Risk Prediction Model for Development of Heart Failure or Cardiomyopathy After Breast Cancer Treatment.

Area of Science:

• Cardiology
• Oncology
• Preventive Medicine

Background:

• Women undergoing breast cancer (BC) treatment face increased risk of heart failure or cardiomyopathy (HF/CM).
• Currently, no standardized method exists to identify high-risk women for proactive cardiac surveillance and intervention during and post-treatment.

Purpose of the Study:

• To develop and validate a predictive model for 10-year HF/CM risk in women receiving systemic treatment for early-stage invasive BC.
• To inform cardiac risk management strategies for this patient population.

Main Methods:

• A longitudinal cohort study was conducted within Kaiser Permanente Southern California.
• A multivariable elastic-net Cox proportional hazards model was used to predict 10-year HF/CM risk.
• The cohort included 26,044 women aged 18-79 with newly diagnosed invasive BC (2008-2020), randomly split into derivation (60%) and validation (40%) groups, with a median follow-up of 5.2 years.

Main Results:

• The risk model demonstrated good calibration and high accuracy in predicting HF/CM risk across low-, moderate-, and high-risk subgroups.
• In the validation cohort, the model accurately estimated 10-year HF/CM risk: 1.7% for low-risk and 19.4% for high-risk women.
• The model's discrimination ability was good, with a time-dependent area under the curve of 0.79 at 10 years in the validation cohort.

Conclusions:

• A validated risk prediction model can prospectively identify women with early-stage BC at risk for HF/CM over a 10-year period.
• The model utilizes BC treatment and clinical variables available at diagnosis.
• This tool can guide risk-stratified cardiac management for women undergoing breast cancer treatment.