Background and Purpose Malignant middle cerebral artery infarction(MMI) affects 5–10% of acute ischaemic stroke patients. Mortality of ∼80% if treated conservatively leaves decompressive surgery the choice treatment, halving mortality if undertaken early. We aimed to determine early clinical and CT based imaging factors that predict MMI.
Methods We undertook a single centre retrospective study using an acute ischaemic stroke database that included CT head, CT angiography(CTA) and CT perfusion(CTP) scans <6 h after stroke onset and follow up CT and CTP at 24 h. Patients were classified as MMI blind to baseline data, using definition (NIHSS score: ≥14 or ≥19 for right and left hemispheres respectively and level of consciousness ≥1, hypodensity >50% of MCA territory and space-occupying oedema on 24 h CT. Univariate and multivariate binary logistic regression identified predictive factors.
Results Of 210 patients, 22 developed MMI. These had higher baseline NIHSS score, lower ASPECTS, larger ischaemic and infarct volumes, more large vessel occlusion and poor collateral circulation (all p=<0.0001). A multivariate model including; large vessel occlusion and total ischaemic volume >75 ml on admission CTP, correctly classified 91.5% of patients.
Conclusions CT imaging within 6 h of stroke onset can predict MMI. Multimodal CT allows early identification of those at risk of MMI.