Prevalence of dosage modifications and impact on response milestone among patients with CML treated with imatinib using RWD.

e23335 Background: Continuous treatment with BCR-ABL1 TKIs are the cornerstone of chronic myeloid leukemia (CML) treatment. This study aims to assess the prevalence of TKI dosage modifications, including interruptions and reductions, and their impact on treatment effectiveness. Methods: This retrospective observational study utilized the UNC Clinical Data Warehouse electronic health record (EHR) database. Eligible patients include adults (≥18 years) diagnosed with CML receiving imatinib as first-line therapy for >1 month between 2010-2023. Patients with incomplete dosing data were excluded. Descriptive statistics include the types and prevalence of dosage modifications, including dose reduction, dosage interruption, and drug discontinuation. Efficacy analyses were evaluated using BCR-ABL transcripts and early response milestones at 3, 6, and 12 months. Statistical analyses were performed using Fisher’s exact test, with a two-sided α level of 0.05. Results: A total of 118 patients were included with a median age of 61 years, mostly female (55%), and racially diverse (59% White; 26% Black; 2% Asian; 1% American Indian; 7% Hispanic; 5% Other/Unknown). Of these patients, 59 patients (50%) experienced a total of 112 all-cause dosage modification events during the first year of use: dose reductions (26%), dosage interruptions (46%), dose discontinuations (25%), and others (3%). Fifty-two percent (58/112) of dose modification events occurred in the first 6 months. Additionally, 39 patients (66%) experienced dosage modifications due to toxicities, while 4 patients (3%) initiated treatment at a reduced dose of imatinib. Among the patients with evaluable response data, 54/84 (64%) patients achieved an early molecular response (BCR-ABL <10%) at month 3, 48/72 (67%) patients achieved <1% at month 6, and 39/67 (58%) patients achieved <0.1% at month 12. Dosage modifications were associated with reduced likelihood of achieving 6-month milestones (p < 0.05) but were not significantly associated with outcomes at 3 (p =0.33) or 12 months (p = 0.52). Conclusions: We demonstrated the feasibility of leveraging a complex EHR data source from an integrated health system to study the impact of dose modifications on clinical outcomes for self-administered chemotherapy. Nearly half of the patients in our study cohort experienced dosage modifications. While these modifications were associated with a lower likelihood of achieving 6-month molecular response milestones, this finding has various limitations due to the sample size, retrospective nature, and lack of adjustment for multiple comparisons in this study. Our cohort of patients experienced more treatment discontinuations due to toxicities when compared to clinical trials. Additional analyses are needed to further evaluate the long-term impact and causal effect of dosage modifications due to toxicities on clinical outcomes.