A Unified Framework for Joint Detection of Lacunes and Enlarged Perivascular Spaces

2026-03-04

AI Summary

This study addresses the challenge of jointly detecting lacunes and enlarged perivascular spaces (PVS) in cerebral small vessel disease imaging, where their morphological similarity causes feature interference and extreme class imbalance. To resolve this, the authors propose a morphology-decoupling framework that employs a zero-initialized gated cross-task attention mechanism to leverage dense PVS contextual information for guiding sparse lacune detection. The approach integrates an exclusive constraint with a centerline Dice-based hybrid loss function and introduces tissue-semantic-aware dynamic false-positive suppression to enhance anatomical plausibility. Evaluated on the VALDO 2021 dataset, the method achieves a lacune detection precision of 71.1% and an F1-score of 62.6%, significantly outperforming the winning entries of the challenge. Its robustness at scale is further validated on the EPAD cohort (N=1762).

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Cerebral small vessel disease (CSVD) markers, specifically enlarged perivascular spaces (EPVS) and lacunae, present a unique challenge in medical image analysis due to their radiological mimicry. Standard segmentation networks struggle with feature interference and extreme class imbalance when handling these divergent targets simultaneously. To address these issues, we propose a morphology-decoupled framework where Zero-Initialized Gated Cross-Task Attention exploits dense EPVS context to guide sparse lacune detection. Furthermore, biological and topological consistency are enforced via a mixed-supervision strategy integrating Mutual Exclusion and Centerline Dice losses. Finally, we introduce an Anatomically-Informed Inference Calibration mechanism to dynamically suppress false positives based on tissue semantics. Extensive 5-folds cross-validation on the VALDO 2021 dataset (N=40) demonstrates state-of-the-art performance, notably surpassing task winners in lacunae detection precision (71.1%, p=0.01) and F1-score (62.6%, p=0.03). Furthermore, evaluation on the external EPAD cohort (N=1762) confirms the model's robustness for large-scale population studies. Code will be released upon acceptance.

Problem

• cerebral small vessel disease
• enlarged perivascular spaces
• lacunes
• medical image analysis
• class imbalance

Innovation

• morphology-decoupled framework
• gated cross-task attention
• mixed-supervision strategy
• anatomically-informed inference calibration
• cerebral small vessel disease