Background: Endovascular therapy (ET) benefits patients with proximal anterior large vessel occlusions (LVO) provided ischemic core volume (CV) is not large. CT ASPECTS is a surrogate of CV and along with age and NIHSS, strongly predicts 3-month clinical outcomes and features in prognostic scores (PRE, HIAT-2). CV measured by CT perfusion (CTP) may overcome the low inter-rater reliability and subjectivity of ASPECTS.
Objective: To determine whether the prognostic accuracy of the PRE score for good outcome (3-month mRS 0-2) is improved if CT ASPECTS is replaced by CV (cerebral blood flow <30%) measured by Rapid processing of Perfusion and Diffusion (RAPID) software.
Methods: Consecutive patients (Oct 2010-Feb 2015) with carotid terminus, M1 and M2 MCA occlusions who had CTP with RAPID processing prior to ET were included. A multivariable model with age, NIHSS and CT ASPECTS as covariates was compared to a model containing age, NIHSS and CV measured by RAPID to predict good outcome. After weighting each variable by regression coefficients, a revised score with RAPID CV instead of CT ASPECTS (PRE-2) was derived. AUCs of the PRE and PRE-2 scores for good outcome were compared.
Results: 188 patients were included (mean age 62.5±15 years, median NIHSS 18 [IQR 13-21], median ASPECTS 8 [IQR 7-9]). TICI2B/3 reperfusion was achieved in 82.4% and good outcome in 59.3% of patients. Age, NIHSS and ASPECTS independently predicted good outcome (p<0.001). After replacing CT ASPECTS with CV, age, NIHSS and CV were found to independently predict good outcome (p<0.05). Predicted probabilities for good outcome in both models were strongly correlated. The PRE-2 score had similar predictive power for good outcome (AUC=0.79) as compared to the original PRE score (AUC=0.81). PRE-2≥130 predicted 0% (0/18) chance of good outcome.
Conclusions: CV measured by RAPID can substitute CT ASPECTS in the PRE score without any change in predictive accuracy. Validation of the PRE-2 score is warranted.