Current cardio-oncology risk assessment primarily focuses on treatment-related cardiotoxicity after initiation of anticancer therapy. However, many patients with breast cancer already exhibit substantial cardiovascular and metabolic vulnerability before treatment, which remains insufficiently characterized. We developed and internally validated a Pretreatment Cardio-Oncology Risk Score (PCORS) integrating cardiometabolic and tumour-related variables to improve baseline risk stratification in treatment-naïve breast cancer.
This retrospective single-centre cohort included women with newly diagnosed, treatment-naïve breast cancer between 2014 and 2024. Patients with implausible survival times were excluded. Baseline demographic, cardiovascular, metabolic, biochemical, electrocardiographic and tumour-related variables were incorporated into PCORS. Five-year all-cause mortality was used for internal validation. The cohort was randomly divided into training and validation datasets (70:30) using stratified sampling. Model discrimination, calibration, clinical utility and survival stratification were evaluated.
A total of 5,219 patients were included after data cleaning, including 381 all-cause deaths. The 5-year prediction cohort comprised 2,934 patients with 333 deaths. The integrated PCORS model demonstrated superior predictive performance compared with tumour-only and cardiometabolic-only models. The final model achieved an AUC of 0.954 in the training cohort and 0.926 in the validation cohort (95% CI 0.898–0.948), with good calibration and a validation Brier score of 0.065. Decision curve analysis demonstrated consistent clinical net benefit. Five-year mortality increased progressively across PCORS risk groups, with significant separation of Kaplan–Meier survival curves (log-rank P<0.001).
PCORS provides an integrated framework for pretreatment cardio-oncology risk assessment by combining tumour biology with baseline cardiovascular-metabolic vulnerability before systemic therapy. This approach may facilitate personalised baseline cardiovascular evaluation, identify patients requiring intensified cardio-oncology surveillance, and support individualized treatment planning in breast cancer.