Multi-task learning improves automatic scoring of brain perivascular spaces
Multi-task learning for the automatic grading of enlarged perivascular space burden using MRI
Computer Vision and Pattern RecognitionMachine Learning
Summary
Enlarged perivascular spaces (PVS) in brain scans may indicate poor brain health, but manually scoring them is slow and varies between doctors. The authors developed machine learning models that can automatically score the amount of PVS from MRI images. They found that a model trained to both locate PVS and score their amount performed best, showing better precision and sensible confidence levels. This could help speed up brain health assessments and make them more consistent.
What this means in practice
- •For radiology teams: Automatically score perivascular space burden in brain MRIs to accelerate clinical assessments and reduce variability between observers.
- •For medical imaging software developers: Integrate multi-task deep learning models that predict PVS scores and segment PVS to enhance brain health MRI analysis tools.$Commercial implications: Enables advanced diagnostic software for hospitals and clinics, offering automated, consistent PVS scoring to improve neurological evaluations.
Authors
Jesse Phitidis, William N. Whiteley, Joanna M. Wardlaw, Miguel O. Bernabeu, Yajun Cheng, Xiaodi Liu, Junfang Zhang, Una Clancy, Stephen Makin, Roberto Duarte Coello, Susana Muñoz Maniega, Mark E. Bastin, Simon R. Cox, Maria del C. Valdés Hernández
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.