Automated Evans index measurement using deep learning in acute subarachnoid hemorrhage: reliability, agreement with experts, and association with external ventricular drainage

1

The Evans index (EI) is used to assess ventricular enlargement and inform external ventricular drainage (EVD) decisions after subarachnoid hemorrhage (SAH).
2

TotalSegmentator (TS), a deep learning tool, was utilized to automate EI measurements from CT scans of 364 patients with spontaneous SAH.
3

TS demonstrated excellent reproducibility (ICC = 0.996) and good agreement with expert measurements (ICC = 0.76) for EI assessment.
4

TS identified more cases with EI > 0.30 compared to expert assessment (29% vs. 17%), indicating moderate agreement (Cohen’s kappa = 0.57).
5

Automated EI measurements using TS were independently associated with EVD placement, demonstrating clinical relevance in acute SAH management.

Frontiers In Neurology

by Yanrui Cai, Huansong Wang, Huanhuan Yu, Yuxia Li, Baobao Meng, Qi Liu, Yuting Wang, Feiyu Qiao
June 19, 2026

1