DWI/FLAIR mismatch assessment for ischemic stroke patients shows promising results in determining if patients are eligible for recombinant tissue-type plasminogen activator (r-tPA) treatment. However, the mismatch criteria suffer from two major issues: binary classification of a non-binary problem and the subjectiveness of the assessor. In this article, we present a simple automatic method for segmenting stroke-related parenchymal hyperintensities on FLAIR, allowing for an automatic and continuous DWI/FLAIR mismatch assessment. We further show that our method’s segmentations have comparable inter-rater agreement (DICE 0.820, SD 0.12) compared to that of two neuro-radiologists (DICE 0.856, SD 0.07), that our method appears robust to hyper-parameter choices (suggesting good generalizability), and lastly, that our methods continuous DWI/FLAIR mismatch assessment correlates to mismatch assessments made for a cohort of wake-up stroke patients at hospital submission. The proposed method shows promising results in automating the segmentation of parenchymal hyperintensity within ischemic stroke lesions and could help reduce inter-observer variability of DWI/FLAIR mismatch assessment performed in clinical environments as well as offer a continuous assessment instead of the current binary one.

Background: High inter-rater variability in DWI-FLAIR mismatch assessments for acute ischemic stroke (AIS) has spurred interest in assisting automated imaging measures. In this study, we explored whether DWI-FLAIR mismatch assessment and automated quantification of FLAIR hyperintensity, DWI-FLAIR volume ratio, and DWI volume could predict changes in NIHSS score following intravenous thrombolysis (IVT) in patients with AIS of unknown time since onset (TSO). We also exploratively compared radiological DWI-FLAIR mismatch assessments and imaging measures between patients who received IVT and those who did not. Methods: We conducted a retrospective, exploratory, single-center study analyzing brain MRIs from a consecutive cohort of patients with suspected AIS and unknown TSO admitted over two years. Patients with DWI hyperintensity lesions related to ischemia, identified automatically and subsequently verified radiologically, were included. We examined the correlation between automated imaging measures, retrospective DWI-FLAIR mismatch assessments, and changes in NIHSS score from baseline to 24 h post-treatment. Results: Of 333 patients included, 109 received IVT (mean age 68.9 ± 15.2 years) and 224 did not (mean age 70.8 ± 13.8 years). The median baseline NIHSS score was 5 in both groups, improving to 2 after IVT. The DWI volume significantly correlated with changes in NIHSS score (p = 0.002); FLAIR intensity demonstrated borderline significance (p = 0.056); and DWI-FLAIR volume ratio showed no statistically significant association in this cohort (p = 0.511). We did not find statistical evidence that the retrospective binary mismatch assessment was correlated with differences in outcome (p > 0.145). Conclusions: This study supports moving beyond binary DWI-FLAIR mismatch, suggesting that continuous, automated imaging parameters could potentially assist the radiologist in AIS management. As evidence remains preliminary, large-scale research is needed to establish clinical utility.