White matter hyperintensities regress at a high rate at three months after minor ischemic stroke or transient ischemic attack

Background: The potential for early white matter hyperintensities(WMH) regression and associated contributory factors remains uncertain. The purpose of this study is to investigate whether WMH regress at early time of three months after minor ischemic stroke (MIS) or transient ischemic attack (TIA), while also identifying factors that may influence this outcome. Methods: A retrospective analysis of a prospective subcohort from the CHANCE trial comprising individuals with MIS and TIA was conducted. All patients underwent brain MRI at the onset and at three months. Deep learning algorithms were employed for the automatic segmentation of WMH volumes in four distinct regions. Scores for lacunes, cerebral microbleeds (CMB), perivascular spaces (PVS), WMH, and overall cerebral small vessel disease (CSVD) burden were quantified. Patients were divided into the stable, regression and progression groups according to change in WMH volume. The demographic, clinical, and imaging data of the participants in the three groups were collected and statistically analyzed. Results: A total of 98 patients with minor ischemic stroke or TIA were included. There were 22 patients in the stable group, 41 patients in the regression group and 35 patients in the progression group. Age and hypertension status were significantly different among the three groups. The lacunes, CMB,WMH, and total CSVD burden scores differed notably among groups, with all the CSVD markers being severely elevated in the progression group, moderately elevated in the regression group, and subtly elevated in the stable group. Conclusion: The findings suggest that WMH could exhibit regression within three months following minor ischemic stroke or TIA. Patients under the age of 65, without a hypertension history, and with a low CSVD burden are more likely to experience WMH regression. (c) 2024 Elsevier Masson SAS. All rights are reserved, including those for text and data mining, AI training, and similar technologies.