Study confirms EHR‑CAT score for VTE risk

By Litha Mfiki

A nationwide US study published in JAMA Network Open has validated the electronic health record cancer-associated thrombosis (EHR-CAT) score for predicting venous thromboembolism (VTE) in patients with cancer receiving systemic therapy. Analyzing data from more than 700,000 patients, the researchers found that the EHR-CAT score outperformed the widely used Khorana score.

Ang Li (Baylor College of Medicine, Houston, Texas, USA) and colleagues evaluated 732,594 patients with newly diagnosed, invasive, solid or hematologic malignant neoplasms treated between 2018 and 2023 across 184 US health systems. Using standardized EHR data, the team compared the performance of the EHR‑CAT score with the Khorana score for predicting 6‑month VTE risk.

The EHR‑CAT model demonstrated strong discrimination (C statistic, 0.70) compared with 0.63 for the Khorana score, and reclassified 20% of patients into more accurate risk categories. “The EHR‑CAT risk score demonstrated a robust performance across different sites and accurately classified patients into groups with 6‑month incidences of overall VTE ranging from 1.3% to 11.1%,” the authors note.

At 6 months, VTE incidence was 4.7%, bleeding incidence was 3.7%, and mortality was 8.4%. Notably, the model remained reliable even after exclusion of patients at high bleeding risk, a factor that has traditionally limited the use of thromboprophylaxis in routine practice. These included elevated bilirubin (HR, 2.78), acute leukemia (HR, 2.33), metastatic brain cancer (HR, 2.17), prior bleeding (HR, 2.24), and reduced renal function (glomerular filtration rate <30 mL/min/1.73 m²; HR, 2.19). The investigators state that “the EHR‑CAT risk score performed well even after excluding 26% of patients at risk for bleeding using common anticoagulant exclusion criteria.”

The authors also reported that VTE incidence remained stable at 4.5–5.0% over the 6‑year study period, with no evidence of data drift. Calibration remained strong for pulmonary embolism and lower‑extremity deep vein thrombosis outcomes.

The authors write, “These results suggest that standardized structured EHR data from different health systems can support scalable validation and implementation of CAT risk assessment.”

Developed by EPG Health for Medthority, independently of any sponsor.