Machine Learning to Predict Delayed Cerebral Ischemia and Outcomes in Subarachnoid Hemorrhage

Abstract

Objective:To determine whether machine learning (ML) algorithms can improve the prediction of delayed cerebral ischemia (DCI) and functional-outcomes after subarachnoid hemorrhage (SAH).Methods:ML models and standard models (SM) were trained to predict DCI and functional-outcomes with data collected within 3 days of admission. Functional-outcomes at discharge and at 3-months were quantified using the modified Rankin scale (mRS) for neurological disability (dichotomized as ‘good’ (mRS≤3) vs ‘bad’ (mRS≥4) outcomes). Concurrently, clinicians prospectively prognosticated 3-month outcomes of patients. The performance of ML, SM and clinicians are retrospectively compared.Results:DCI status, discharge, and 3-month outcomes were available for 399, 393 and 240 subjects respectively. Prospective clinician (an attending, a fellow and a nurse) prognostication of 3-month outcomes was available for 90 subjects. ML models yielded predictions with the following AUC (area under the receiver operating curve) scores: 0.75 ± 0.07 (95% CI: 0.64 to 0.84) for DCI, 0.85 ± 0.05 (95% CI: 0.75 to 0.92) for discharge outcome, and 0.89 ± 0.03 (95% CI: 0.81 to 0.94) for 3-month outcome. ML outperformed SMs, improving AUC by 0.20 (95% CI: -0.02–0.4) for DCI, by 0·07 ± 0.03 (95% CI: -0.0018–0.14) for discharge outcomes, by 0.14 (95% CI: 0.03 –0.24) for 3-month outcomes and matched physician’s performance in predicting 3-month outcomes.Conclusion:ML models significantly outperform SMs in predicting DCI and functional-outcomes and has the potential to improve SAH management.

Cerebrospinal fluid hemoglobin drives subarachnoid hemorrhage-related secondary brain injury

“…While often used in clinical practice, established clinical (WFNS, Hunt & Hess) and radiological (BNI, modified Fisher) scores show a limited predictive accuracy for SAH-SBI (19)(20)(21)(22). The performance of combined scores or machinelearning-based models for predicting SAH-SBI is also only slightly better than that of the established clinical and radiological scores (23,24). The value of daily assessments of arterial flow velocity in the large cerebral arteries using TCD is controversial because of its limited diagnostic precision and high interrater variability (25,26) it cannot be applied in one-fifth of patients due to anatomical reasons (27).…”

Easily Created Prediction Model Using Automated Artificial Intelligence Framework (Prediction One, Sony Network Communications Inc., Tokyo, Japan) for Subarachnoid Hemorrhage Outcomes Treated by Coiling and Delayed Cerebral Ischemia

“…Park reported an AI-based model with an AUC of 0.77 using many variables, including vital signs and baseline characteristics with minimum redundancy maximum relevance algorithm [22]. Savarraj also reported an AI-based model with an AUC of 0.75 using clinical features [19].…”

“…items have been studied, and the AUCs for DCI occurrence are around 0.7 [18]. AI-based prediction models for DCI occurrence had AUCs of around 0.80 [12,[19][20][21][22]. However, reports on AI-based prediction models for DCI occurrence remain few.…”

Mast Cells Mediate Inflammatory Injury and Aggravate Neurological Impairment in Experimental Subarachnoid Hemorrhage Through Microglial PAR-2 Pathway

“…Subarachnoid hemorrhage (SAH), mainly caused by ruptured intracranial aneurysm, is a serious cerebrovascular disease with high mortality and chronic disability worldwide ( Macdonald and Schweizer, 2017 ; Hostettler et al, 2020 ). Although studies on SAH have been carried out for decades, the prognosis of patients with SAH remains unsatisfactory ( Rautalin et al, 2020 ; Savarraj et al, 2020 ). Among the multiple pathological events that occur after SAH, prolonged inflammation has been identified to be the critical factor accounting for the poor prognosis of SAH.…”