Deep learning can estimate cardiovascular (CV) risk from a chest X-ray image that is similar to the current clinical standard, according to a study presented at the annual meeting of the Radiological Society of North America. Jakob Weiss, MD, and colleagues developed the deep learning model CXR-CVD risk to estimate 10-year CV risk from a routine chest radiograph. Validation occurred in an independent cohort of 11,430 outpatients potentially eligible for primary prevention (LDL cholesterol, 70-190 mg/dL; no prevalent diabetes; no prior
major adverse CV events [MACE]). In the validation cohort, 1,096 MACE occurred during a median follow-up of 10.3 years (9.6% of the cohort). In statin-eligible patients, there was a significant association between CXR-CVD risk and MACE (HR, 2.03), which remained significant when adjusting for CV risk factors (adjusted HR [aHR], 1.63). CXR-CVD risk performed similarly to the atherosclerotic CV disease (ASCVD) risk score and added to the ASCVD risk score (aHR, 1.58). “Based on a single existing chest X-ray image, our deep learning model predicts future major adverse cardiovascular events with similar performance
and incremental value to the established clinical standard,” Dr. Weiss said in a statement.

EHR Data Can Predict Readmission in Children 

A risk prediction model may identify infants and children at risk for hospital readmission, according to a study published in JAMA Network Open. Denise M. Goodman, MD, and
colleagues developed and validated a tool for identifying patients before hospital discharge who are at risk for subsequent readmission for all ages. The derivation set was based on 29,988 patients (48,019 hospitalizations). Among children aged 6 months and older with one or more prior hospitalizations within the last 6 months (recent admission), prior utilization, current or prior procedures indicating illness severity (transfusion, ventilation, or central venous catheter), commercial insurance, and prolonged length of stay (LOS) were associated with readmission. Among children older than 6 months with no prior hospitalizations in the last 6 months (new admission), procedures, prolonged LOS, and an ED visit in the past 6 months were associated with reduced readmission. Among children younger than 6 months (young infants), LOS, prior visits, and critical procedures were associated with readmission. “These models may allow future improvements in tailored discharge preparedness to prevent high-risk readmissions,” Dr. Goodman and colleagues wrote.