Evaluating the atherosclerosis cardiovascular disease risk score in patients with brain metastases: Associations with overall survival and high-value care outcomes

Abstract

Objective

Brain metastases (BM) constitute the most common intracranial tumor in adults. Prior literature indicates the 10-year atherosclerotic cardiovascular disease (ASCVD) risk score is associated with increased risk of cancer, potentially attributable to shared risk factors. Understanding the role of ASCVD risk scores in BM may help optimize their care and inform clinical decision-making. Our aim was to explore associations between ASCVD risk score in BM patients and their overall survival, hospital charges, and non-routine discharge disposition.

Methods

Electronic medical records were reviewed to collect clinical data for BM patients undergoing surgery at a single institution (2017–2021). Regression analyses were performed accordingly and maximally selected rank statistics were employed to identify an optimal cutoff for ASCVD risk scores. The random survival forest (RSF) machine learning technique identified the most important variable associated with survival outcomes in BM patients.

Results

A total of 139 patients were included with average age 62.93±9.29 years, 48.2 % male, 25.2 % with high hospital charges, and 23.7 % experiencing non-routine discharge. Among these patients, 32.3 % had prior history of an ASCVD event, while 67.7 % did not. Overall, this cohort had an average 10-year ASCVD risk score of 12.51±12.98, indicating intermediate risk of ASCVD among all BM patients. On multivariate logistic regression, prior history of ASCVD was associated with higher odds of high hospital charges (OR=3.670, p=0.018), and higher ASCVD risk scores were associated with greater odds of non-routine discharge (OR=1.059, p=0.012). On the multivariate Cox regression model, higher ASCVD risk scores correlated with worse overall survival (HR=1.031, p=0.014). A threshold of 25.1 was identified for high-risk ASCVD scores. Patients with ASCVD scores >25.1 exhibited reduced overall survival in Kaplan-Meier analysis (p=0.015) and multivariate Cox regression (HR: 2.811, p=0.016). Notably, ASCVD risk scores were found to be the most important variable in predicting worse survival outcomes in BM patients compared to other established frailty indices.

Conclusion

This study indicates higher ASCVD risk scores in BM patients are associated with worse overall survival. Integrating ASCVD assessment into clinical workflow may facilitate more informed risk-based decision-making.