Diffusion-Weighted Imaging Volume as the Best Predictor of the Diffusion–Perfusion Mismatch in Acute Stroke Patients within 8 Hours of Onset

BACKGROUND Diagnostic accuracies of standard NCCT, CTA, CTA-SI, FLAIR, and DWI to detect the diffusion‐perfusion mismatch (DPM) were compared. METHODS Stroke patients considered for endovascular therapy within 8 hours of onset were enrolled. DPM was defined as at least 160% mismatch between DWI and PWI volume. RESULTS DPM was seen in 35 (71%) of 49 patients. ASPECTS on NCCT, CTA-SI, and DWI was 9 (8-9), 8 (6-9), and 7 (5-9) in mismatch group, and 6 (4-9), 6 (2-7), 5 (2-6) in nonmismatch group, respectively (P = .027, .006, and .001). Ischemic volume on CTA-SI and DWI was 4.6 (.2-13.0) cm 3 and 21.5 (9.7-44.0) cm 3 in mismatch group, and 61.5 (6.6-101.1) cm 3 and 94.9 (45.7-139.8) cm 3 in nonmismatch group (P = .003 and <.001). Significant collateralization on CTA-SI and FLAIR was seen in 80% and 88% in mismatch group, and 42% and 58% in nonmismatch group (P = .026 and .039). Odds ratios (95% CI) of DWI volume of 70 cm 3 to predict the mismatch was 30.17 (2.06-442.41) after adjusting for ASPECTSs on NCCT, CTA-SI, and DWI, 44.90 (2.75-732.73) for ischemic volume on CTA-SI, and 42.80 (3.05-601.41) for significant collateralization on CTA-SI and FLAIR (P = .013, .008, and .005). CONCLUSIONS DWI volume was the best predictor of DPM.