PHILA-Score: retrospective evaluation of cardiotoxicity risk in spanish Chronic Myeloid Leukemia patients treated with first-line BCR-ABL Tyrosine Kinase Inhibitors
J Ducay Rico, N Rahma Almaraz, V Garcia Gutierrez, A Luna De Abia, I Gamez Guijarro, F J Lopez Jimenez, J L Zamorano Gomez, P Agudo QuilezAbstract
Background
BCR-ABL tyrosine kinase inhibitors (TKIs) are standard therapy for chronic myeloid leukemia (CML), improving survival but potentially increasing cardiovascular (CV) risk. Data on CV events, baseline risk, and predictive performance of risk scores in Spain CML patients remain limited.
Objectives
To assess the predictive value of SCORE2/SCORE2-OP and HFA-ICOS BCR-ABL scores for cardiotoxicity in CML patients treated with first-line TKIs in a predominantly low CV risk population.
Methods
This single-center observational study retrospectively included consecutive CML patients treated with TKIs from January 2014 to December 2024. Baseline demographics, CV comorbidities were collected. Cardiotoxicity was defined according to HFA-ICOS criteria. SCORE2/SCORE2-OP and HFA-ICOS for BCR-ABL TKI scores were calculated when possible. Associations with cardiotoxicity were assessed using chi-square and Mann-Whitney U tests.
Results
A total of 141 patients were included (mean age 51±17 years; 56% male). Baseline characteristics showed 25.5% hypertension, 8.5% diabetes mellitus (DM), 2.9% ischemic heart disease, 2.2% atrial fibrillation/flutter, and 6.4% chronic kidney disease. 3.5% of patients had a combined history of heart failure, cardiomyopathy, or cancer therapy–related cardiac dysfunction (CTRCD). Most patients had low (46.8%) to intermediate (32.3%) Sokal risk. First-line TIs were imatinib (66%), nilotinib (20.6%), dasatinib (9.9%), bosutinib (2.8%), and ponatinib (0.7%). Overall cardiotoxicity occurred in 19.9% patients, mostly mild congestion/heart failure (13.5%) and hypertension (5%). Severe events were less common: significant LV dysfunction (1.4%), arrhythmias (3%), arterial thrombosis (1.4%), and pulmonary hypertension (0.7%). Cardiotoxicity rates varied by TKI but were not statistically significant (p=0.374). SCORE2/SCORE2-OP was calculable in 30.7% of patients and was not associated with cardiotoxicity (p=0.855). The HFA-ICOS BCR-ABL score was calculable in 138 patients, categorizing 17.4% as low, 51.4% as moderate, 23.2% as high, and 8.0% as very high risk. Cardiotoxicity incidence increased from moderate (12.7%) to very high risk (36.4%) but did not reach statistical significance (p=0.116, chi-square). Analysis of individual HFA-ICOS components showed significant associations for hypertension (p=0.019) and DM (p=0.048). Other classical CV factors, including age and sex, did not show statistically significant associations. There was a trend toward higher cardiotoxicity in patients with a history of any of heart failure, cardiomyopathy, or CTRCD (p=0.054).
Conclusions
In this presumed low CV risk population, first-line BCR-ABL TKIs cause cardiotoxicity in ~20%, mostly mild. Trends in HFA-ICOS scores suggest potential predictive value. These findings highlight the importance of baseline CV assessment and monitoring. Larger prospective studies are needed to validate these findings and optimize CV risk stratification in this population.