Clinically-aligned ischemic stroke segmentation and ASPECTS scoring on NCCT imaging using a slice-gated loss on foundation representations
This work addresses acute ischemic stroke management by improving segmentation and ASPECTS scoring on NCCT imaging, representing an incremental advance through the integration of existing methods with clinical priors.
The authors tackled the problem of rapid infarct assessment on non-contrast CT for acute ischemic stroke by proposing a clinically aligned framework that integrates foundation representations with structured clinical priors, achieving a Dice score of 0.6385 on AISD and improving mean Dice from 0.698 to 0.767 on a proprietary ASPECTS dataset.
Rapid infarct assessment on non-contrast CT (NCCT) is essential for acute ischemic stroke management. Most deep learning methods perform pixel-wise segmentation without modeling the structured anatomical reasoning underlying ASPECTS scoring, where basal ganglia (BG) and supraganglionic (SG) levels are clinically interpreted in a coupled manner. We propose a clinically aligned framework that combines a frozen DINOv3 backbone with a lightweight decoder and introduce a Territory-Aware Gated Loss (TAGL) to enforce BG-SG consistency during training. This anatomically informed supervision adds no inference-time complexity. Our method achieves a Dice score of 0.6385 on AISD, outperforming prior CNN and foundation-model baselines. On a proprietary ASPECTS dataset, TAGL improves mean Dice from 0.698 to 0.767. These results demonstrate that integrating foundation representations with structured clinical priors improves NCCT stroke segmentation and ASPECTS delineation.