Semi-Supervised Deep Learning for Multi-Tissue Segmentation from Multi-Contrast MRI

Segmentation of thigh tissues (muscle, fat, inter-muscular adipose tissue (IMAT), bone, and bone marrow) from magnetic resonance imaging (MRI) scans is useful for clinical and research investigations in various conditions such as aging, diabetes mellitus, obesity, metabolic syndrome, and their associated comorbidities. Towards a fully automated, robust, and precise quantification of thigh tissues, herein we designed a novel semi-supervised segmentation algorithm based on deep network architectures. Built upon Tiramisu segmentation engine, our proposed deep networks use variational and specially designed targeted dropouts for faster and robust convergence, and utilize multi-contrast MRI scans as input data. In our experiments, we have used 150 scans from 50 distinct subjects from the Baltimore Longitudinal Study of Aging (BLSA). The proposed system made use of both labeled and unlabeled data with high efficacy for training, and outperformed the current state-of-the-art methods with dice scores of 97.52%, 94.61%, 80.14%, 95.93%, and 96.83% for muscle, fat, IMAT, bone, and bone marrow tissues, respectively. Our results indicate that the proposed system can be useful for clinical research studies where volumetric and distributional tissue quantification is pivotal and labeling is a significant issue. To the best of our knowledge, the proposed system is the first attempt at multi-tissue segmentation using a single end-to-end semi-supervised deep learning framework for multi-contrast thigh MRI scans.

[1]  Brian B. Avants,et al.  N4ITK: Improved N3 Bias Correction , 2010, IEEE Transactions on Medical Imaging.

[2]  Vladimir Braverman,et al.  Multiparametric Deep Learning Tissue Signatures for Muscular Dystrophy: Preliminary Results , 2019, ArXiv.

[3]  Syed Muhammad Anwar,et al.  Segmentation of Glioma Tumors in Brain Using Deep Convolutional Neural Network , 2017, Neurocomputing.

[4]  Ariel D. Procaccia,et al.  Variational Dropout and the Local Reparameterization Trick , 2015, NIPS.

[5]  Josien P. W. Pluim,et al.  Not‐so‐supervised: A survey of semi‐supervised, multi‐instance, and transfer learning in medical image analysis , 2018, Medical Image Anal..

[6]  Quoc V. Le,et al.  Adding Gradient Noise Improves Learning for Very Deep Networks , 2015, ArXiv.

[7]  T. Manini,et al.  Longitudinal study of muscle strength, quality, and adipose tissue infiltration. , 2009, The American journal of clinical nutrition.

[8]  Syed Muhammad Anwar,et al.  Medical Image Analysis using Convolutional Neural Networks: A Review , 2017, Journal of Medical Systems.

[9]  Jayaram K. Udupa,et al.  New variants of a method of MRI scale standardization , 2000, IEEE Transactions on Medical Imaging.

[10]  Ulas Bagci,et al.  A Novel Extension to Fuzzy Connectivity for Body Composition Analysis: Applications in Thigh, Brain, and Whole Body Tissue Segmentation , 2018, IEEE Transactions on Biomedical Engineering.

[11]  Andreas K. Maier,et al.  Robust Spectral Denoising for Water-Fat Separation in Magnetic Resonance Imaging , 2015, MICCAI.

[12]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[13]  I. Illa,et al.  Muscle MRI in muscular dystrophies , 2015, Acta myologica : myopathies and cardiomyopathies : official journal of the Mediterranean Society of Myology.

[14]  Geoffrey E. Hinton,et al.  Learning Sparse Networks Using Targeted Dropout , 2019, ArXiv.

[15]  K. Hollingsworth,et al.  Improving highly accelerated fat fraction measurements for clinical trials in muscular dystrophy: origin and quantitative effect of R2* changes. , 2015, Radiology.

[16]  A. Falini,et al.  Automatic muscle and fat segmentation in the thigh from T1‐Weighted MRI , 2016, Journal of magnetic resonance imaging : JMRI.

[17]  Jörg Weule,et al.  Non-Linear Gaussian Filters Performing Edge Preserving Diffusion , 1995, DAGM-Symposium.

[18]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[19]  Yoshua Bengio,et al.  The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[20]  J. Lexell,et al.  Aging of human muscle: structure, function and adaptability , 1995, Scandinavian journal of medicine & science in sports.

[21]  L. Ferrucci The Baltimore Longitudinal Study of Aging (BLSA): a 50-year-long journey and plans for the future. , 2008, The journals of gerontology. Series A, Biological sciences and medical sciences.

[22]  Daniel P. Huttenlocher,et al.  Comparing Images Using the Hausdorff Distance , 1993, IEEE Trans. Pattern Anal. Mach. Intell..

[23]  Dino Samartzis,et al.  Reliability of quantifying the spatial distribution of fatty infiltration in lumbar paravertebral muscles using a new segmentation method for T1-weighted MRI , 2016, BMC Musculoskeletal Disorders.

[24]  M. McGee-Lawrence,et al.  Fatty Infiltration of Skeletal Muscle: Mechanisms and Comparisons with Bone Marrow Adiposity , 2016, Front. Endocrinol..

[25]  Michael Gadermayr,et al.  A comprehensive study on automated muscle segmentation for assessing fat infiltration in neuromuscular diseases. , 2018, Magnetic resonance imaging.

[26]  Kexin Li,et al.  Domain‐specific data augmentation for segmenting MR images of fatty infiltrated human thighs with neural networks , 2019, Journal of magnetic resonance imaging : JMRI.

[27]  Luigi Ferrucci,et al.  Image-Based Tissue Distribution Modeling for Skeletal Muscle Quality Characterization , 2016, IEEE Transactions on Biomedical Engineering.

[28]  Ulas Bagci,et al.  Lung and Pancreatic Tumor Characterization in the Deep Learning Era: Novel Supervised and Unsupervised Learning Approaches , 2018, IEEE Transactions on Medical Imaging.

[29]  Ronald M. Summers,et al.  Holistic Segmentation of Intermuscular Adipose Tissues on Thigh MRI , 2017, MICCAI.

[30]  T. Yousry,et al.  MRI biomarker assessment of neuromuscular disease progression: a prospective observational cohort study , 2016, The Lancet Neurology.

[31]  S. Kritchevsky,et al.  Association between regional adipose tissue distribution and both type 2 diabetes and impaired glucose tolerance in elderly men and women. , 2003, Diabetes care.