AUTHOR=Fast Lea , Temuulen Uchralt , Villringer Kersten , Kufner Anna , Ali Huma Fatima , Siebert Eberhard , Huo Shufan , Piper Sophie K. , Sperber Pia Sophie , Liman Thomas , Endres Matthias , Ritter Kerstin 

TITLE=Machine learning-based prediction of clinical outcomes after first-ever ischemic stroke

JOURNAL=Frontiers in Neurology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2023.1114360

DOI=10.3389/fneur.2023.1114360

ISSN=1664-2295

ABSTRACT=Background and Purpose: Accurate prediction of clinical outcomes in individual patients following acute stroke is vital for healthcare providers to optimize treatment strategies and plan further patient care. Here, we use advanced machine learning (ML) techniques to systematically compare the prediction of functional recovery, cognitive function, depression, and mortality for first-ever stroke patients and to identify their specific prognostic factors. 

Methods: We predicted clinical outcomes for 307 patients (151 females, 156 males; 68±14 years) from the PROSpective Cohort with Incident Stroke Berlin study using 43 baseline features. Outcomes included modified Rankin Scale (mRS), Barthel Index (BI), Mini-Mental State Examination (MMSE), Modified Telephone Interview for Cognitive Status (TICS-M), Center for Epidemiologic Studies Depression Scale (CES-D) and survival. The ML models included a Support Vector Machine with a linear kernel and a radial basis function kernel as well as a Gradient Boosting Classifier based on repeated 5-fold nested cross-validation. Key features for prognosis were identified using Shapley additive explanations. 

Results: The ML models achieved significant prediction performance for mRS at baseline and after one year, BI at baseline, MMSE at baseline, TICS-M after one and three years and CES-D after one year. Additionally, we showed that NIHSS was the top predictor for most functional recovery outcomes as well as education for cognitive function and depression.
 
Conclusion: Using a thorough ML analysis, we demonstrated not only that clinical outcomes after first-ever stroke can be significantly predicted but also identified key prognostic factors facilitating the prediction.