Carotid intima media thickness measurements coupled with stroke severity strongly predict short-term outcome in patients with acute ischemic stroke: a machine learning study.

Acute ischemic stroke (IS) is one of the leading causes of morbidity, functional disability and mortality worldwide. The objective was to evaluate IS risk factors and imaging variables as predictors of short-term disability and mortality in IS. Consecutive 106 IS patients were enrolled. We examined the accuracy of IS severity using the National Institutes of Health Stroke Scale (NIHSS), carotid intima-media thickness (cIMT) and carotid stenosis (both assessed using ultrasonography with doppler) predicting IS outcome assessed with the modified Rankin scale (mRS) three months after hospital admission. Poor prognosis (mRS ≥ 3) at three months was predicted by carotid stenosis (≥ 50%), type 2 diabetes mellitus and NIHSS with an accuracy of 85.2% (sensitivity: 90.2%; specificity: 81.8%). The mRS score at three months was strongly predicted by NIHSS (β = 0.709, p < 0.001). Short-term mortality was strongly predicted using a neural network model with cIMT (≥ 1.0 mm versus < 1.0 mm), NIHSS and age, yielding an area under the receiving operator characteristic curve of 0.977 and an accuracy of 94.7% (sensitivity: 100.0%; specificity: 90.9%). High NIHSS (≥ 15) and cIMT (≥ 1.0 mm) increased the probability of dying with hazard ratios of 7.62 and 3.23, respectively. Baseline NIHSS was significantly predicted by the combined effects of age, large artery atherosclerosis stroke, sex, cIMT, body mass index, and smoking. In conclusion, high values of cIMT and NIHSS at admission strongly predict short-term functional impairment as well as mortality three months after IS, underscoring the importance of those measurements to predict clinical IS outcome.
(© 2021. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.)