Validation of the HFA-ICOS cardiovascular toxicity risk tool in chronic myeloid leukemia patients receiving tyrosine kinase inhibitors: Real-world data from the Saudi population.

Abstract

e24009 Background: Tyrosine kinase inhibitors (TKIs) are the cornerstone treatment in patients with chronic myeloid leukemia (CML) but are associated with cardiotoxic effects. This study aimed to validate the Heart Failure Association (HFA) and the International Cardio-Oncology Society (ICOS), a tool aims to assess cardiotoxicity due to TKI use. The use of this tool is recommended by the European Cardio-Oncology guidelines, however, the evidence regarding its effectiveness in a real world cohort and its applicability to the Saudi population is limited. Methods: This is a retrospective chart review study conducted at one single tertiary care hospital in Saudi Arabia. Data from 2016 to 2024 were collected for adult CML patients treated with TKIs for at least six months. Patients were stratified by the HFA-ICOS score; low, medium, high, and very high risk. C-statistics were calculated to evaluate the model's discriminative ability, while incidence rates were analyzed to assess the association between risk stratification categories and the occurrence of cardiotoxicity events. Statistical significance was determined with a p-value of <0.001. Statistical analysis was performed using SPSS 22.0. Results: Of 304 CML patients, 93 patients were included, of which 55.9% were females and with a median age of 59 (IQR 46-67). According to the HFA-ICOS criteria, 41 patients (44.1%) were classified as low risk, 15 (16.1%) as moderate risk, 31(33.2%) as high risk, and 6 (6.5%) as very high risk. Additionally, 14% had cardiovascular diseases before receiving TKI. Over a median of 60 month period of follow-up, the C-statistic for predicting cardiotoxicity was 0.862 (95% CI: 0.781–0.944), indicating excellent discriminative ability. This demonstrates that the predictive model effectively distinguishes between individuals who experienced cardiotoxicity events and those who did not, highlighting its strong performance in risk stratification. Incidence rates of cardiotoxicity events revealed a statistically significant association with the HFA-ICOS risk stratification categories: 0% in the low-risk group, 13.3% in the medium-risk group, 29% in the high-risk group, and 83.3% in the very high-risk group (p-value <0.001). These findings highlight the strong performance of the risk stratification model in predicting cardiotoxicity events. Risk factors for developing cardiotoxicity events were older age, and pre-existing cardiovascular disease. Conclusions: This study demonstrates the excellent discriminative ability of the predictive model and a significant association between HFA-ICOS risk categories and cardiotoxic events in CML patients receiving TKIs. The model effectively stratifies patients by cardiovascular risk, highlighting the importance of older age and pre-existing cardiovascular disease as key risk factors.