Multi-task learning for the automatic grading of enlarged perivascular space burden using MRI

2026-09-29

AI Summary

This study addresses the inefficiency and poor inter-rater consistency of manual scoring for perivascular spaces (PVS) in brain MRI by proposing a multi-task convolutional neural network framework for automated PVS grading in the basal ganglia and centrum semiovale. The core innovation lies in introducing semi-automatically generated "silver standard" segmentation masks as auxiliary supervisory signals to jointly optimize segmentation and rating prediction tasks. Experimental results demonstrate that the proposed model achieves an average accuracy of 64.08%, significantly outperforming baseline methods. Furthermore, it provides both subject-level PVS localization capabilities and reliable probabilistic confidence estimates.

Abstract

Enlarged perivascular spaces (PVS) visible in brain magnetic resonance imaging (MRI) are increasingly thought to be linked to poor brain health. PVS are elongated structures of less than 3 mm in diameter and can be numerous. To reflect the incidence of PVS, radiologists visually score their burden following a clinical grading scale - a task that would benefit from automation to accelerate analyses and overcome the influence of inter-observer differences. We developed and evaluated methods for training machine learning models to score PVS incidence in the basal ganglia (BG) and centrum semiovale (CSO) leveraging the Potters/Wardlaw scale. The novelty in our work lies in the use of imperfect, semi-automatically generated "silver-standard" PVS segmentation masks during training, in addition to PVS radiological scores. We comparatively evaluated a conditional convolutional neural network (CNN) which accepts PVS masks as an extra input channel, a multi-task CNN which performs both PVS segmentation and scoring, and a logistic regression model which utilises features derived from PVS masks to predict PVS scores. Multi-task learning was the most effective method, achieving a mean average precision of 64.08% compared to 60.22% for the conditional CNN, 52.11% for a baseline CNN trained only to predict PVS scores, and 49.32% for the logistic regression model. The multi-task model showed an ability to localise individual PVS not shown by the other CNNs, and behaved in a probabilistically sensible way, predicting with lower confidence on inherently harder classes. Age, sex, hypertension status, white matter hyperintensity volume, and ischaemic stroke lesion status were shown to be associated with the multi-task model's PVS score predictions and the ground truth in a similar way.

Problem

• Enlarged perivascular spaces
• Automatic grading
• Magnetic resonance imaging
• Multi-task learning

Innovation

• Multi-task learning
• Enlarged perivascular spaces
• Silver-standard segmentation
• Convolutional neural network
• MRI grading