Faculty Bibliography

ABSTRACT/UNASSIGNED:Radiomics and machine learning-based methods offer exciting opportunities for improving diagnostic performance and efficiency in musculoskeletal radiology for various tasks, including acute injuries, chronic conditions, spinal abnormalities, and neoplasms. While early radiomics-based methods were often limited to a smaller number of higher-order image feature extractions, applying machine learning-based analytic models, multifactorial correlations, and classifiers now permits big data processing and testing thousands of features to identify relevant markers. A growing number of novel deep learning-based methods describe magnetic resonance imaging- and computed tomography-based algorithms for diagnosing anterior cruciate ligament tears, meniscus tears, articular cartilage defects, rotator cuff tears, fractures, metastatic skeletal disease, and soft tissue tumors. Initial radiomics and deep learning techniques have focused on binary detection tasks, such as determining the presence or absence of a single abnormality and differentiation of benign versus malignant. Newer-generation algorithms aim to include practically relevant multiclass characterization of detected abnormalities, such as typing and malignancy grading of neoplasms. So-called delta-radiomics assess tumor features before and after treatment, with temporal changes of radiomics features serving as surrogate markers for tumor responses to treatment. New approaches also predict treatment success rates, surgical resection completeness, and recurrence risk. Practice-relevant goals for the next generation of algorithms include diagnostic whole-organ and advanced classification capabilities. Important research objectives to fill current knowledge gaps include well-designed research studies to understand how diagnostic performances and suggested efficiency gains of isolated research settings translate into routine daily clinical practice. This article summarizes current radiomics- and machine learning-based magnetic resonance imaging and computed tomography approaches for musculoskeletal disease detection and offers a perspective on future goals and objectives.

Magnetic resonance imaging (MRI) is the keystone of modern musculoskeletal imaging; however, long pulse sequence acquisition times may restrict patient tolerability and access. Advances in MRI scanners, coil technology, and innovative pulse sequence acceleration methods enable 4-fold turbo spin echo pulse sequence acceleration in clinical practice; however, at this speed, conventional image reconstruction approaches the signal-to-noise limits of temporal, spatial, and contrast resolution. Novel deep learning image reconstruction methods can minimize signal-to-noise interdependencies to better advantage than conventional image reconstruction, leading to unparalleled gains in image speed and quality when combined with parallel imaging and simultaneous multislice acquisition. The enormous potential of deep learning-based image reconstruction promises to facilitate the 10-fold acceleration of the turbo spin echo pulse sequence, equating to a total acquisition time of 2-3 minutes for entire MRI examinations of joints without sacrificing spatial resolution or image quality. Current investigations aim for a better understanding of stability and failure modes of image reconstruction networks, validation of network reconstruction performance with external data sets, determination of diagnostic performances with independent reference standards, establishing generalizability to other centers, scanners, field strengths, coils, and anatomy, and building publicly available benchmark data sets to compare methods and foster innovation and collaboration between the clinical and image processing community. In this article, we review basic concepts of deep learning-based acquisition and image reconstruction techniques for accelerating and improving the quality of musculoskeletal MRI, commercially available and developing deep learning-based MRI solutions, superresolution, denoising, generative adversarial networks, and combined strategies for deep learning-driven ultra-fast superresolution musculoskeletal MRI. This article aims to equip radiologists and imaging scientists with the necessary practical knowledge and enthusiasm to meet this exciting new era of musculoskeletal MRI.

Decades of technical innovations have propelled musculoskeletal radiology through an astonishing evolution. New artificial intelligence and deep learning methods capitalize on many past innovations in magnetic resonance imaging (MRI) to reach unprecedented speed, image quality, and new contrasts. Similarly exciting developments in computed tomography (CT) include clinically applicable molecular specificity and substantially improved spatial resolution of musculoskeletal structures and diseases. This special issue of Investigative Radiology comprises a collection of expert summaries and reviews on the most impactful innovations and cutting-edge topics in musculoskeletal radiology, including radiomics and deep learning methods for musculoskeletal disease detection, high-resolution MR neurography, deep learning-driven ultra-fast musculoskeletal MRI, MRI-based synthetic CT, quantitative MRI, modern low-field MRI, 7.0 T MRI, dual-energy CT, cone beam CT, kinematic CT, and synthetic contrast generation in musculoskeletal MRI.

Image-guided minimally invasive radiofrequency ablation (RFA) of sensory nerves has emerged as a treatment option for pain and swelling associated with advanced symptomatic joint and spine degeneration to bridge the gap between optimal medical therapy and surgical treatments. RFA of articular sensory nerves and the basivertebral nerve use image-guided percutaneous approaches resulting in faster recovery time and minimal risks. The current published evidence indicates clinical effectiveness; however, further research must be performed comparing other conservative treatments with RFA to understand further its role in different clinical settings, such as osteonecrosis. This review article discusses and illustrates the applications of RFA for treating symptomatic joint and spine degeneration.

Accurately detecting and characterizing articular cartilage defects is critical in assessing patients with osteoarthritis. While radiography is the first-line imaging modality, magnetic resonance imaging (MRI) is the most accurate for the noninvasive assessment of articular cartilage. Multiple semiquantitative grading systems for cartilage lesions in MRI were developed. The Outerbridge and modified Noyes grading systems are commonly used in clinical practice and for research. Other useful grading systems were developed for research, many of which are joint-specific. Both two-dimensional (2D) and three-dimensional (3D) pulse sequences are used to assess cartilage morphology and biochemical composition. MRI techniques for morphological assessment of articular cartilage can be categorized into 2D and 3D FSE/TSE spin-echo and gradient-recalled echo sequences. T2 mapping is most commonly used to qualitatively assess articular cartilage microstructural composition and integrity, extracellular matrix components, and water content. Quantitative techniques may be able to label articular cartilage alterations before morphological defects are visible. Accurate detection and characterization of shallow low-grade partial and small articular cartilage defects are the most challenging for any technique, but where high spatial resolution 3D MRI techniques perform best. This review article provides a practical overview of commonly used 2D and 3D MRI techniques for articular cartilage assessments in osteoarthritis.

This article provides a focused overview of emerging technology in musculoskeletal MRI and CT. These technological advances have primarily focused on decreasing examination times, obtaining higher quality images, providing more convenient and economical imaging alternatives, and improving patient safety through lower radiation doses. New MRI acceleration methods using deep learning and novel reconstruction algorithms can reduce scanning times while maintaining high image quality. New synthetic techniques are now available that provide multiple tissue contrasts from a limited amount of MRI and CT data. Modern low-field-strength MRI scanners can provide a more convenient and economical imaging alternative in clinical practice, while clinical 7.0-T scanners have the potential to maximize image quality. Three-dimensional MRI curved planar reformation and cinematic rendering can provide improved methods for image representation. Photon-counting detector CT can provide lower radiation doses, higher spatial resolution, greater tissue contrast, and reduced noise in comparison with currently used energy-integrating detector CT scanners. Technological advances have also been made in challenging areas of musculoskeletal imaging, including MR neurography, imaging around metal, and dual-energy CT. While the preliminary results of these emerging technologies have been encouraging, whether they result in higher diagnostic performance requires further investigation.

OBJECTIVES/OBJECTIVE:To test whether a 4-fold accelerated 3D T2-weighted (T2) CAIPIRINHA SPACE TSE sequence with isotropic voxel size is equivalent to conventional 2DT2 TSE for the evaluation of intrinsic and perilesional soft tissue tumors (STT) characteristics. METHODS:For 108 patients with histologically-proven STTs, MRI, including 3DT2 (CAIPIRINHA SPACE TSE) and 2DT2 (TSE) sequences, was performed. Two radiologists evaluated each sequence for quality (diagnostic, non-diagnostic), tumor characteristics (heterogeneity, signal intensity, margin), and the presence or absence of cortical involvement, marrow edema, and perilesional edema (PLE); tumor size and PLE extent were measured. Signal-to-noise (SNR) and contrast-to-noise (CNR) ratios and acquisition times for 2DT2 in two planes and 3DT2 sequences were reported. Descriptive statistics and inter-method agreement were reported. RESULTS:Image quality was diagnostic for all sequences (100% [108/108]). No difference was observed between 3DT2 and 2DT2 tumor characteristics (p < 0.05). There was no difference in mean tumor size (3DT2: 2.9 ± 2.5 cm, 2DT2: 2.8 ± 2.6 cm, p = 0.4) or PLE extent (3DT2:0.5 ± 1.2 cm, 2DT2:0.5 ± 1.0 cm, p = 0.9) between the sequences. There was no difference in the SNR of tumors, marrow, and fat between the sequences, whereas the SNR of muscle was higher (p < 0.05) on 3DT2 than 2DT2. CNR measures on 3DT2 were similar to 2DT2 (p > 0.1). The average acquisition time was shorter for 3DT2 compared with 2DT2 (343 ± 127 s vs 475 ± 162 s, respectively). CONCLUSION/CONCLUSIONS:Isotropic 3DT2 MRI offers higher spatial resolution, faster acquisition times, and equivalent assessments of STT characteristics compared to conventional 2DT2 MRI in two planes. 3DT2 is interchangeable with a 2DT2 sequence in tumor protocols. KEY POINTS/CONCLUSIONS:• Isotropic 3DT2 CAIPIRINHA SPACE TSE offers higher spatial resolution than 2DT2 TSE and is equivalent to 2DT2 TSE for assessments of soft tissue tumor intrinsic and perilesional characteristics. • Multiplanar reformats of 3DT2 CAIPIRINHA SPACE TSE can substitute for 2DT2 TSE acquired in multiple planes, thereby reducing the acquisition time of MRI tumor protocols. • 3DT2 CAIPIRINHA SPACE TSE and 2DT2 TSE had similar CNR of tissues.

A 54-year-old woman presented with progressive right hip pain after hip arthroplasty 9 years earlier. The emerging role of metal artifact reduction MRI in the noninvasive diagnosis of infectious synovitis as the surrogate marker for periprosthetic hip joint infection and differentiation from other synovitis types is discussed.

OBJECTIVE:Deep learning has the potential to automatically triage orthopedic emergencies, such as joint dislocations. However, due to the rarity of these injuries, collecting large numbers of images to train algorithms may be infeasible for many centers. We evaluated if the Internet could be used as a source of images to train convolutional neural networks (CNNs) for joint dislocations that would generalize well to real-world clinical cases. METHODS:We collected datasets from online radiology repositories of 100 radiographs each (50 dislocated, 50 located) for four joints: native shoulder, elbow, hip, and total hip arthroplasty (THA). We trained a variety of CNN binary classifiers using both on-the-fly and static data augmentation to identify the various joint dislocations. The best-performing classifier for each joint was evaluated on an external test set of 100 corresponding radiographs (50 dislocations) from three hospitals. CNN performance was evaluated using area under the ROC curve (AUROC). To determine areas emphasized by the CNN for decision-making, class activation map (CAM) heatmaps were generated for test images. RESULTS:The best-performing CNNs for elbow, hip, shoulder, and THA dislocation achieved high AUROCs on both internal and external test sets (internal/external AUC): elbow (1.0/0.998), hip (0.993/0.880), shoulder (1.0/0.993), THA (1.0/0.950). Heatmaps demonstrated appropriate emphasis of joints for both located and dislocated joints. CONCLUSION/CONCLUSIONS:With modest numbers of images, radiographs from the Internet can be used to train clinically-generalizable CNNs for joint dislocations. Given the rarity of joint dislocations at many centers, online repositories may be a viable source for CNN-training data.

Many surgical procedures and operations are used to treat ankle and foot disorders. Radiography is the first-line imaging for postoperative surveillance and evaluation of pain and dysfunction. Computed tomography scans and MR imaging are used for further evaluation. MR imaging is the most accurate test for soft tissues assessments. MR imaging protocol adjustments include basic and advanced metal artifact reduction. We chose a surgical approach to select the common types of procedures and discuss the normal and abnormal postoperative MR imaging appearances, highlighting potential complications. This article reviews commonly used surgical techniques and their normal and abnormal MR imaging appearances.