Deep learning for automatic segmentation of thigh and leg muscles

Abramo Agosti,Enzo Ricci,Luca Diamanti,Giovanni Savini,Shaun I Muzic,Matteo Paoletti,Giulia Colelli,Niels Bergsland,Enea Shaqiri,Mauro Monforte,Anna Pichiecchio,Xeni Deligianni,Francesco Santini,Giorgio Tasca,Francesca Solazzo,Stefano Bastianello

doi:10.1007/s10334-021-00967-4

Abstract

ObjectiveIn this study we address the automatic segmentation of selected muscles of the thigh and leg through a supervised deep learning approach.Material and methodsThe application of quantitative imaging in neuromuscular diseases requires the availability of regions of interest (ROI) drawn on muscles to extract quantitative parameters. Up to now, manual drawing of ROIs has been considered the gold standard in clinical studies, with no clear and universally accepted standardized procedure for segmentation. Several automatic methods, based mainly on machine learning and deep learning algorithms, have recently been proposed to discriminate between skeletal muscle, bone, subcutaneous and intermuscular adipose tissue. We develop a supervised deep learning approach based on a unified framework for ROI segmentation.ResultsThe proposed network generates segmentation maps with high accuracy, consisting in Dice Scores ranging from 0.89 to 0.95, with respect to “ground truth” manually segmented labelled images, also showing high average performance in both mild and severe cases of disease involvement (i.e. entity of fatty replacement).DiscussionThe presented results are promising and potentially translatable to different skeletal muscle groups and other MRI sequences with different contrast and resolution.

Highlights

Recent technical advances of muscle MRI imaging have led to an evolution from traditional qualitative evaluation into what is currently known as quantitative imaging, in which a large amount of diagnostically relevant information can be quantified and extracted from muscles of subjects affected by neuromuscular diseases [7, 23, 28]
As a further step towards the automatization of muscle regions of interest (ROI) drawing, we aimed to develop an automatic segmentation tool based on deep learning techniques to create single-muscle segmentation maps at thigh and leg level, starting from manually segmented multi-contrast quantitative muscle MRI scans of both healthy subjects and patients affected by two different neuromuscular diseases
With the aim to standardize and accelerate the process of ROI drawing we developed a deep neural network architecture, consisting of a classifier and two segmentation networks with residual units and contracting and expanding topologies inserted in a tree-like structure, which gave a unified framework for the automatic segmentation of both thigh and leg muscles

Summary

Introduction

Recent technical advances of muscle MRI imaging have led to an evolution from traditional qualitative evaluation into what is currently known as quantitative imaging (qMRI), in which a large amount of diagnostically relevant information (such as fat substitution and edema) can be quantified and extracted from muscles of subjects affected by neuromuscular diseases [7, 23, 28]. To extract quantitative data, drawing precise regions of interest (ROI) on selected muscles is crucial. The acquisition of multiple sequences on the same region potentially requires registering ROIs to different datasets; such a process adds the further task to manually correct the registered ROIs in the final space where data are eventually extracted for statistical analysis

Methods

Results

Discussion

Conclusion