A deep learning model for diagnosing dystrophinopathies on thigh muscle MRI images

Abstract

Background: Dystrophinopathies are the most common type of inherited muscular diseases. Muscle biopsy and genetic tests are effective to diagnose the disease but cost much more than primary hospitals can reach. The more available muscle MRI is promising but its diagnostic results highly depends on doctors' experiences. This study intends to explore a way of deploying a deep learning model for muscle MRI images to diagnose dystrophinopathies.

Methods: This study collected 2536 T1WI images from 432 cases who had been diagnosed by genetic analysis and/or muscle biopsy, including 148 cases with dystrophinopathies and 284 cases with other diseases. The data was randomly divided into three sets: the data from 233 cases were used to train the CNN model, the data from 97 cases for the validation experiments, and the data from 102 cases for the test experiments. We also validated our models expertise at diagnosing by comparing the model's results on the 102 cases with those of three skilled radiologists.

Results: The proposed model achieved 91% (95% CI: 0.88, 0.93) accuracy on the test set, higher than the best accuracy of 84% in radiologists. It also performed better than the skilled radiologists in sensitivity : sensitivities of the models and the doctors were 0.89 (95% CI: 0.85 0.93) versus 0.79 (95% CI:0.73, 0.84; p = 0.190).

Conclusions: The deep model achieved excellent accuracy and sensitivity in identifying cases with dystrophinopathies. The comparable performance of the model and skilled radiologists demonstrates the potential application of the model in diagnosing dystrophinopathies through MRI images.

Keywords: Computer-Assisted Diagnosis; Deep Learning; Magnetic Resonance Imaging; Muscular Diseases.

Fig. 1

Workflow Diagram. a Flowchart of obtaining the images of three data sets. b Flowchart of constructing the CNN model using the three datasets

Fig. 2

Schematic of a convolutional neural networks with transfer learning. First, the ResNet50 model is trained on the ImageNet dataset of 1000 categories; Second, the convolutional layers are frozen and transferred into a new network; Third, the fully connected layers are retrained through the input of dystrophinopathies T1W1 images; Finally, the model outputs binary classification results

Fig. 3

Various metrics of experts and networks for dystrophinopathies diagnosis. a, Accuracy. b, F1 score. c, Sensitivity. d, Specificity. e, Positive LR. f, Negative LR. The legend of each subplot reports the detailed numerical results

Fig. 4

Receiver operating characteristic curve obtained using the convolution network. The receiver operator characteristic (ROC) area under the curve (AUC) is 0.98, and the orange triangle refers to the average performance of the experts. Circles and the triangle are all below the curve

Fig. 5

Saliency maps of the correctly diagnosed/the misdiagnosed dystrophinopathy/non- dystrophinopathy samples. Colors ranging from red to blue indicates the importance of image regions from high to low. a original images of samples correctly diagnosed as dystrophinopathy by the CNNs; c original images of correctly diagnosed as non-dystrophinopathy samples; e original images of incorrectly diagnosed as dystrophinopathy samples; g original images of incorrectly diagnosed as non-dystrophinopathy samples (b, d, f, h) are the corresponding saliency maps