Abstract Accurate risk stratification is crucial in cardio-oncology to identify patients at high risk of cancer therapy-related cardiac dysfunction (CTRCD) and implement early preventive strategies. While strain imaging during follow-up has proven valuable in predicting CTRCD, its role in baseline risk assessment remains unclear. Evaluating left ventricular, right ventricular and left atrial strain at baseline could enhance initial risk prediction and refine the HFA-ICOS score, improving patient selection for closer monitoring and early cardio-oncology referrals. This study aimed to assess whether incorporating baseline strain measurements could improve the predictive accuracy of the HFA-ICOS risk score for anthracycline cardiotoxicity in the CARDIOTOX registry. A single-center, observational, retrospective case-control study was conducted, including randomly selected patients with and without clinical relevant CRTCD (defining cases as symptomatic or moderate to severe asymptomatic CRTCD) who had an adequate baseline echocardiogram for strain analysis. The new Strain-Integrated Cardio-Oncology (SICOS) score was developed by modifying the HFA-ICOS score, adding one grade if any baseline strain parameter was abnormal and subtracting one if all were withing normal values. Baseline deformation parameters included left ventricle global longitudinal strain (LVGLS), free-wall right ventricle strain (FWRVS), 4-chamber right ventricle strain (4CRVS), left atrial reservoir, conduct and contraction (LAr, LAcd, LAct respectively). Strain was measure using Phillips TOMTEC/Qlabâ software. Data were analysed using R software. 178 Patients were selected from CARDIOTOX-Registry. We included 57 cases and 121 randomly selected matched controls. Both groups were well matched, with no significant differences in age, gender, cardiovascular risk factor, baseline treatment, tumor type, cumulative doxorubicin-equivalent dose or HFA-ICOS score (Table 1). Baseline GLS and LAcd were significantly higher in the control group (GLS -19% vs. -18%, p=0.03; LAcd -23% vs. -18%, p=0.015), while LAr and 4CRVS showed a non-significant trend toward better values in controls (42% vs. 37%, p=0.065; -20% vs. -19%, p=0.061). LAcd demonstrated the best AUC for all strain measures (0.64(0.52-0.7)), similar to the left ventricle ejection fraction (AUC 0.64(0.55-0.73)). 4CRVS showed better AUC (0.59(0.5-0.69)) than FWRVS (0.57(0.48-0.66)). GLS, LAcd and 4CRVS were included in the modified score. The SICOS score reclassified 5 cases from moderate to high risk and 7 controls, 6 from high to moderate risk and 1 from moderate to high risk (Figure 1). Baseline LAcd emerged as the most predictive imaging parameter for CTRCD. Integrating LAcd, GLS, and 4CRVS into de SICOS score enhances risk stratification, allowing for a more precise discrimination between moderate- and high-risk patients.
Gonzalez et al. (Fri,) studied this question.