Aromatase Inhibitor Therapy With or Without Fulvestrant for the Treatment of HR Positive Metastatic Breast Cancer With an ERS1 Activating Mutation, the INTERACT Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Aromatase Inhibitor Therapy With or Without Fulvestrant for the Treatment of HR Positive Metastatic Breast Cancer With an ERS1 Activating Mutation, the INTERACT Study (NCT04256941), a randomized phase 2 study. The primary outcome is Progression-free Survival (PFS).

2. Clinical Question

In practical statistical terms, the question was whether the PFS experience differed between a fulvestrant-based strategy and an aromatase-inhibitor-based strategy in the intended mutation-positive population. The record does not establish that a particular combination was consistently given within either strategy: each arm description allows multiple CDK4/6 inhibitors, and the aromatase-inhibitor arm allows either letrozole or anastrozole. Details of the specific regimens received are not reported in the registry record.

3. Trial Design

INTERACT is registered as a phase II, randomized, open-label, parallel-group trial sponsored by M.D. Anderson Cancer Center. The registry gives a start date of 2019-05-31, a primary completion date of 2023-09-22, an enrollment of 4, and a status of terminated. The registry does not report the reasons for termination beyond noting poor accrual despite modifications and that the trial closed.

The treatment strategies were distinguished by endocrine therapy. Arm I included fulvestrant, with ribociclib, palbociclib, and/or abemaciclib. Arm II included letrozole or anastrozole, also with ribociclib, palbociclib, and/or abemaciclib. Thus, the registered comparison is between treatment strategies rather than a comparison of a single fixed drug combination against another single fixed combination. The actual treatment choices and their distribution among participants are not reported in the registry record.

The registry describes the study as randomized but does not report the randomization method, allocation ratio, or concealment procedures. Masking is listed as none, meaning that the study is registered as open label. The record does not describe whether outcome assessment was conducted without knowledge of assigned treatment. It also does not report a target enrollment or a sample-size justification.

The primary endpoint was PFS, with a stated timeframe of up to 30 days. The record says that no data were collected at the protocol-specified timepoint because accrual was poor and the trial closed. It lists secondary outcomes concerning circulating tumor DNA, ESR1 mutation emergence, cancer antigen 15-3, overall survival, and time to next treatment, but does not provide analyzed results for these outcomes. Further details about trial conduct, including any changes to the protocol, 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.

Before assignment: Of the 49 patients screened, 6 patients harbored ESR1 mutation. 4 patients were randomized to the study. 1 patient withdrew consent due to frequent travel/visits after randomization (AI arm). 1 patient was a screen fail due to elevated LFTs, and 1 patient passed away prior to study intervention)

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: No data was collected since despite modifications, accrual was poor and the closed. Outcome Measure has zero total analyzed at the protocol specified timepoint.

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 Progression-free Survival (PFS), measured over Up to 30 days.

Progression-free survival measures the time until disease progression or death, as defined by a trial’s protocol. It is used in cancer trials because it captures how long participants remain alive without documented worsening of disease. The exact event definitions, assessment procedures, and censoring rules for INTERACT are not reported in the registry record, so those details cannot be described more specifically.

The registry lists PFS as the primary endpoint, with a timeframe of up to 30 days. Its description proposes estimating survival curves with the Kaplan-Meier method, comparing groups with a log-rank test, and estimating a treatment comparison with Cox proportional hazards regression. A 95% confidence interval and p-value were planned for the regression comparison. The record also says the proportional hazards assumption would be assessed using graphs of rescaled Schoenfeld residuals and related tests.

These methods would ordinarily summarize the timing of events while accounting for participants whose follow-up ends before an event is observed. However, the registry states that no data were collected at the protocol-specified timepoint and that the endpoint has zero total analyzed. Consequently, no PFS curve, treatment comparison, or estimate of precision is available to interpret. The endpoint’s scientific purpose remains clear, but its planned analysis cannot yield a result without observed outcome data.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The planned PFS analysis combines methods for describing event timing with methods for comparing groups. The registry names Kaplan-Meier estimation, a log-rank test, and Cox proportional hazards regression. These are general statistical methods; the record does not report any fitted model, test result, survival estimate, or treatment-effect estimate for INTERACT because no primary outcome data were collected at the specified timepoint.

What does Kaplan-Meier estimation do?

In general, the Kaplan-Meier method estimates the probability of remaining event-free over follow-up. It uses each participant’s observed follow-up time and event status. Participants who have not had an event by their last observation can contribute information up to that point and are then treated as censored. This approach allows follow-up duration to vary, provided the assumptions behind censoring and the data collection process are reasonable.

A Kaplan-Meier curve is a description of the observed event pattern over time, not a guarantee about future outcomes. Its interpretation depends on clear event definitions, accurate timing, adequate follow-up, and an appropriate account of censoring. The registry does not report the detailed event, censoring, or assessment rules for this trial. With no analyzed PFS data, there is no curve to inspect for event patterns, follow-up, or separation between groups.

What does the log-rank test assess?

In general, the log-rank test compares event-time experience between groups across follow-up. It evaluates whether the pattern of observed events is compatible with the hypothesis of no difference in the survival distributions. It does not by itself describe the size of a treatment effect, and it does not explain why any difference might occur. Its usefulness depends on having observed event-time information and a meaningful group comparison.

The registry planned a log-rank test, but reports no analyzed PFS data. Consequently, there is no test statistic or p-value to interpret. A missing test result is not a nonsignificant result: it means the planned comparison was not reported as an analysis of observed endpoint data.

What would a Cox model add?

In general, Cox proportional hazards regression estimates a relative event rate over follow-up while allowing the underlying event rate to vary over time. The registry states that it planned to estimate the treatment comparison using this model and report a 95% confidence interval and p-value. A confidence interval conveys the range of effect estimates compatible with the data and model at the chosen confidence level; its width reflects statistical precision as well as the amount and pattern of information available.

The Cox model relies on the proportional hazards assumption: broadly, the relative event rate between groups is assumed to remain constant over time. The registry planned to assess that assumption using graphs of rescaled Schoenfeld residuals and related tests. In general, residual-based checks can identify patterns that may be inconsistent with the assumption, although such diagnostics require observed event and follow-up data. Since the primary analysis was not conducted, neither the model estimate nor its diagnostics are available.

Why does the absence of outcome data matter?

Randomization can support a fair comparison by balancing prognostic factors in expectation, but it does not create outcome observations. Statistical inference still requires data on what happened after assignment. With the primary endpoint unobserved at the specified timepoint, the intended randomized comparison cannot be estimated. Even if treatment groups were assigned, their assignment alone cannot establish the direction, size, or precision of a PFS effect.

Small amounts of information can also make estimates unstable and confidence intervals wide. That general concern is distinct from the central limitation in this record: the registry reports no analyzed PFS observations at all. It would be inappropriate to treat the absence of an estimate as an estimate near no effect, or to infer a result from the trial’s termination status.

How should planning illustrations be understood?

In general, sample-size calculations describe the amount of information that might be needed under specified assumptions. The registry’s assumptions describe a planning illustration based on the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. The stated effect to detect is a fraction of the observed effect. Such calculations are planning illustrations, not observed power and not results from the trial. The registry does not report a trial-specific sample-size justification. No planning calculation can supply the missing PFS observations or establish what this trial would have found.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists secondary outcomes covering circulating tumor DNA ESR1 mutant allele fraction and kinetics, emergence of ESR1 mutations, changes in cancer antigen 15-3, overall survival, and time to next treatment. It does not provide analyzed results for these outcomes. The primary PFS endpoint also has no analyzed data at the protocol-specified timepoint. This means there is no set of reported outcome estimates from which to judge consistency across endpoints.

When a study tests multiple outcomes or multiple comparisons, the chance of at least one apparently unusual finding can increase. General approaches to multiplicity include designating a primary endpoint in advance, adjusting statistical tests, or treating secondary findings as exploratory. The registry identifies PFS as primary, but a detailed multiplicity strategy is not reported in the registry record. The handling of multiple comparisons among treatment options and secondary endpoints is also not reported.

The record attributes the absence of PFS data to poor accrual despite modifications and closure. It does not provide a detailed account of missingness, follow-up completion, withdrawals, or the status of each secondary outcome. Those details are not reported in the registry record. There is therefore no basis for determining whether missing outcome information was related to prognosis, treatment assignment, or other factors.

The analysis population is not described beyond the statement that the primary endpoint had zero total analyzed. The registry does not say whether an intention-to-treat, per-protocol, or other population was planned for the final analysis. Nor does it report how participants with incomplete follow-up would have been handled. Without these details and without endpoint observations, neither a missing-data analysis nor an assessment of analysis-population choices can be made.

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 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports safety counts under the group labels “Fulvestrant” and “Continuing AI After Randomization.” In the Fulvestrant group, the serious adverse event fields record 1 at risk and 0 affected; the other adverse event fields record 1 at risk and 1 affected; deaths are recorded as 1. In the Continuing AI After Randomization group, the serious adverse event fields record 3 at risk and 0 affected; the other adverse event fields record 3 at risk and 3 affected; deaths are recorded as 0.

These are counts as recorded, not comparative rates or adjusted estimates. The registry does not provide event descriptions for serious adverse events in these entries, and the listed serious terms are empty. The “other” category should not be read as a particular type or severity of event when the record does not specify one. The meaning of the death counts in relation to cause or timing is not reported in the registry record.

The at-risk counts are small, so the counts alone offer limited information about safety patterns and cannot establish a reliable difference between groups. They also do not give a complete account of exposure duration or treatment received. No rate, percentage, or comparative risk estimate can be inferred from these entries without additional data and definitions.

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

The registry record documents a planned randomized comparison of fulvestrant-based and aromatase-inhibitor-based treatment strategies, each with CDK4/6 inhibitor options, for HR-positive metastatic breast cancer with an ESR1 activating mutation. It also documents that the trial was terminated and that poor accrual persisted despite modifications. Most importantly for interpretation, the record says that no PFS data were collected at the protocol-specified timepoint and that zero were analyzed for that endpoint.

These registry entries describe why the planned primary comparison is unavailable; they do not tell readers which strategy had better PFS. There is no reported treatment-effect estimate, confidence interval, p-value, or survival curve. The absence of those results cannot be interpreted as showing no difference. It also cannot be used to claim benefit, harm, equivalence, or non-inferiority.

The record’s enrollment entry is 4, but it does not report arm allocation or treatment exposure details. The adverse-event counts provide limited descriptive information under the registry’s group labels, but they do not replace the PFS analysis and should not be treated as comparative rates. The secondary outcomes are named, but analyzed results for them are not presented in the information recorded.

In general, a randomized trial can help estimate a causal treatment contrast when treatment assignment, follow-up, and outcomes are observed and analyzed appropriately. Randomization is a design feature, not an outcome. In INTERACT, the registered design identifies the intended comparison, while the missing primary endpoint data leave that comparison unanswered. The record therefore supports a conclusion about the availability of evidence from this trial, not a conclusion about the relative clinical effects of its treatment strategies.

13. Why This Trial Matters Statistically

INTERACT illustrates the difference between a well-defined research question and an answerable statistical comparison. The registry identifies a phase II randomized, parallel-group study and describes distinct endocrine strategies, with CDK4/6 inhibitor options in each. It also names PFS as the primary endpoint and specifies methods that would ordinarily support a time-to-event comparison. Those design elements clarify what investigators intended to learn.

But statistical methods cannot overcome the absence of observed outcome information. The registry reports poor accrual despite modifications, trial closure, and no PFS data collected at the specified timepoint. The result is not an imprecise estimate of the treatment effect; it is no reported primary estimate at all. This distinction matters when interpreting trial records, because a lack of a statistically significant result and a lack of an analysis are fundamentally different situations.

The record also underscores why treatment-strategy definitions matter. The arms allow multiple CDK4/6 inhibitor options, and the aromatase-inhibitor strategy permits letrozole or anastrozole. Without information on which regimens participants actually received, readers cannot describe the observed exposure pattern in detail. Even if outcome results had been available, those regimen details would be relevant to understanding what contrast the data represented.

Finally, the trial demonstrates why registration details about analysis populations, missing data, and multiplicity are useful. These elements help readers assess how planned analyses relate to reported results. In this case, several details are not reported in the registry record, and the primary endpoint has no analyzed data. The statistically responsible interpretation is consequently narrow: the study posed a clinically specific question, but its posted record does not provide an outcome-based answer.

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 principal limitation is the lack of primary endpoint data. The registry attributes this to poor accrual despite modifications and closure, and states that PFS had zero total analyzed at the protocol-specified timepoint. No effect estimate or uncertainty interval is available. The enrollment entry is 4, but group allocation and outcome follow-up are not described, so that entry does not clarify how much information was available for any comparison.

The treatment strategies were flexible rather than single fixed regimens in the arm descriptions. Both included ribociclib, palbociclib, and/or abemaciclib; one used fulvestrant, while the other used letrozole or anastrozole. The actual drug combinations received are not reported in the registry record. This limits the detail with which the intended contrast can be characterized as an exposure contrast.

The record does not report the randomization method, allocation ratio, concealment procedures, endpoint assessment rules, or analysis population. It also does not give detailed reasons for termination beyond poor accrual despite modifications and trial closure. These gaps prevent a fuller appraisal of conduct and analysis plans.

Secondary outcomes are listed, but their analyzed results are not reported in the registry information. The safety entries have small at-risk counts and limited event detail; causes and timing for the recorded deaths are not reported. The multiplicity strategy and detailed missing-data approach are also not reported. These limitations mean that the record cannot support a broad assessment of efficacy, safety, or the robustness of planned analyses.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Aromatase Inhibitor Therapy With or Without Fulvestrant for the Treatment of HR Positive Metastatic Breast Cancer With an ERS1 Activating Mutation, the INTERACT Study — NCT04256941.
• Registry record last updated 2024-12-27; status last verified 2024-12.
• Results first posted 2024-12-27.
• 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

INTERACT is registered as a phase II randomized, open-label, parallel-group trial sponsored by M.D. Anderson Cancer Center. It enrolled 4 and is listed as terminated. The registered start date is 2019-05-31 and the primary completion date is 2023-09-22.

The intended comparison was fulvestrant-based treatment versus letrozole- or anastrozole-based treatment, with ribociclib, palbociclib, and/or abemaciclib options in both strategies. The primary endpoint was PFS, with a timeframe of up to 30 days and planned Kaplan-Meier, log-rank, and Cox regression methods.

The registry states that poor accrual persisted despite modifications and that the trial closed without PFS data collected at the protocol-specified timepoint. The primary endpoint has zero total analyzed. No comparative PFS estimate or analyzed secondary outcome results are reported. The available safety entries are descriptive counts and do not provide a comparative treatment-effect analysis.