Clinically-Informed Preprocessing Improves Stroke Segmentation in Low-Resource Settings

2025-08-21

AI Summary

To address the clinical challenge of accurate ischemic stroke lesion segmentation from CT alone—without MRI—in low-resource settings, this study proposes a clinical-knowledge-driven multimodal CT segmentation framework. Methodologically: (1) we adapt nnU-Net as the backbone and jointly process non-contrast CT (NCCT), CT angiography (CTA), and CT perfusion (CTP) sequences; (2) we leverage follow-up DWI-derived annotations as weak supervision to refine CT-based learning objectives; and (3) we integrate clinical preprocessing pipelines and automatically extract vascular structures from CTA to provide anatomical priors. In 10-fold cross-validation, our method achieves a 38% Dice score improvement over the baseline. Further integration of a dedicated CTA vessel segmentation module yields an additional 21% Dice gain in 5-fold validation. The framework significantly enhances acute-phase stroke lesion boundary delineation in MRI-absent scenarios, offering a deployable, rapid, and reliable diagnostic solution for resource-constrained regions.

Technology Category

• Computer Vision: Segmentation
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Stroke is among the top three causes of death worldwide, and accurate identification of ischemic stroke lesion boundaries from imaging is critical for diagnosis and treatment. The main imaging modalities used include magnetic resonance imaging (MRI), particularly diffusion weighted imaging (DWI), and computed tomography (CT)-based techniques such as non-contrast CT (NCCT), contrast-enhanced CT angiography (CTA), and CT perfusion (CTP). DWI is the gold standard for the identification of lesions but has limited applicability in low-resource settings due to prohibitive costs. CT-based imaging is currently the most practical imaging method in low-resource settings due to low costs and simplified logistics, but lacks the high specificity of MRI-based methods in monitoring ischemic insults. Supervised deep learning methods are the leading solution for automated ischemic stroke lesion segmentation and provide an opportunity to improve diagnostic quality in low-resource settings by incorporating insights from DWI when segmenting from CT. Here, we develop a series of models which use CT images taken upon arrival as inputs to predict follow-up lesion volumes annotated from DWI taken 2-9 days later. Furthermore, we implement clinically motivated preprocessing steps and show that the proposed pipeline results in a 38% improvement in Dice score over 10 folds compared to a nnU-Net model trained with the baseline preprocessing. Finally, we demonstrate that through additional preprocessing of CTA maps to extract vessel segmentations, we further improve our best model by 21% over 5 folds.

Problem

• Improving stroke lesion segmentation from CT images
• Enhancing diagnostic accuracy in low-resource healthcare settings
• Developing clinically-informed preprocessing for better model performance

Innovation

• Clinically-informed preprocessing enhances CT segmentation
• Combining CT inputs with DWI-annotated follow-up volumes
• Vessel segmentation from CTA maps boosts model performance