It is shown that transformer-based models with the proposed control-image regularisation achieve higher segmentation performance than U-Net-based approaches on clinically realistic data, while substantially reducing false positives in lesion-negative images.
Abstract
Ischaemic stroke, a leading cause of death and disability, relies on neuroimaging for characterising the anatomical pattern of injury. Diffusion-weighted MRI (DWI) provides the most anatomically specific signal in acute ischaemic stroke but poses substantial challenges for automated lesion segmentation due to susceptibility artefacts, lesion morphological heterogeneity, comorbid pathology, instrumental variability, and limited labelled data. Current U-Net-based models therefore underperform, a problem accentuated by evaluation metrics that neglect anatomical, subpopulation and acquisition-dependent variability.
We train 3D vision transformer–based segmentation models on a multi-site DWI dataset comprising 3563 annotated lesion-positive and 6900 lesion-negative volumes, using balanced cross-validation splits. We compare these models with U-Net baselines and an nnU-Net configuration under harmonised augmentation and training schemes, and introduce an evaluation framework that quantifies fidelity, anatomical precision, robustness to image corruption and equity across demographic and lesion-defined subtypes.
Here, we show that transformer-based models with our proposed control-image regularisation achieve higher segmentation performance than U-Net-based approaches on clinically realistic data, while substantially reducing false positives in lesion-negative images. They exhibit more stable performance across lesion sizes, anatomical territories, image quality and patient subgroups, indicating improved epistemic equity relative to convolutional architectures.
This work reconciles model expressivity with domain-specific challenges and redefines performance benchmarks to prioritise equity and generalisability–critical for personalised medicine and mechanistic research. These findings establish vision transformer architectures, combined with equity-aware validation, as a powerful approach for ischaemic stroke lesion segmentation on DWI.
When combined with human review and correction, MAESTRO offers a practical approach for generating standardized, high-quality lesion annotations, helping reduce a major practical barrier to large-scale stroke imaging studies.
M. H. Khan, O. Marin-Pardo, S. Chakraborty et al.· medRxiv· 0 citations
Objective: Diffusion-weighted MRI (DWI-MRI) is the gold standard for visualizing and quantifying acute ischaemic stroke (AIS). Although deep learning methods can accurately segment AIS lesions, the optimal image inputs and model architecture remain uncertain. We evaluated whether accurate AIS lesion segmentation can be...
Atle Bjørnerud, T. Schellhorn, T. Skattør et al.· 0 citations
Fast and accurate segmentation of Acute Ischemic Stroke (AIS) lesions is essential for stroke prognosis and treatment planning. Non-contrast CT (NCCT), the first-line imaging modality for diagnosing ischemic infarcts, exhibits subtle infarct contrast, making manual delineation slow and labor-intensive. Motivated by thi...
Cerebral microbleeds (CMBs) and cortical superficial siderosis (cSS) are imaging markers of cerebral small vessel disease, but their automated segmentation is limited by the scarcity of positive cases and voxel-level annotations. We propose a synthetic training framework for long-tail haemorrhagic lesion segmentation t...
Yuan Cao, Sumeet Dash, Antonia Zachariadis et al.· 0 citations
This work presents Stroke CT Analysis and Natural Language Reporting (SCAN-R), a unified end-to-end framework that integrates multiclass stroke detection, Transformerenhanced U-Net segmentation with task-specific pre-trained backbones, and Retrieval-Augmented Generation for evidence-based clinical report generation.
Le Minh Toan Truong, X. Nguyen, Dang Khanh Tran· International Conference on...· 0 citations
Practical recommendations for automated segmentation and future validation requirements are outlined: favor self-configuring frameworks like nnU-Net when compute resources permit, consider lightweight 2D models such as YOLOv8 for fast screening, and ensure rigorous cross-site validation before potential clinical use.
Patricia García-Berlanga, Juan Zapata, J. Martínez-Alajarín et al.· Journal of Intelligent Syste...· 0 citations