A rotator cuff tear (RCT) is an injury in adults that causes difficulty in moving, weakness, and pain. Only limited diagnostic tools such as magnetic resonance imaging (MRI) and ultrasound Imaging (UI) systems can be utilized for an RCT diagnosis. Although UI offers comparable performance at a lower cost to other diagnostic instruments such as MRI, speckle noise can occur the degradation of the image resolution. Conventional vision-based algorithms exhibit inferior performance for the segmentation of diseased regions in UI. In order to achieve a better segmentation for diseased regions in UI, deep-learning-based diagnostic algorithms have been developed. However, it has not yet reached an acceptable level of performance for application in orthopedic surgeries. In this study, we developed a novel end-to-end fully convolutional neural network, denoted as Segmentation Model Adopting a pRe-trained Classification Architecture (SMART-CA), with a novel integrated on positive loss function (IPLF) to accurately diagnose the locations of RCT during an orthopedic examination using UI. Using the pre-trained network, SMART-CA can extract remarkably distinct features that cannot be extracted with a normal encoder. Therefore, it can improve the accuracy of segmentation. In addition, unlike other conventional loss functions, which are not suited for the optimization of deep learning models with an imbalanced dataset such as the RCT dataset, IPLF can efficiently optimize the SMART-CA. Experimental results have shown that SMART-CA offers an improved precision, recall, and dice coefficient of 0.604% (+38.4%), 0.942% (+14.0%) and 0.736% (+38.6%) respectively. The RCT segmentation from a normal ultrasound image offers the improved precision, recall, and dice coefficient of 0.337% (+22.5%), 0.860% (+15.8%) and 0.484% (+28.5%), respectively, in the RCT segmentation from an ultrasound image with severe speckle noise. The experimental results demonstrated the IPLF outperforms other conventional loss functions, and the proposed SMART-CA optimized with the IPLF showed better performance than other state-of-the-art networks for the RCT segmentation with high robustness to speckle noise.

Background: Although simple posterolateral or posterior elbow dislocations are relatively common and usually stable after closed reduction, simple posteromedial dislocations are extremely rare and poorly characterized. We investigated the clinical characteristics, soft-tissue injury patterns, treatments, and outcomes of a series of posteromedial elbow dislocations without relevant osseous lesions. Methods: We retrospectively reviewed 20 cases of simple posteromedial elbow dislocations without relevant osseous lesions that were treated at 7 fellowship training hospitals during a 10-year period. Soft-tissue injury patterns in 15 cases were investigated with use of magnetic resonance imaging. Clinical outcomes were evaluated after an average of 56.1 months (range, 24 to 93 months) with use of the Mayo Elbow Performance Score (MEPS) and the Quick-DASH (an abbreviated version of the Disabilities of the Arm, Shoulder and Hand [DASH]) score. Complications were also evaluated. Results: On magnetic resonance imaging, significant tears of the lateral collateral ligament complex and common extensor group were observed in all cases. Seventeen cases (85%) required surgical treatment for acute instability. Fourteen cases underwent only lateral complex repair and 3 underwent repair of both the medial and lateral complexes. At the time of the latest follow-up, the mean MEPS and Quick-DASH scores were 85.8 ± 15.0 and 10.5 ± 16.3, respectively. Seventeen patients (85%) had a satisfactory clinical outcome. Complications following treatment included 4 patients with heterotopic ossification; 2 of these patients also experienced posttraumatic elbow stiffness, which was treated with arthrolysis at 8 and 18 months after the initial operation. Conclusions: Posteromedial elbow dislocations without relevant osseous lesions are associated with a more severe soft-tissue injury, especially to the lateral complex, resulting in a high rate of surgical treatment. With careful post-reduction evaluation, either operative or nonoperative treatment provided satisfactory clinical outcomes. Level of Evidence: Therapeutic Level IV . See Instructions for Authors for a complete description of levels of evidence.

Background: This study was performed to compare glenoid version and inclination measured using two-dimensional (2D) images from computed tomography (CT) scans or three-dimensional (3D) reconstructed bone models.Methods: Thirty patients who had undergone conventional CT scans were included. Two orthopedic surgeons measured glenoid version and inclination three times on 2D images from CT scans (2D measurement), and two other orthopedic surgeons performed the same measurements using 3D reconstructed bone models (3D measurement). The 3D-reconstructed bone models were acquired and measured with Mimics and 3-Matics (Materialise).Results: Mean glenoid version and inclination in 2D measurements were –1.705º and 9.08º, respectively, while those in 3D measurements were 2.635º and 7.23º. The intra-observer reliability in 2D measurements was 0.605 and 0.698, respectively, while that in 3D measurements was 0.883 and 0.892. The inter-observer reliability in 2D measurements was 0.456 and 0.374, respectively, while those in 3D measurements was 0.853 and 0.845.Conclusions: The difference between 2D and 3D measurements is not due to differences in image data but to the use of different tools. However, more consistent results were obtained in 3D measurement. Therefore, 3D measurement can be a good alternative for measuring glenoid version and inclination.

Rotator cuff tear (RCT) is a common injury that causes pain and disability in adults. The quantitative diagnosis of the RCT can be crucial in determining a treatment plan or monitoring treatment efficacy. Currently, only a few diagnosis tools, such as magnetic resonance imaging (MRI) and ultrasound imaging (US), are utilized for the diagnosis. Specifically, US exhibited comparable performance with MRI while offering a readily available diagnosis of RCTs at a lower cost. However, three-dimensional(3D) US and analysis of the regions are necessary to enable a better diagnosis of RCTs. Therefore, we developed a wide-field 3D US platform with a semi-automatic 3D image segmentation algorithm for 3D quantitative diagnosis of RCTs. The 3D US platform is built based on a conventional 2D US system and obtains 3D US images via linear scanning. With respect to 3D segmentation algorithm based on active contour model, frequency compounding and anisotropic diffusion methods were applied, and their effects on segmentation were discussed. The platform was used for clinical examination after evaluating the platform via the RCT-mimicking phantoms. As verified by the Dice coefficient(average DC: 0.663, volume DC: 0.723), which was approximately up to 50% higher than that obtained with conventional algorithms, the RCT regions segmented by the developed algorithm significantly matched the ground truth. The results indicated that the wide-field 3D US platform with the 3D segmentation algorithm can constitute a useful tool for improving the accuracy in the diagnosis of RCTs, and can eventually lead to better determination of treatment plans and surgical planning.