Unsupervised learning based on CT imaging in identifying subtypes of acute ischemic stroke severity
- 1Department of Radiology, The Third People’s Hospital of Chengdu, Chengdu, Sichuan PR China
Abstract
Purpose : To explore unsupervised learning based on CT imaging for identifying severity subtypes of acute ischemic stroke (AIS) in the middle cerebral artery territory.
Methods: One hundred seventy-three AIS patients were retrospectively enrolled. K-means clustering was performed on clinical and CT features to identify severity subtypes. NIHSS scores were compared between subtypes. Hemorrhagic transformation (HT) served as an adverse outcome to compare associations of cluster subtypes versus NIHSS scores.
Results: Two subtypes were identified. Subtype 1 (severe) showed larger infarct volumes, higher rates of basal ganglia involvement and dense middle cerebral artery sign, longer onset time, and younger age than Subtype 2 (non-severe) (all p < 0.001). Median NIHSS score was higher in Subtype 1 [17.0 (14.0,19.0) vs. 12.0 (8.0,17.0), p < 0.001]. HT identification accuracy was 70% for cluster subtypes vs. 67% for NIHSS scores. AUC for HT was 0.692 (subtypes), 0.679 (NIHSS), and 0.741 (combined). For poor 90-day functional outcome (mRS ≥ 3), AUC was 0.71 (subtypes), 0.69 (NIHSS), and 0.76 (combined).
Conclusions: Unsupervised learning based on CT imaging effectively identifies AIS severity subtypes and may serve as an objective method for assessing post-stroke neurological damage.
Related
- Difference of onset rhythm in acute ischemic stroke patients with different TOAST subtypes
- Unsupervised Classification of Wide-Field Swept-Source Optical Coherence Tomography Reveals Patterns Associated With Glaucoma Severity and Progression.
- Unsupervised Global CBC/RUO/CPD Phenotyping Identifies Haematological Clusters Enriched for Thrombocytopenia Severity and Mechanisms
- Retinal nerve fibre layer, and ganglion cell layer thicknesses in children with ADHD: associations with subtypes and symptom severity
- Identifying Cardiac Autoantibody Biomarkers that Predict the Severity of Chagas Cardiomyopathy.
BibTeX
@article{pan2026unsupervised,
title = {{Unsupervised learning based on CT imaging in identifying subtypes of acute ischemic stroke severity}},
author = {Dun Pan and Gang Xie},
journal = {BMC Neurology},
year = {2026},
month = aug,
doi = {10.1186/s12883-026-05250-4},
url = {https://doi.org/10.1186/s12883-026-05250-4},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Unsupervised learning based on CT imaging in identifying subtypes of acute ischemic stroke severity
AU - Dun Pan
AU - Gang Xie
PY - 2026
DA - 2026/08/01
JO - BMC Neurology
DO - 10.1186/s12883-026-05250-4
UR - https://doi.org/10.1186/s12883-026-05250-4
AB - Purpose : To explore unsupervised learning based on CT imaging for identifying severity subtypes of acute ischemic stroke (AIS) in the middle cerebral artery territory. Methods: One hundred seventy-three AIS patients were retrospectively enrolled. K-means clustering was performed on clinical and CT features to identify severity subtypes. NIHSS scores were compared between subtypes. Hemorrhagic transformation (HT) served as an adverse outcome to compare associations of cluster subtypes versus NIHSS scores. Results: Two subtypes were identified. Subtype 1 (severe) showed larger infarct volumes, higher rates of basal ganglia involvement and dense middle cerebral artery sign, longer onset time, and younger age than Subtype 2 (non-severe) (all p < 0.001). Median NIHSS score was higher in Subtype 1 [17.0 (14.0,19.0) vs. 12.0 (8.0,17.0), p < 0.001]. HT identification accuracy was 70% for cluster subtypes vs. 67% for NIHSS scores. AUC for HT was 0.692 (subtypes), 0.679 (NIHSS), and 0.741 (combined). For poor 90-day functional outcome (mRS ≥ 3), AUC was 0.71 (subtypes), 0.69 (NIHSS), and 0.76 (combined). Conclusions: Unsupervised learning based on CT imaging effectively identifies AIS severity subtypes and may serve as an objective method for assessing post-stroke neurological damage.
SN - 1471-2377
ER -