391P - Detrimental effect on overall survival of CDK4/6 inhibitor dose reduction if immortal time bias is considered

Background

Since 2017, the combination of CDK4/6 inhibitors and endocrine therapy has been the standard first-line treatment for patients diagnosed with hormone-positive/HER2-negative metastatic breast cancer (mBC) in Denmark. Previous analyses, based on randomized phase III studies, have shown that the clinical benefit remains after CDK4/6i dose reduction, but analyses may have been susceptible to immortal time bias.

Methods

This real-world cohort study is based on electronic health records from 695 patients from the capitol region of Denmark who initiated first-line treatment for mBC with CDK4/6i between May 2017 and October 2022. We used landmark Kaplan-Meier estimates and Cox regression to examine the association of dose reduction with treatment duration and overall survival. Patients were categorized based on dose reduction from the recommended dose prior to the landmark, defined at 3 months after initiation of treatment. Hazard ratios were adjusted for CDK4/6i agent, age, comorbidity, BMI, LDH > ULN, ECOG performance status and visceral metastases.

Results

A total of 530 patients were eligible for inclusion in the landmark analysis and 198 (37%) were categorized into the dose reduction group. Patients with dose reduction were significantly older and more had received prior endocrine therapy. Dose reduction was significantly associated with overall survival and shorter treatment duration, 39.6 vs 55.6 months (95%CI: 35.5-NE vs 49.9-NE) and 19.6 vs. 30.3 months, respectively. Adjusted hazard ratio for death was 1.49 for patients with dose reduction (95% CI: 1.01-2.22).

Conclusions

Dose reduction of CDK4/6i within the first 3 months was associated with significantly higher mortality and shorter treatment duration. These findings contrast previous analyses that have shown no effect of dose reduction on treatment efficacy. Consideration of immortal time bias in this study may be a decisive reason for the disparate findings. Table: 391P

¹n (%); Median (IQR). ²Modified to exclude solid tumors. ³Manual review of a random sample of 75 patients in each group.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.