itself estimate a randomized treatment effect. |
17. Related Tutorials
Learn more about the methods used in this trial:
18. Related Statistical Calculators
19. Sources
- ClinicalTrials.gov: ENESTop, NCT01698905.
- Linked publication: PubMed record for PMID 29459949.
Continue through Clinical Biostats
Explore the statistical concepts behind clinical-trial endpoints, study designs, estimation, and analysis methods.
20. Record Summary
ENESTop provides a focused example of statistical analysis for a single-group clinical trial with a binary primary endpoint. The registry identifies 163 enrolled participants, a phase 2 design, nilotinib as the intervention, and treatment-free remission as the primary endpoint within the first 48 weeks following nilotinib cessation. The endpoint itself is clinically specific: TFR requires no confirmed loss of MR4 or loss of MMR and no restarting of nilotinib within the 12-month criterion stated in the registry definition.
The ClinicalTrials.gov record reports that results are posted, but it does not provide a formal statistical-analysis record. Instead, the appropriate analytical framework is to classify the binary endpoint according to the registry definition, estimate the relevant proportion in the prespecified analysis population, and quantify uncertainty with an appropriate binomial confidence interval. The serious-adverse-event data should likewise be interpreted with care because the registry-reported groups have different denominators and represent different treatment-status categories rather than randomized comparison arms.