Faculty Bibliography

Purpose: Ultrasound is increasingly utilized as the first-line diagnostic evaluation of superficial soft tissue masses. With growing health care costs, there is increasing pressure to develop cost-effective methods to triage patients with palpable masses. Deep convolutional neural networks (CNNs) have demonstrated the ability to classify images with good accuracy. We hypothesize that using a limited dataset, a CNN can be trained to classify benign versus malignant soft tissue masses seen on ultrasound.
Material(s) and Method(s): Ultrasound exams from 227 patients were selected with up to two pairs of gray scale and Doppler images extracted per patient. Pairs of gray scale and Doppler images were concatenated to create a single image for a total of 344 combined images. Images from 49 patients (96 images) were withheld for a pathology enriched test set (56 benign and 40 malignant). The remaining 248 images were used to train a CNN using an 80/20 training-validation split with five-fold crossvalidation. The model was trained on Keras using a pretrained VGG-16 architecture on a Nvidia GTX 1070 GPU. The withheld test set was used for a reader study which consists of two experienced musculoskeletal radiologists to assess the performance of the model.
Result(s): The CNN achieved an average accuracy of 0.87+/-0.07 on fivefold cross validation. The best performing model in the five folds was selected for comparison against two musculoskeletal radiologists on the pathology enriched test data set. The model achieved an accuracy 0.73 on the test data set and an AUC of 0.78 which was comparable to the performance of the two musculoskeletal radiologists (0.76 and 0.65 accuracy).
Conclusion(s): Using a relatively small data set, a CNN can be trained to differentiate between benign and malignant soft tissue masses seen on ultrasound with its performance approaching that of two experienced musculoskeletal radiologists

OBJECTIVE:The purpose of this study was to determine whether SWE can detect biomechanical changes in the supraspinatus muscle that occur with increasing supraspinatus tendon abnormality prior to morphologic gray-scale changes. MATERIALS AND METHODS/METHODS:An IRB approved, HIPAA compliant retrospective study of shoulder ultrasounds from 2013-2018 was performed. The cohort consisted of 88 patients (mean age 55 ± 15 years old) with 110 ultrasounds. Images were acquired in longitudinal orientation to the supraspinatus muscle with shear wave velocity (SWV) point quantification. The tendon and muscle were graded in order of increasing tendinosis/tear (1-4 scale) and increasing fatty infiltration (0-3 scale). Mixed model analysis of variance, analysis of covariance, and Spearman rank correlation were used for statistical analysis. RESULTS:There was no statistically significant age or sex dependence for supraspinatus muscle SWV (p = 0.314, 0.118, respectively). There was no significant correlation between muscle SWV and muscle or tendon grade (p = 0.317, 0.691, respectively). In patients with morphologically normal muscle on gray-scale ultrasound, there were significant differences in muscle SWV when comparing tendon grade 3 with grades 1, 2, and 4 (p = 0.018, 0.025, 0.014, respectively), even when adjusting for gender and age (p = 0.044, 0.028, 0.018, respectively). Pairwise comparison of tendon grades other than those mentioned did not achieve statistical significance (p > 0.05). CONCLUSION/CONCLUSIONS:SWE can detect biomechanical differences within the supraspinatus muscle that are not morphologically evident on gray-scale ultrasound. Specifically, supraspinatus tendon partial tears with moderate to severe tendinosis may correspond to biomechanically distinct muscle properties compared to both lower grades of tendon abnormality and full-thickness tears.

Most first-time anterior glenohumeral dislocations occur as the result of trauma. Many patients suffer recurrent episodes of anterior shoulder instability (ASI). The anatomy and biomechanics of ASI is addressed, as is the pathophysiology of capsulolabral injury. The roles of imaging modalities are described, including computed tomography (CT) and MR imaging with the additional value of arthrography and specialized imaging positions. Advances in 3D CT and MR imaging particularly with respect to the quantification of humeral and glenoid bone loss is discussed. The concepts of engaging and nonengaging lesions as well as on-track and off-track lesions are examined.

RATIONALE AND OBJECTIVES/OBJECTIVE:The medial plantar proper digital nerve, also called Joplin's nerve, arises from the medial plantar nerve, courses along the medial hallux metatarsophalangeal joint, and can be a source of neuropathic pain due to various etiologies, following acute injury including bunion surgery and repetitive microtrauma. We describe our clinical experience with diagnostic ultrasound assessment of Joplin's neuropathy and technique for ultrasound-guided therapeutic intervention including both injection and cryoablation over a 6-year period. MATERIALS AND METHODS/METHODS:Retrospective review of all diagnostic studies performed for Joplin's neuropathy and therapeutic Joplin's nerve ultrasound-guided injections and cryoablations between 2012 and 2018 was performed. Indications for therapeutic injection and cryoablation, were recorded. Studies were assessed for sonographic abnormalities related to the nerve and perineural soft tissues. Post-treatment outcomes including immediate pain scores, clinical follow-up, and periprocedural complications were documented. RESULTS:Twenty-four ultrasound-guided procedures were performed, including 15 perineural injections and nine cryoablations. With respect to sonographic abnormalities, nerve thickening (33%) and perineural hypoechoic scar tissue (27%) were the most common findings. The mean pain severity score prior to the therapeutic injection was 6.4/10 (range 4-10) and 0.25/10 (range 0-2) following the procedure; mean follow-up was 26.2 months (range 3-63 months). All of the cryoablation patients experienced sustained pain relief with a mean length follow-up of 3.75 months (range 0.2-10 months). CONCLUSION/CONCLUSIONS:Therapeutic injection of Joplin's nerve is a safe and easily performed procedure under ultrasound guidance, with high rates of immediate symptom improvement. For those experiencing a relapse or recurrent symptoms, cryoablation offers an effective secondary potential treatment option.

PURPOSE/OBJECTIVE:To determine if magnetic resonance angiography (MRA) and/or magnetic resonance imaging (MRI) could accurately determine the width of the labrum. METHODS:Consecutively enrolled patients between the ages of 18 and 65 indicated for hip arthroscopy for femoroacetabular impingement were included between December 2017 and June 2018. Inclusion criteria for preoperative MRIs included: MRI availability in picture archiving and communication system; performance on a 1.5T or 3T MRI or 3T MRA; and adequate quality and lack of labrum ossification. Intraoperative labral width measurements were taken at standardized locations using an established acetabular "clockface" paradigm. Measurement was performed using a calibrated probe. The labral width was defined as the distance from the labrum extended laterally from the acetabular rim. MRI measurements were taken by 2 blinded musculoskeletal fellowship-trained radiologists at the same positions. Measurements were made at the 11:30 o'clock position (indirect rectus) on coronal proton density (PD) sequence, at 3 o'clock position (psoas-U) on axial oblique PD sequence, and at 1:30 (a point halfway between the 2) on sagittal fat-suppressed PD. The surgeons were blinded to the radiologists' measurements and vice versa. Intraoperative and radiographic labral width measurements were compared using an intraclass correlation coefficients (ICC), absolute agreement, and 2-way random effects model. The 2 radiologists' measurements were compared for interrater reliability using the same ICC model. RESULTS:Fifty-one patients were included (30 females, 26 right hips). Average labrum width at the 3:00, 11:30, and 1:30 o'clock positions by arthroscopic measurement were 5.8 mm (range; standard deviation, 2-8; ±1.4), 6.3 mm (2-10; ±1.5) and 6.0 mm (2-9; ±1.5), and by MRI were 6.3 mm (2-10; ±1.5), 6.7 mm (3-10; ±1.4), and 6.1 mm (2-9; ±1.6), respectively. When including all MRI modalities, ICC agreement between intraoperative assessment, and radiologist assessment at the 3:00 o'clock, 11:30, and point halfway between was 0.82 (P < .001), 0.78 (P < .001), 0.84 (P < .001), respectively. Radiologist interrater ICC agreement at the same points was 0.88 (P < .001), 0.93 (P < .001), and 0.88 (P < .001). CONCLUSIONS:Strong agreement was found between radiologic and arthroscopic measurement of labrum width when using MRI, suggesting MRI is an accurate way to measure labral width. There was not a significant difference between different MRI modalities. Accurately measuring labral width preoperatively with MRI may aid in surgical decision making. LEVEL OF EVIDENCE/METHODS:Level II, diagnostic study.

Osteoarthritis (OA) is the most common joint disease in the United States. The prevalence of OA is rising due to an aging population and increasing rates of obesity. Magnetic resonance imaging (MRI) allows an incomparable noninvasive assessment of all joint structures. Irreversible and progressive degradation of the articular cartilage remains the hallmark feature of OA. To date, attempts at developing disease-modifying drugs or biomechanical interventions for treating OA have proven unsuccessful. MRI-based cartilage imaging techniques have continued to advance, however, and will likely play a central role in the development of these joint preservation methods of the future. In this narrative review, we describe clinical MR image acquisition and assessment of cartilage. We discuss the semiquantitative cartilage scoring methods used in research. Lastly, we review the quantitative MRI techniques that allow assessment of changes in the biochemical composition of cartilage, even before the morphological changes are evident.

PURPOSE/OBJECTIVE:To compare femoroacetabular motion in a series of consecutive symptomatic patients with hip pain throughout the range of motion of the hip using a real-time radial gradient echo (GRE) sequence in addition to the routine hip protocol sequences for magnetic resonance (MR) arthrographic assessment of patients with and without clinical femoroacetabular impingement (FAI) syndrome. In particular, we sought to assess whether the additional dynamic sequence could differentiate between patients with and without a positive physical exam maneuver for FAI syndrome. METHODS:Patients with hip pain referred for conventional hip MR arthrogram including those with and without a positive physical exam maneuver for FAI syndrome were imaged using routine hip MR arthrogram protocol and an additional real-time radial 2-dimensional GRE acquisition at 3 Tesla in an axial oblique plane with continuous scanning of a 9 mm thick slice through the center of the femoral head-neck axis. Patients who were unable to move through the range of motion were excluded (n = 3). Patients with acetabular dysplasia (defined by a lateral center-edge angle [CEA] of 20°) were also excluded, as were patients had Kellgren and Lawrence scores of > 0. The real-time cine sequence was acquired with the patient actively moving through neutral, flexion, flexion-abduction external-rotation, and flexion-adduction internal rotation (FADIR) positions aiming for 40° of abduction, then 25° of adduction at 80° to 90° flexion. Due to the placement of the coil over the hip, a true FADIR was precluded. Images were evaluated independently by 2 musculoskeletal radiologists measuring the joint space in the anterior, central, and posterior positions at each point during range of motion for femoroacetabular cortical space (FACS). Anterior FACS narrowing was calculated as the ratio of joint space in FADIR:neutral position, with lower ratios indicating greater narrowing. Static metrics including alpha angle, CEA, grade of cartilage loss according the Outerbridge classification, and patient demographics were also recorded. RESULTS:Twenty-two painful hips in 22 patients (11 males and 11 females) with mean age 36 years (range, 15-67) were included. Twelve patients had a positive physical exam maneuver for FAI syndrome. The time to perform the dynamic sequence was 3 to 6 minutes. Interobserver agreement was strong, with intraclass correlation 0.91 and concordance correlation 0.90. According to results from both readers, patients with impingement on clinical exam had significantly lower anterior FACS ratios compared with those without clinical impingement (reader 1: 0.39 ± 0.10 vs 0.69 ± 0.20, P = .001; reader 2: 0.36 ± 0.07 vs 0.70 ± 0.17, P < .001). Decreased anterior FACS ratio was found to be significantly correlated to increased alpha angle by both readers (reader 1: R = -0.63, P = .002; reader 2: R = -0.67, P = .001) but not significantly correlated to CEA (reader 1: R = 0.13, P = .561; reader 2: R = 0.20, P = .378) or cartilage loss (reader 1: R = 0.03, P = .885; reader 2: R = -0.06, P = .784). Both readers found patients with an anterior FACS ratio of 1/2 to have significantly higher mean alpha angle (reader 1: 62.88 vs 52.79, P = .038; reader 2: 63.50 vs 50.58, P = .006); however, there were no significant differences in cartilage loss (reader 1: P = .133; reader 2: P = .882) or CEA (reader 1: P = .340; reader 2: P = .307). CONCLUSIONS:A dynamic radial 2-dimensional-GRE sequence can be added to standard hip MR arthrogram protocols in <6 minutes, allowing assessment of dynamic femoroacetabular motion with strong interreader agreement. Patients with impingement on clinical exam had significantly lower anterior FACS ratios between FADIR and neutral positions, compared with those without clinical impingement. LEVEL OF EVIDENCE/METHODS:Level III, comparative diagnostic investigation.