Nomogram predicting the cardiovascular disease mortality for older patients with colorectal cancer: A real-world population-based study.

Abstract

Background: Cardio-oncology has received increasing attention especially among older patients with colorectal cancer (CRC). Cardiovascular disease (CVD)-specific mortality is the second-most frequent cause of death. The risk factors for CVD-specific mortality among older patients with CRC are still poorly understood.

Aim: To identify the prognostic factors and construct a nomogram-based model to predict the CVD-specific mortality among older patients with CRC.

Methods: The data on older patients diagnosed with CRC were retrieved from The Surveillance, Epidemiology, and End Results database from 2004 to 2015. The prognostic factors and a nomogram-based model predicting the CVD-specific mortality were assessed using least absolute shrinkage and selection operator and Cox regression.

Results: A total of 141251 eligible patients with CRC were enrolled, of which 41459 patients died of CRC and 12651 patients died of CVD. The age at diagnosis, sex, marital status, year of diagnosis, surgery, and chemotherapy were independent prognostic factors associated with CVD-specific mortality among older patients with CRC. We used these variables to develop a model to predict CVD-specific mortality. The calibration curves for CVD-specific mortality probabilities showed that the model was in good agreement with actual observations. The C-index value of the model in the training cohort and testing cohort for predicting CVD-specific mortality was 0.728 and 0.734, respectively.

Conclusion: The proposed nomogram-based model for CVD-specific mortality can be used for accurate prognostic prediction among older patients with CRC. This model is a potentially useful tool for clinicians to identify high-risk patients and develop personalized treatment plans.