Artificial Intelligence Applications in Musculoskeletal Imaging.

This narrative review focuses on clinical applications of artificial intelligence (AI) in musculoskeletal imaging. A range of musculoskeletal disorders are discussed using a clinical-based approach, including trauma, bone age estimation, osteoarthritis, bone and soft-tissue tumors, and orthopedic implant-related pathology. Several AI algorithms have been applied to fracture detection and classification, which are potentially helpful tools for radiologists and clinicians. In bone age assessment, AI methods have been applied to assist radiologists by automatizing workflow, thus reducing workload and inter-observer variability. AI may potentially aid radiologists in identifying and grading abnormal findings of osteoarthritis as well as predicting the onset or progression of this disease. Either alone or combined with radiomics, AI algorithms may potentially improve diagnosis and outcome prediction of bone and soft-tissue tumors. Finally, information regarding appropriate positioning of orthopedic implants and related complications may be obtained using AI algorithms. In conclusion, rather than replacing radiologists, the use of AI should instead help them to optimize workflow, augment diagnostic performance, and keep up with ever-increasing workload.Relevance statement This narrative review provides an overview of AI applications in musculoskeletal imaging. As the number of AI technologies continues to increase, it will be crucial for radiologists to play a role in their selection and application as well as to fully understand their potential value in clinical practice.Key points• AI may potentially assist musculoskeletal radiologists in several interpretative tasks.• AI applications to trauma, age estimation, osteoarthritis, tumors, and orthopedic implants are discussed.• AI should help radiologists to optimize workflow and augment diagnostic performance.Graphical

Deep learning (DL) models have achieved remarkable performance in musculoskeletal (MSK) medical imaging research, yet their clinical integration remains hindered by their black-box nature and the absence of reliable confidence measures. Uncertainty quantification (UQ) seeks to bridge this gap by providing each DL prediction with a calibrated estimate of uncertainty, thereby fostering clinician trust and safer deployment. We conducted a targeted narrative review, performing expert-driven searches in PubMed, Scopus, and arXiv and mining references from relevant publications in MSK imaging utilizing UQ, and a thematic synthesis was used to derive a cohesive taxonomy of UQ methodologies. UQ approaches encompass multi-pass methods (e.g., test-time augmentation, Monte Carlo dropout, and model ensembling) that infer uncertainty from variability across repeated inferences; single-pass methods (e.g., conformal prediction, and evidential deep learning) that augment each individual prediction with uncertainty metrics; and other techniques that leverage auxiliary information, such as inter-rater variability, hidden-layer activations, or generative reconstruction errors, to estimate confidence. Applications in MSK imaging, include highlighting uncertain areas in cartilage segmentation and identifying uncertain predictions in joint implant design detections; downstream applications include enhanced clinical utility and more efficient data annotation pipelines. Embedding UQ into DL workflows is essential for translating high-performance models into clinical practice. Future research should prioritize robust out-of-distribution handling, computational efficiency, and standardized evaluation metrics to accelerate the adoption of trustworthy AI in MSK medicine. Not applicable.

To describe and evaluate a new fully automated musculoskeletal tissue segmentation method using deep convolutional neural network (CNN) and three-dimensional (3D) simplex deformable modeling to improve the accuracy and efficiency of cartilage and bone segmentation within the knee joint. A fully automated segmentation pipeline was built by combining a semantic segmentation CNN and 3D simplex deformable modeling. A CNN technique called SegNet was applied as the core of the segmentation method to perform high resolution pixel-wise multi-class tissue classification. The 3D simplex deformable modeling refined the output from SegNet to preserve the overall shape and maintain a desirable smooth surface for musculoskeletal structure. The fully automated segmentation method was tested using a publicly available knee image data set to compare with currently used state-of-the-art segmentation methods. The fully automated method was also evaluated on two different data sets, which include morphological and quantitative MR images with different tissue contrasts. The proposed fully automated segmentation method provided good segmentation performance with segmentation accuracy superior to most of state-of-the-art methods in the publicly available knee image data set. The method also demonstrated versatile segmentation performance on both morphological and quantitative musculoskeletal MR images with different tissue contrasts and spatial resolutions. The study demonstrates that the combined CNN and 3D deformable modeling approach is useful for performing rapid and accurate cartilage and bone segmentation within the knee joint. The CNN has promising potential applications in musculoskeletal imaging. Magn Reson Med 79:2379-2391, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Dual-energy CT stands out as a robust and innovative imaging modality, which has shown impressive advancements and increasing applications in musculoskeletal imaging. It allows to obtain detailed images with novel insights that were once the exclusive prerogative of magnetic resonance imaging. Attenuation data obtained by using different energy spectra enable to provide unique information about tissue characterization in addition to the well-established strengths of CT in the evaluation of bony structures. To understand clearly the potential of this imaging modality, radiologists must be aware of the technical complexity of this imaging tool, the different ways to acquire images and the several algorithms that can be applied in daily clinical practice and for research. Concerning musculoskeletal imaging, dual-energy CT has gained more and more space for evaluating crystal arthropathy, bone marrow edema, and soft tissue structures, including tendons and ligaments. This article aims to analyze and discuss the role of dual-energy CT in musculoskeletal imaging, exploring technical aspects, applications and clinical implications and possible perspectives of this technique.

In musculoskeletal imaging, CT is used in a wide range of indications, either alone or in a synergistic approach with MRI. While MRI is the preferred modality for the assessment of soft tissues and bone marrow, CT excels in the imaging of high-contrast structures, such as mineralized tissue. Additionally, the introduction of dual-energy CT in clinical practice two decades ago opened the door for spectral imaging applications. Recently, the advent of photon-counting detectors (PCDs) has further advanced the potential of CT, at least in theory. Compared to conventional energy-integrating detectors (EIDs), PCDs providesuperior spatial resolution, reduced noise, and intrinsic spectral imaging capabilities. This review briefly describes the technical advantages of PCDs. For each technical feature, the corresponding applications in musculoskeletal imaging will be discussed, including high-spatial resolution imaging for the assessment of bone and crystal deposits, low-dose applications such as whole-body CT, as well as spectral imaging applications including the characterization of crystal deposits and imaging of metal hardware. Finally, we will highlight the potential of PCD-CT in emerging applications, underscoring the need for further preclinical and clinical validation to unleash its full clinical potential.

Artificial intelligence (AI) is increasingly used in clinical practice for musculoskeletal imaging tasks, such as disease diagnosis and image reconstruction. AI applications in musculoskeletal imaging have focused primarily on radiography, CT, and MRI. Although musculoskeletal ultrasound stands to benefit from AI in similar ways, such applications have been relatively underdeveloped. In comparison with other modalities, ultrasound has unique advantages and disadvantages that must be considered in AI algorithm development and clinical translation. Challenges in developing AI for musculoskeletal ultrasound involve both clinical aspects of image acquisition and practical limitations in image processing and annotation. Solutions from other radiology subspecialties (e.g., crowdsourced annotations coordinated by professional societies), along with use cases (most commonly rotator cuff tendon tears and palpable soft-tissue masses), can be applied to musculoskeletal ultrasound to help develop AI. To facilitate creation of high-quality imaging datasets for AI model development, technologists and radiologists should focus on increasing uniformity in musculoskeletal ultrasound performance and increasing annotations of images for specific anatomic regions. This Expert Panel Narrative Review summarizes available evidence regarding AI's potential utility in musculoskeletal ultrasound and challenges facing its development. Recommendations for future AI advancement and clinical translation in musculoskeletal ultrasound are discussed.

Shoulder pain is a common musculoskeletal complaint requiring accurate imaging for diagnosis and management. Ultrasound is favored for its accessibility, dynamic imaging, and high-resolution soft tissue visualization. However, its operator dependency and variability in interpretation present challenges. Recent advancements in artificial intelligence (AI), particularly deep learning algorithms like convolutional neural networks, offer promising applications in musculoskeletal imaging, enhancing diagnostic accuracy and efficiency. This narrative review explores AI integration in shoulder ultrasound, emphasizing automated pathology detection, image segmentation, and outcome prediction. Deep learning models have demonstrated high accuracy in grading bicipital peritendinous effusion and discriminating rotator cuff tendon tears, while machine learning techniques have shown efficacy in predicting the success of ultrasound-guided percutaneous irrigation for rotator cuff calcification. AI-powered segmentation models have improved anatomical delineation; however, despite these advancements, challenges remain, including the need for large, well-annotated datasets, model generalizability across diverse populations, and clinical validation. Future research should optimize AI algorithms for real-time applications, integrate multimodal imaging, and enhance clinician-AI collaboration.

The potential of Artificial intelligence (AI) is increasingly recognized in musculoskeletal radiology, offering solutions to challenges posed by increasing imaging volumes and fellowship trained radiologist shortages. The integration of AI is not intended to replace radiologists but to augment their capabilities, improving workflow efficiency and diagnostic accuracy. This narrative review examines the current landscape of AI applications in musculoskeletal imaging, focusing on both general‐purpose multimodal models and specialized foundation models. AI has proven effective in musculoskeletal imaging, enhancing fracture detection, scoliosis assessment, and lower limb alignment analysis. In osteoarthritis, AI aids early detection by identifying subtle structural changes. AI‐accelerated MRI reconstruction reduces scan times by up to 90% while maintaining diagnostic quality, improving efficiency and accessibility. Emerging multimodal models further integrate imaging with clinical data, advancing precision medicine. Technical challenges persist, particularly in addressing motion artifacts and anatomical complexity. Ethical considerations, including data privacy, algorithmic bias, and model transparency, remain crucial for responsible implementation. While challenges remain in clinical validation and implementation, the combination of broad and narrow AI models shows promise in advancing precision medicine and democratizing quality care.Level of EvidenceLevel V.

Artificial Intelligence and Machine Learning Applications in Musculoskeletal Imaging

In recent years, technological advances have allowed manufacturers to implement dual-energy computed tomography (DECT) on clinical scanners. With its unique ability to differentiate basis materials by their atomic number, DECT has opened new perspectives in imaging. DECT has been used successfully in musculoskeletal imaging with applications ranging from detection, characterization, and quantification of crystal and iron deposits; to simulation of noncalcium (improving the visualization of bone marrow lesions) or noniodine images. Furthermore, the data acquired with DECT can be postprocessed to generate monoenergetic images of varying kiloelectron volts, providing new methods for image contrast optimization as well as metal artifact reduction. The first part of this article reviews the basic principles and technical aspects of DECT including radiation dose considerations. The second part focuses on applications of DECT to musculoskeletal imaging including gout and other crystal-induced arthropathies, virtual noncalcium images for the study of bone marrow lesions, the study of collagenous structures, applications in computed tomography arthrography, as well as the detection of hemosiderin and metal particles.

In recent years, technological advances have allowed manufacturers to implement dual-energy computed tomography (DECT) on clinical scanners. With its unique ability to differentiate basis materials by their atomic number, DECT has opened new perspectives in imaging. DECT has been successfully used in musculoskeletal imaging with applications ranging from detection, characterization, and quantification of crystal and iron deposits, to simulation of noncalcium (improving the visualization of bone marrow lesions) or noniodine images. Furthermore, the data acquired with DECT can be postprocessed to generate monoenergetic images of varying kiloelectron volts, providing new methods for image contrast optimization as well as metal artifact reduction. The first part of this article reviews the basic principles and technical aspects of DECT including radiation dose considerations. The second part focuses on applications of DECT to musculoskeletal imaging including gout and other crystal-induced arthropathies, virtual noncalcium images for the study of bone marrow lesions, the study of collagenous structures, applications in computed tomography arthrography, as well as the detection of hemosiderin and metal particles.

Deep learning-based segmentation has become a key tool for the precise and automated analysis of anatomical structures, such as bones, cartilage, and muscles, in musculoskeletal (MSK) imaging. This study examined the research trends by analyzing the number of related publications in PubMed since 2016 from both clinical and technical perspectives. Early studies primarily focused on the segmentation of major anatomical structures such as the spine and knee using large-scale datasets. However, recent studies have expanded to include the extremities and shoulders. In lesion segmentation, traditional topics such as body composition analysis, fractures, and tumors remain prominent, whereas deep learning-based detection and classification methods are increasingly integrated, leading to applications in newer areas. In addition, this study explored commonly used segmentation techniques and various applications of deep learning in MSK imaging. By systematically analyzing trends in deep learning-based segmentation research, we aim to provide insights into future directions for this rapidly evolving field.

The purpose of this study is to identify the specialty characteristics of providers referring musculoskeletal (MSK) extremity imaging examinations to radiologists, so as to better understand the drivers of MSK imaging utilization and potentially improve the appropriateness of such imaging examinations. Data on provider referral for MSK extremity imaging services were extracted from the 2014 Medicare Referring Provider Utilization for Procedures public use file, which aggregates data on diagnostic procedures according to referring provider identities and service codes. MSK extremity imaging services were identified using Neiman Institute Types of Service codes. The referring provider specialty was identified from cross-linked Medicare provider characteristics files. For 4,275,647 MSK extremity imaging examinations ordered, the most common specialties of the referring providers were orthopedic surgery (37.6% of ordered examinations), internal medicine (20.2%), family practice (14.8%), emergency medicine (7.9%), and rheumatology (5.7%). Orthopedic surgery was the referring specialty that most commonly ordered MSK extremity CT (33,465 ordered examinations; for all other specialties, < 2000 examinations), MRI (325,485 examinations; for all other specialities, < 20,000 examinations), and radiography (1,249,748 examinations; for all other specialities, < 850,000 examinations), whereas internal medicine was the referring specialty that most commonly ordered MSK extremity ultrasound examinations (8052 ordered examinations; for all other specialties, < 6000 examinations). Among the select specialties most relevant to MSK imaging, the most frequent referrers after orthopedic surgeons were rheumatologists, for radiography (236,057 ordered examinations) and ultrasound (2034 examinations), and podiatrists, for CT (1201 examinations) and MRI (19,159 examinations). The most commonly ordered individual MSK extremity imaging services were knee radiography, with 190,354 examinations ordered by orthopedic surgeons; hand radiography, with 66,167 examinations ordered by rheumatologists; foot radiography, with 137,042 examinations ordered by podiatrists; shoulder radiography, with 11,299 examinations ordered by sports medicine specialists; and hip radiography, with 9838 examinations ordered by physiatrists. Referral patterns for MSK imaging vary considerably by provider specialty. Referral pattern insights may guide targeted efforts by radiologists to ensure the appropriateness of such examinations.

Focus issue: Artificial intelligence in medical physics.

Two of the most exciting areas of development in MR imaging in the past several years have been rapid imaging and fat-suppression techniques. This article reviews the most widely available techniques for performing rapid imaging and fat suppression and summarizes current clinical applications in musculoskeletal imaging.