A Retrospective Study of Outcome Predictive Models for Anterior Circulation Stroke Post-Endovascular Thrombectomy: Are CT-Derived Morphomics Useful?

introductionIn acute ischaemic stroke secondary to a large vessel occlusion, endovascular thrombectomy (EVT) is the standard of care. However, prediction of functional outcomes following EVT remains challenging. We sought to determine whether CT-derived morphomic data on body composition can improve outcome prediction following EVT.

methodsA retrospective analysis of anterior circulation stroke patients who underwent EVT between 2021 and 2023 at a single centre was conducted. Demographic, clinical, and morphomic data including neck muscle and fat area and density were collected. Multivariate logistic regression was used to develop pre-procedure and 24-h post-EVT models for predicting 90-day outcomes. A positive outcome was defined as a modified Rankin Scale (mRS) score of ≤2 at 90 days post-EVT. These models were subsequently tested on 2 separate validation cohorts.

resultsHigh pre-procedure ASPECTS scores, lower baseline NIHSS scores, absence of post-procedure haemorrhage, younger age, and early neurological improvement were the strongest predictors of a positive outcome. None of the morphomic variables were significantly correlated with outcome. The 24-h model was highly accurate in predicting post-EVT outcomes in both validation cohorts.

conclusionCT-derived morphomic variables did not significantly improve prediction of stroke outcomes post-EVT. Clinical and imaging variables (especially 24-h post-procedure data) showed high accuracy in predicting functional outcomes at 90 days.