AUTHOR=Rajashekar Deepthi , Hill Michael D. , Demchuk Andrew M. , Goyal Mayank , Fiehler Jens , Forkert Nils D. 

TITLE=Prediction of Clinical Outcomes in Acute Ischaemic Stroke Patients: A Comparative Study

JOURNAL=Frontiers in Neurology

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2021.663899

DOI=10.3389/fneur.2021.663899

ISSN=1664-2295

ABSTRACT=Clinical stroke rehabilitation decision-making relies on multi-modal data, including imaging and other clinical assessments. However, most previously described methods for predicting long-term stroke outcomes do not make use of the multi-modal data available. The aim of this work was to develop and evaluate the nested regression models that utilize clinical assessments as well as image-based biomarkers to model 30-day NIHSS.

221 subjects were pooled from two prospective trials with follow-up MRI or CT scans, and NIHSS assessed at baseline, 48-hours, and 30-days after symptom onset. Three prediction models for 30-day NIHSS were developed using a support vector regression model: one clinical model based on modifiable and non-modifiable risk factors (MCLINICAL) and two nested regression models that aggregate clinical and image-based features and differed with respect to the method used to select the important brain regions for the modeling task. The first model used the widely accepted RreliefF machine learning method (MRELIEF), while the other employed a lesion-symptom mapping technique (MLSM) often used in neuroscience to investigate structure-function relationships and identify eloquent regions in the ischemic brain.

The two nested models achieved a similar performance while outperforming MCLINICAL. However, MRELIEF required fewer brain regions and achieved a lower mean absolute error than MLSM while being less computationally expensive.

Aggregating clinical and imaging information improves outcome prediction models. While LSM is a useful tool to investigate structure-function relationships of the brain, it does not lead to better outcome predictions compared to the less computationally expensive and easier to implement, feature-selection approach.