White Matter Hyperintensity Load Independent From the Stroke Lesion Is Associated With Chronic Aphasia Severity and Treatment Outcome

BACKGROUND: Although many studies have suggested that white matter hyperintensity (WMH) severity predicts naming and aphasia severity in chronic poststroke aphasia, there are inconsistencies in the literature. WMHs are typically symmetrical in neurotypical controls, and measuring WMH in the contralateral hemisphere is likely the best option to estimate brain health independently from the stroke lesion and avoid measurement contamination from stroke-related gliosis. In this study, we aimed to clarify the discrepancies in the literature by testing whether WMH rating methods are related to clinical outcomes. METHODS: Ninety-five participants with chronic aphasia at least 12 months after their left-hemisphere stroke completed a baseline Western Aphasia Battery and the Philadelphia Naming Test. All participants then underwent 6 weeks of phonological and semantic naming treatments focused on improving lexical processing, and the Philadelphia Naming Test was readministered immediately following treatment. Using the Fazekas scale, WMH severity was independently rated on the whole brain and the right hemisphere only. Their relationship of WMH behavior was calculated by accounting for age, lesion volume, time poststroke, years of education, and sex. RESULTS: There were significant positive correlations between whole-brain and right-hemisphere ratings of WMH, but Wilcoxon signed-rank tests revealed that whole-brain ratings were consistently higher (whole brain M=3.337, right hemisphere M=2.899; P <0.001). Right hemisphere ratings were more strongly correlated with behavioral measures. There were significant negative correlations between all WMH ratings and behavior at baseline (Western Aphasia Battery Aphasia Quotient and Philadelphia Naming Test) except for whole brain periventricular ratings. However, only right-hemisphere periventricular WMH was significantly associated with therapy-related naming improvements ( r =−0.226, R 2 =0.05, P =0.044). CONCLUSIONS: Given the regional effect of the stroke lesion, whole-brain ratings can overestimate the WMH burden and thus reduce the accuracy in evaluating small vessel disease effects on stroke recovery and aphasia severity, particularly therapy-related neuroplasticity. This is an important detail that should be considered in mechanistic studies of small vessel disease, brain health, and stroke recovery.