Predicting Thrombectomy Recanalization from CT Imaging Using Deep Learning Models

TL;DR

Predicting final MTB recanalization (mTICI) before intervention using pre-treatment CT/CTA imaging, the paper presents SCANet, a vision-transformer–based architecture with SAT and CAT modules that localize to stroke-relevant regions. The method processes slice-wise CT/CTA data through ResNet-based neighborhood branches to output binary recanalization predictions. On 177 patients, SCANet achieves ROC-AUC of 0.7732 ± 0.039, outperforming prior fully automatic and semi-automated approaches as well as radiomics and standard CNN baselines. This work demonstrates the potential of transformer-based imaging representations to inform treatment decisions for acute ischemic stroke, with future work targeting larger multi-site validation and linking immediate predictions to long-term outcomes.

Abstract

For acute ischemic stroke (AIS) patients with large vessel occlusions, clinicians must decide if the benefit of mechanical thrombectomy (MTB) outweighs the risks and potential complications following an invasive procedure. Pre-treatment computed tomography (CT) and angiography (CTA) are widely used to characterize occlusions in the brain vasculature. If a patient is deemed eligible, a modified treatment in cerebral ischemia (mTICI) score will be used to grade how well blood flow is reestablished throughout and following the MTB procedure. An estimation of the likelihood of successful recanalization can support treatment decision-making. In this study, we proposed a fully automated prediction of a patient's recanalization score using pre-treatment CT and CTA imaging. We designed a spatial cross attention network (SCANet) that utilizes vision transformers to localize to pertinent slices and brain regions. Our top model achieved an average cross-validated ROC-AUC of 77.33 $\pm$ 3.9\%. This is a promising result that supports future applications of deep learning on CT and CTA for the identification of eligible AIS patients for MTB.

Figures (1)

• Figure 1: Overview of the proposed deep learning architecture. The figure details the entire architecture, neighborhood branch modules, spatial attention transformer module (SAT), and cross attention transformer module (CAT).