Abstract
To provide objective visualization and pattern analysis of neck muscle boundaries to inform and monitor treatment of cervical dystonia. We recorded transverse cervical ultrasound (US) images and whole-body motion analysis of sixty-one standing participants (35 cervical dystonia, 26 age matched controls). We manually annotated 3,272 US images sampling posture and the functional range of pitch, yaw, and roll head movements. Using previously validated methods, we used 60-fold cross validation to train, validate and test a deep neural network (U-net) to classify pixels to 13 categories (five paired neck muscles, skin, ligamentum nuchae, vertebra). For all participants for their normal standing posture, we segmented US images and classified condition (Dystonia/Control), sex and age (higher/lower) from segment boundaries. We performed an explanatory, visualization analysis of dystonia muscle-boundaries. For all segments, agreement with manual labels was Dice Coefficient (64 ± 21%) and Hausdorff Distance (5.7 ± 4mm). For deep muscle layers, boundaries predicted central injection sites with average precision 94 ± 3%. Using leave-one-out cross-validation, a support-vector-machine classified condition, sex, and age from predicted muscle boundaries at accuracy 70.5%, 67.2%, 52.4% respectively, exceeding classification by manual labels. From muscle boundaries, Dystonia clustered optimally into three sub-groups. These sub-groups are visualized and explained by three eigen-patterns which correlate significantly with truncal and head posture. Using US, neck muscle shape alone discriminates dystonia from healthy controls. Using deep learning, US imaging allows online, automated visualization, and diagnostic analysis of cervical dystonia and segmentation of individual muscles for targeted injection.
Highlights
C ERVICAL Dystonia (CD), called spasmodic torticollis, is a painful condition in which the neck muscles contract involuntarily, causing the head to twist, turn, and pull into an abnormal posture
We report five main findings: (i) accuracy of extracted boundaries and of injection points within neck muscles, (ii) classification of condition, sex and age from muscle boundaries, (iii) the optimal clustering of dystonia into sub-groups, (iv) reduction to eigen-patterns of muscle shape associated with cervical dystonia and (v) the association of neck muscle eigenpatterns with whole body posture
This study reports the first application of deep learning to the segmentation, analysis and visualization of axial neck US images to participants with cervical dystonia
Summary
C ERVICAL Dystonia (CD), called spasmodic torticollis, is a painful condition in which the neck muscles contract involuntarily, causing the head to twist, turn, and pull into an abnormal posture. This neurological movement disorder affects an estimated 18,000 adults in the UK [1]. Clinical experience shows the main causes for treatment failure are suboptimal neck muscle selection or BoNT dosing, indicating the importance of appropriate targeting of overactive muscles [4], [5]. Monitoring the effectiveness of treatment is confounded by use of differing rating scales and assessment methods [6]. There is a clinical need to diagnose CD more promptly, to improve analysis and identification of dystonic muscles, to improve delivery of injection and dose to specific muscles, to provide objective recording of injection sites for retention within medical records and to track longitudinally the effect of injections on individual muscles [7]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
