Deep learning method for segmentation of rotator cuff muscles on MR images.

OBJECTIVE: To develop and validate a deep convolutional neural network (CNN) method capable of (1) selecting a specific shoulder sagittal MR image (Y-view) and (2) automatically segmenting rotator cuff (RC) muscles on a Y-view. We hypothesized a CNN approach can accurately perform both tasks compared with manual reference standards. MATERIAL AND METHODS: We created 2 models: model A for Y-view selection and model B for muscle segmentation. For model A, we manually selected shoulder sagittal T1 Y-views from 258 cases as ground truth to train a classification CNN (Keras/Tensorflow, Inception v3, 16 batch, 100 epochs, dropout 0.2, learning rate 0.001, RMSprop). A top-3 success rate evaluated model A on 100 internal and 50 external test cases. For model B, we manually segmented subscapularis, supraspinatus, and infraspinatus/teres minor on 1048 sagittal T1 Y-views. After histogram equalization and data augmentation, the model was trained from scratch (U-Net, 8 batch, 50 epochs, dropout 0.25, learning rate 0.0001, softmax). Dice (F1) score determined segmentation accuracy on 105 internal and 50 external test images. RESULTS: Model A showed top-3 accuracy > 98% to select an appropriate Y-view. Model B produced accurate RC muscle segmentations with mean Dice scores > 0.93. Individual muscle Dice scores on internal/external datasets were as follows: subscapularis 0.96/0.93, supraspinatus 0.97/0.96, and infraspinatus/teres minor 0.97/0.95. CONCLUSIONS: Our results show overall accurate Y-view selection and automated RC muscle segmentation using a combination of deep CNN algorithms.

A Retrospective Outcomes Study of 20 Sacroiliac Joint Fusion Patients.

STUDY DESIGN: Retrospective case series. PURPOSE: To report a novel approach to open posterior sacroiliac (SI) joint arthrodesis using a threaded titanium cage containing rhBMP--2. MATERIALS & METHODS: Twenty consecutive patients with a mean age of 57.7 years (range: 33--84) underwent posterior SI joint fusion. Two closely related novel posterior oblique approaches were employed. Enrolled subjects included 17 females and three males. The mean follow--up time for CT to assess fusion was 27 months (range: 17--45 months). Insurance included a mixture of public and private payers. One of the patients (patient 19) was on worker's compensation. During follow--up, patients were assessed radiologically for radiographic bony union and asked to rate their satisfaction with the procedure. The Oswestry Disability Index (ODI) was applied on a one-time basis upon follow-up. All patients were diagnosed with sacroiliac joint pain based primarily on a positive response to fluoroscopically guided injections into the sacroiliac joint. RESULTS: Out of 20 patients, 33 SI joints were considered symptomatic and operated, and 32/33 joints successfully fused radiographically (a 96.9% fusion success rate). The average procedure satisfaction rating (PSR) was 7.25 out of a maximum 10 (range 1--10). Seventeen patients responded to post-surgery -follow-up questions, and 13 patients (76%), indicated they would elect to have the surgery again as well as recommend it to others. Average estimated blood loss was less than 50 mL, and average length of stay was one day. CONCLUSIONS: Preliminary experiences with these novel posterior approaches to the SI joint described here seem to be safe and effective. The novel posterior approaches to the SI joint described here appear, preliminarily, to have many advantages over previously described procedures including markedly reduced surgical morbidity.