Introduction: Malignant edema in large MCA infarct (MMI) is known for high rate of mortality and morbidity. Hemicraniectomy was showed to decrease mortality but not improve functional outcome. Further studies are required for better patient selection for surgery. Among several clinical biomarkers, DWI lesion volume > 82 ml within 6 hours from symptoms onset is the most used predictor for MMI. There is also evidence suggesting brain edema is correspond with severity of tissue ischemia, which could be measured by semiquantitative cerebral blood volume (CBV) and cerebral blood flow (CBF) on MRP map. We hypothesized that DWI volume combining with MRP parameters could provide more predictive value.
Methods: Patients in MMI-MRI study (a prospective, multicenter, observational MRI/MRP study of MCA ischemic stroke within 6 hours of symptom onset) and STIR (Stroke Imaging Repository) were included if they met the inclusion criteria: age ≥ 18, MCA infarct with NIHSS ≥ 15, and MRI/MRP study within 12 hours from symptoms onset. Patients underwent thrombectomy, with poor quality images or without follow up images, with bilateral strokes, or developed hemorrhagic transformation were excluded. MMI was defined as NIHSS score >18 with suppressed level of consciousness, and at least 2/3 of the MCA territory infarct on follow-up MRI or CT.
Results: Of 89 patients included, 17 (19.1%) developed MMI. A malignant edema predicting score (MEPS) of maximum three points was developed: one point for DWI lesion volume > 82 ml, one point for more than 40% of DWI volume with CBV < 40% of contralateral normal side, and one point for more than 50% DWI volume with CBF < 30% of contralateral normal side. The threshold of a DWI lesion volume > 82 ml predicted malignant edema with sensitivity (0.65, 95% CI 0.39-0.85), specificity (0.86, 95% CI 0.75-0.93), positive predictive value (0.52, 95% CI 0.30-0.74), and negative predictive value (0.91, 95% CI 0.81-0.96). MEPS ≥ 2 predicted malignant edema with sensitivity (0.65, 95% CI 0.39-0.85), specificity (0.98, 95% CI 0.91-1.00), PPV (0.92, 95% CI 0.60-1.00), and NPV (0.92, 95% CI 0.83-0.97).
Conclusions: DWI volume is reliable in predicting MMI. The addition of information on tissue ischemia in a multivariate score improves the prediction of MMI over DWI alone.