Development and clinical validation of deep learning for auto-diagnosis of supraspinatus tears

Abstract

Background Accurately diagnosing supraspinatus tears based on magnetic resonance imaging (MRI) is challenging and time-combusting due to the experience level variability of the musculoskeletal radiologists and orthopedic surgeons. We developed a deep learning-based model for automatically diagnosing supraspinatus tears (STs) using shoulder MRI and validated its feasibility in clinical practice. Materials and methods A total of 701 shoulder MRI data (2804 images) were retrospectively collected for model training and internal test. An additional 69 shoulder MRIs (276 images) were collected from patients who underwent shoulder arthroplasty and constituted the surgery test set for clinical validation. Two advanced convolutional neural networks (CNN) based on Xception were trained and optimized to detect STs. The diagnostic performance of the CNN was evaluated according to its sensitivity, specificity, precision, accuracy, and F1 score. Subgroup analyses were performed to verify its robustness, and we also compared the CNN’s performance with that of 4 radiologists and 4 orthopedic surgeons on the surgery and internal test sets. Results Optimal diagnostic performance was achieved on the 2D model, from which F1-scores of 0.824 and 0.75, and areas under the ROC curves of 0.921 (95% confidence interval, 0.841–1.000) and 0.882 (0.817–0.947) were observed on the surgery and internal test sets. For the subgroup analysis, the 2D CNN model demonstrated a sensitivity of 0.33–1.000 and 0.625–1.000 for different degrees of tears on the surgery and internal test sets, and there was no significant performance difference between 1.5 and 3.0 T data. Compared with eight clinicians, the 2D CNN model exhibited better diagnostic performance than the junior clinicians and was equivalent to senior clinicians. Conclusions The proposed 2D CNN model realized the adequate and efficient automatic diagnoses of STs, which achieved a comparable performance of junior musculoskeletal radiologists and orthopedic surgeons. It might be conducive to assisting poor-experienced radiologists, especially in community scenarios lacking consulting experts.

Artificial Intelligence and Machine Learning in Rotator Cuff Tears

“…This algorithm quantified FI and displayed clear distinctions of Goutallier grades. Other authors developed a model that outperformed the Goutallier grading system, yielding a significant correlation between 3D FI measurements provided by the AI algorithm and the Goutallier grades assigned by clinicians 5 In addition, multiple DL models have shown high segmentation agreement among human observers, with additional high accuracy in detecting tears within the supraspinatus muscle 9,24,29 . Diagnosing RCT through DL-based supraspinatus MRI analysis has shown high objectivity and reliability.…”

Enabling Personalized Medicine in Orthopaedic Surgery Through Artificial Intelligence

“…DL has become one of the most exciting and popularized areas for clinical applications of AI in orthopaedics 28,68,74 as evidenced by the considerable increase in number of investigators applying this technology for common clinical problems 29,75–96 (Table III). Examples of clinically relevant applications of AI in imaging include the surveillance for prosthesis loosening/failure and brand recognition of different prosthetic implants 29,75–96 (Table III).…”

“…DL has become one of the most exciting and popularized areas for clinical applications of AI in orthopaedics 28,68,74 as evidenced by the considerable increase in number of investigators applying this technology for common clinical problems 29,75–96 (Table III). Examples of clinically relevant applications of AI in imaging include the surveillance for prosthesis loosening/failure and brand recognition of different prosthetic implants 29,75–96 (Table III). Although the use of computer vision may not be sufficient as a standalone metric for identification and diagnosis 90 , it does show great potential with investigators reporting near-perfect accuracy 84 and accuracy comparable with that of human raters 93 , further supporting the success in this area thus far.…”

“…Although these uses of AI in imaging have shown promising results thus far, further evidence is required to support their use clinically. Furthermore, all of the studies analyzing the use of computer vision in orthopaedics remain in the derivation 78,81 and validation 79,84,90,92 phases, with none analyzing the true impact of these applications in a clinical setting to the best of our knowledge. Therefore, the authors grade the utilization of computer vision in orthopaedics as a B (Table II).…”

Application of Artificial Intelligence in Shoulder Pathology

“…CNNs play a crucial role in enhancing the analysis and interpretation of shoulder MRI scans. A 2D CNN model was developed by Guo et al to automatically detect supraspinatus tears, trained on 701 shoulder MRIs and validated on 69 arthroplasty MRIs [ 38 ]. The model showed optimal performance, achieving high F1-scores and sensitivity on both surgery and internal test sets.…”