Table 1_Evaluation of small vessel disease burden on MRI and stroke outcomes.docx

PurposeWe aimed to assess whether a composite SVD score derived from MRI features improves stroke outcome prediction when integrated with clinical factors.
Materials and methodsA 2019–2022 retrospective analysis included patients who had MRI prior to stroke admission. A semi-automated approach evaluated SVD MRI markers (white matter hyperintensities (WMH), lacunar infarcts, perivascular spaces (PVS), and cerebral microbleeds (CMBs)) using continuous and categorical measures to create a composite SVD score. Multivariate regression analyses compared performance across three models: (1) SVD score, (2) clinical factors, and (3) SVD score + clinical factors for outcomes, including stroke etiology, ICU and hospital stay, NIH Stroke Scale (NIHSS), 90 day modified Rankin Scale (mRS), functional independence (mRS < 2), and stroke recurrence.
ResultsForty-eight patients were included in this study. The combined SVD + clinical factors model outperformed other models in predicting functional independence with area under the curve (AUC) 0.58 (95% CI 0.41–0.75) and stroke etiologies of large artery atherosclerosis AUC 0.78 (0.62–0.91), small vessel occlusion 0.65 (0.41–0.88), and other determined etiology 0.74 (0.37–0.96). The combined model also better predicted the following outcomes with lower mean absolute error (MAE): NIHSS MAE 5.16 (3.80–6.69), ICU days 1.26 (0.86–1.66), total length of stay 2.62 (2.08–3.17), and 90 day mRS 1.74 (1.39–2.12).
ConclusionCombining an SVD score with clinical variables improved prediction of stroke outcomes when compared to either predictor alone, although the added value is modest and requires further validation. Nonetheless, integrating an SVD score into clinical practice may guide management in acute stroke settings and support risk stratification and prognostication.