Cardiovascular risk-prediction models in cancer survivors: a meta-analysis.

In 30 seconds

This meta-analysis evaluated 34 studies involving 983,438 adult cancer survivors to assess cardiovascular risk-prediction models. Approximately one-third of the models showed acceptable-to-good discrimination (AUC ≥0.70). Notably, the New Zealand CVD score was the only composite model with an AUC ≥0.70, indicating its potential utility in clinical practice.

Key findings

• One-third of evaluated models demonstrated acceptable-to-good discrimination with AUC ≥0.70.
• The New Zealand CVD score was the only composite outcome model with AUC ≥0.70.
• Cancer-specific scores like HFA-ICOS showed high specificity but lower sensitivity for ruling out CTRCD/heart failure.
• Model performance varied significantly by cancer type, with better results in breast cancer cohorts.

Why it matters

Understanding the performance of cardiovascular risk-prediction models in cancer survivors is crucial for clinicians to make informed decisions regarding primary cardiovascular prevention, especially given the unique risks associated with cancer therapies.