Clinical Applications, Challenges, and Recommendations for Artificial Intelligence in Musculoskeletal and Soft-Tissue Ultrasound: AJR Expert Panel Narrative Review.

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.

Diagnostics of infrapatellar saphenous neuralgia-a reversible cause of chronic anteromedial pain following knee surgery.

OBJECTIVES: To evaluate the impact of diagnostic nerve block and ultrasound findings on therapeutic choices and predict the outcome after concomitant surgery in patients with suspected neuropathy of the infrapatellar branch of the saphenous nerve (IPBSN). METHODS: Fifty-five patients following knee surgery with suspicion of IPBSN neuralgia were retrospectively included. Ultrasound reports were assessed for neuroma and postsurgical scarring (yes/no). Responders and non-responders were assigned following anesthetic injection of the IPBSN. The type of procedure (neurectomy/interventional pain procedure/other than nerve-associated therapy) and pain score at initial follow-up were recorded and patients were assigned as positive (full pain relief) or negative (partial/no pain relief) to therapeutic nerve treatment. Factors associated with a relevant visual analog scale (VAS) reduction were assessed using uni- and multivariate logistic regression models and chi-square for quantitative and qualitative variables (p ≤ 0.05). RESULTS: Responders (37/55) more often had an entrapment or an evident neuroma of the IPBSN (97% vs. 6%). A positive Hoffmann-Tinel sign (p = 0.002) and the absence of knee joint instability (p = 0.029) predicted a positive response of the diagnostic nerve block (90%; 26/29). In the follow-up after therapeutic nerve treatment, all patients with full pain relief showed neuromas or entrapment of the IPBSN. Patients negatively responding to therapeutic nerve treatment more frequently showed an additional knee joint instability (25% vs. 4%). CONCLUSION: Selective denervation for neuropathic knee pain is beneficial in selected patients with significant VAS reduction after diagnostic nerve block. Non-responders following diagnostic nerve block but sonographic evidence of IPBSN pathologies need to be evaluated for other causes such as knee joint instability. KEY POINTS: • Sonographic diagnosis of neuroma or entrapment of the IPBSN is frequently seen in patients with anteromedial knee pain and leads to a good response to diagnostic nerve block following knee surgery. • The vast majority of patients with clinical signs of IPBSN neuropathy and response to a diagnostic nerve block sustained full pain relief following therapeutic nerve treatment. • Patients not responding to therapeutic IPBSN treatment have to be evaluated for other causes of anteromedial knee pain such as knee joint instability.

What happens to the lower lumbar spine after marathon running: a 3.0 T MRI study of 21 first-time marathoners.

OBJECTIVE: To better understand the impact of long-distance running on runners' lumbar spines by assessing changes before and after their first marathon run. MATERIALS AND METHODS: The lumbar spines of 28 asymptomatic adults (14 males, 14 females, mean age: 30 years old), who registered for their first marathon, the 2019 London Richmond Marathon, were examined 16 weeks before (time point 1) and 2 weeks after (time point 2) the marathon. Participants undertook a pre-race 16-week training programme. Magnetic resonance imaging (MRI) of high-resolution 3.0 Tesla was used at each time point. Senior musculoskeletal radiologists assessed the lower lumbar spine condition. RESULTS: Out of 28 participants, 21 completed both the training and the race and 7 neither completed the training nor started the marathon but not due to spine-related issues. At time point 1, disc degeneration was detected in 17/28 (61%), most predominantly at spinal segments L4-L5 and L5-S1. No back pain/other symptoms were reported. When compared to time point 2, there was no progression in the extent of disc degeneration, including intervertebral disc (IVD) height (p = 0.234), width (p = 0.359), and intervertebral distance (p = 0.641). There was a regression in 2 out of 8 (25%) participants who had pre-marathon sacroiliac joint bone marrow oedema, and a small increase in the size of a pre-marathon subchondral cyst in one participant, all asymptomatic. CONCLUSION: Running 500 miles over 4 months plus a marathon for the first time had no adverse effects on the lumbar spine, even when early degenerative changes were present. Additionally, there was evidence of regression of sacroiliac joint abnormalities.

AI-based detection and classification of distal radius fractures using low-effort data labeling: evaluation of applicability and effect of training set size.

OBJECTIVES: To evaluate the performance of a deep convolutional neural network (DCNN) in detecting and classifying distal radius fractures, metal, and cast on radiographs using labels based on radiology reports. The secondary aim was to evaluate the effect of the training set size on the algorithm's performance. METHODS: A total of 15,775 frontal and lateral radiographs, corresponding radiology reports, and a ResNet18 DCNN were used. Fracture detection and classification models were developed per view and merged. Incrementally sized subsets served to evaluate effects of the training set size. Two musculoskeletal radiologists set the standard of reference on radiographs (test set A). A subset (B) was rated by three radiology residents. For a per-study-based comparison with the radiology residents, the results of the best models were merged. Statistics used were ROC and AUC, Youden's J statistic (J), and Spearman's correlation coefficient (ρ). RESULTS: The models' AUC/J on (A) for metal and cast were 0.99/0.98 and 1.0/1.0. The models' and residents' AUC/J on (B) were similar on fracture (0.98/0.91; 0.98/0.92) and multiple fragments (0.85/0.58; 0.91/0.70). Training set size and AUC correlated on metal (ρ = 0.740), cast (ρ = 0.722), fracture (frontal ρ = 0.947, lateral ρ = 0.946), multiple fragments (frontal ρ = 0.856), and fragment displacement (frontal ρ = 0.595). CONCLUSIONS: The models trained on a DCNN with report-based labels to detect distal radius fractures on radiographs are suitable to aid as a secondary reading tool; models for fracture classification are not ready for clinical use. Bigger training sets lead to better models in all categories except joint affection. KEY POINTS: • Detection of metal and cast on radiographs is excellent using AI and labels extracted from radiology reports. • Automatic detection of distal radius fractures on radiographs is feasible and the performance approximates radiology residents. • Automatic classification of the type of distal radius fracture varies in accuracy and is inferior for joint involvement and fragment displacement.

In vivo 3D tomography of the lumbar spine using a twin robotic X-ray system: quantitative and qualitative evaluation of the lumbar neural foramina in supine and upright position.

OBJECTIVES: Supine lumbar spine examinations underestimate body weight effects on neuroforaminal size. Therefore, our purpose was to evaluate size changes of the lumbar neuroforamina using supine and upright 3D tomography and to initially assess image quality compared with computed tomography (CT). METHODS: The lumbar spines were prospectively scanned in 48 patients in upright (3D tomographic twin robotic X-ray) and supine (30 with 3D tomography, 18 with CT) position. Cross-sectional area (CSA), cranio-caudal (CC), and ventro-dorsal (VD) diameters of foramina were measured by two readers and additionally graded in relation to the intervertebral disc height. Visibility of bone/soft tissue structures and image quality were assessed independently on a 5-point Likert scale for the 18 patients scanned with both modalities. Descriptive statistics, Wilcoxon's signed-rank test (p < 0.05), and interreader reliability were calculated. RESULTS: Neuroforaminal size significantly decreased at all levels for both readers from the supine (normal intervertebral disc height; CSA 1.25 ± 0.32 cm2; CC 1.84 ± 0.24 cm2; VD 0.88 ± 0.16 cm2) to upright position (CSA 1.12 ± 0.34 cm2; CC 1.78 ± 0.24 cm2; VD 0.83 ± 0.16 cm2; each p < 0.001). Decrease in intervertebral disc height correlated with decrease in foraminal size (supine: CSA 0.88 ± 0.34 cm2; CC 1.39 ± 0.33 cm2; VD 0.87 ± 0.26 cm2; upright: CSA 0.83 ± 0.37 cm2, p = 0.010; CC 1.32 ± 0.33 cm2, p = 0.015; VD 0.80 ± 0.21 cm2, p = 0.021). Interreader reliability for area was fair to excellent (0.51-0.89) with a wide range for cranio-caudal (0.32-0.74) and ventro-dorsal (0.03-0.70) distances. Image quality was superior for CT compared with that for 3D tomography (p < 0.001; κ, CT = 0.66-0.92/3D tomography = 0.51-1.00). CONCLUSIONS: The size of the lumbar foramina is smaller in the upright weight-bearing position compared with that in the supine position. Image quality, especially nerve root delineation, is inferior using 3D tomography compared to CT. KEY POINTS: • Weight-bearing examination demonstrates a decrease of the neuroforaminal size. • Patients with higher decrease in intervertebral disc showed a narrower foraminal size. • Image quality is superior with CT compared to 3D tomographic twin robotic X-ray at the lumbar spine.

Ulnar collateral ligament injuries of the first metacarpophalangeal joint: prevalence of associated injuries on radiographs and MRI.

PURPOSE: To evaluate the prevalence of associated findings at the first metacarpophalangeal joint on radiographs and MRI following acute ulnar collateral ligament (UCL) injuries. MATERIALS AND METHODS: This retrospective study included 25 patients with an injury of the UCL at MRI. Presence of associated injuries to the volar ligaments (checkrein and phalangoglenoid ligaments and volar plate) was assessed on radiographs and MRI independently by two musculoskeletal radiologists. Wilcoxon signed-rank test was used to compare frequencies of injuries between both modalities (p < 0.05). Interreader variability was calculated. RESULTS: Complete tears of the UCL (48%/60%, reader 1/2) were more common than partial tears (24%/16%) on MRI. Dislocation of the UCL ≥ 3 mm was detected in 40%/56% on MRI. UCL avulsion fractures were more frequently seen on MRI (28%) compared with radiographs (12%) for reader 1. Associated avulsion injuries of the phalangoglenoid ligament were evident in 12%/8% on radiographs and in 80%/76% on MRI. Almost all patients (100%/79%) with a dislocated UCL tear showed a concomitant volar ligament injury; and even two-thirds (66%/72%) of the non-displaced UCL tears had an injury to the volar ligaments. Interreader agreement was moderate to excellent (κ = 0.60-1.0). CONCLUSION: UCL tears are often associated with volar ligament injuries, even in lesser degrees of an UCL injury.

Typical MRI-pattern suggests peak maturation of the ACI graft 2 years after third-generation ACI: a systematic review.

PURPOSE: The purpose of the present article was (1) to systematically review the current literature and (2) to collect data regarding the postoperative magnetic resonance imaging (MRI) appearance of third-generation autologous chondrocyte implantation (ACI) grafts and (3) to provide an overview of imaging findings at various postoperative time points. METHODS: A systematic review of the literature in Medline (Pubmed) and Embase was performed using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Articles which reported the post-operative MRI morphological outcomes following the use of third-generation ACI for treatment of knee cartilage lesions were included. All MRI results were allocated to six different time intervals: ≤ 3 months, > 3-6 months, > 6 months-1 year, > 1 year-2 years, > 2-5 years and > 5 years after surgery. RESULTS: A total of 22 studies were included and the study populations ranged from 13 to 180 patients adding up to a total of 951 patients. Parameters such as defect fill, border integration, surface contour, graft morphology and integrity of the subchondral lamina all improve gradually with a peak two years following surgery suggesting complete graft maturation at this time point. After this peak, a statistically insignificant decline is noted for most of the parameters. Signal intensity was found to gradually shift from hyperintense to isointense in the first 36 months and to hypointense later on. Contrarily, subchondral bone edema is not only a postoperative feature of the procedure but also can reappear or persist up to ten years after surgery. As graft failures can appear after two years, consequently, the MRI composite score is also affected. CONCLUSION: Recurring patterns in postoperative MRI appearance were observed in certain parameters including defect filling, graft signal intensity and structure, border integration of the graft while parameters like subchondral bone tend to be unpredictable. Given the heterogenous findings in terms of clinical correlation, and relating that aspect to the patterns found in this review, an MRI is justified at three months, one year, two years and five years after surgery, unless the clinical symptomatology and individual patient needs dictate otherwise. LEVEL OF EVIDENCE: IV.

MR variability of collagen meniscal implant remodelling in patients with good clinical outcome.

PURPOSE: Collagen meniscal implants (CMI) have variable imaging findings on MRI and it remains ambiguous, if alterations may be present in asymptomatic patients. The aim was, to evaluate MR morphological characteristics and volume of CMI in the early, mid- and long-term follow-up. METHODS: Fifty-seven patients (mean age, 43.6±11 years; 41 male/16 female) with good clinical outcome (subjective visual analogue scale for satisfaction < 2 and a Lysholm-score ≥ 84) after CMI were assessed retrospectively using MRI 1, 2 and 3-8 years postoperatively. CMI morphology, signal intensity (SI), homogeneity and size were assessed and a meniscal score was calculated. Chondral defects and bone marrow edema pattern were reported and summarized in a chondral score. Meniscal extrusion and volume were evaluated. Intra- and inter-reader reliabilities were calculated. Spearman and partial correlations were used (p < 0.05). RESULTS: One year postoperatively, the CMI varied in size [10% (4/40) normal, 30% (12/40) small, 60% (24/40) hypertrophic] and was hyperintense and inhomogeneous in all patients. At long-term follow-up, the size of CMI decreased [6% (1/17) resorbed, 18% (3/17) normal, 41% (7/17) small, 35% (6/17) hypertrophic]. The SI of the CMI changed to normal in only 13% (2/16). The majority was still hyperintense [87% (14/16)]. Less meniscal extrusion was present in the follow-up [71% (11/16) versus initially 93% (37/40)]. The meniscal score improved significantly (ρ = 0.28). Over time, full-thickness femoral chondral defects were increasingly present [65% (10/17) versus initially 33% (13/40)] and the chondral score worsened significantly (p = 0.017). CONCLUSION: Abnormal and inhomogeneous SI and irregular margins of the early postoperative CMI on MRI are findings seen in patients with good clinical results and should not necessarily be overcalled when reporting MRI. These features tend to decrease over time. LEVEL OF EVIDENCE: IV.

Avoiding pitfalls in ankle fracture-dislocations: A case report of a dislocated tibialis posterior tendon causing persistent ankle subluxation.

Tibialis posterior tendon dislocation, a rare complication in ankle fracture-dislocations, can impede anatomical reduction of the ankle mortise. We report on a 59-year-old healthy male with an open fracture-dislocation of the right ankle. Despite multiple attempts under direct vision of the anterior syndesmosis, anatomical reduction of the ankle mortise was not possible. Soft tissue windows in a computed tomography (CT) scan revealed the dislocated tibialis posterior tendon to be the impeding structure. At the level of the fibula fracture the tendon passed through the interosseous membrane anterior to the distal tibia and was then incarcerated in the medial talocrural joint before returning to the flexor retinaculum and its insertion on the navicular bone. Understanding the trauma mechanism and the course of the dislocated tendon as well as correct interpretation of CT and magnetic resonance images of the ankle enable surgeons to early diagnose and correctly treat this condition.

Can we use MRI to detect clinically silent recurrent soft-tissue sarcoma?

PURPOSE: The impact of MRI on early detection of local recurrence (LR) in high-grade soft-tissue sarcomas (STS) is unsubstantiated. To identify the contribution of MRI criteria including dynamic contrast-enhanced (DCE) MRI and knowledge of surgical margins that can be used in detecting recurrence prior to obvious proven presence of LR in soft-tissue sarcomas. The secondary aim was to determine causes for misdiagnosing LR. METHODS: MRI of 23 patients (12 men; mean age, 59.7 years ± 16.5 years) with LR of STS and that of 22 age- and histology-matched controls with STS but without LR were retrospectively analyzed by two musculoskeletal radiologists. Preoperative MRI characteristics (conventional and DCE) were compared to those of MRIs made after treatment, but before LR was proven. Likelihood of recurrence was rated on a 5-point Likert scale for morphological and dynamic assessment separately, before and after adding knowledge of surgical margins. Descriptive statistics and receiver operating characteristic analysis were performed. RESULTS: Differentiation of LR from post-therapeutic changes was the highest combining result of conventional MRI, DCE-MRI, and knowledge of surgical margins (area under the curve (AUC) 0.779), followed by DCE-MRI (AUC 0.706) and conventional MRI (AUC 0.648). Suboptimal MRI technique and overcalling post-therapeutic changes in microscopic positive margins were the main reasons for false negative and false positive results, respectively. CONCLUSION: MRI including DCE improves the detection of recurrent, clinically silent soft-tissue sarcoma when combined with knowledge of achieved surgical margins. LR may be missed on inadequate MRI protocols. KEY POINTS: • Dynamic contrast-enhanced MRI is useful in the differentiation of recurrent soft-tissue sarcoma and post-therapeutic fibrosis. • Knowledge of surgical margins substantially increases the value of MRI in detecting recurrent soft-tissue sarcoma. • MR with all three image orientations, covering the entire part of the extremity in at least one sequence and comparison to initial tumor characteristics and location, is beneficial.

Dynamic MR imaging of the skeletal muscle in young and senior volunteers during synchronized minimal neuromuscular electrical stimulation.

OBJECTIVE: Neuromuscular electrical stimulation (NMES)-induced isometric contraction is feasible during MRI and can be combined with acquisition of volumetric dynamic MR data, in a synchronous and controlled way. Since NMES is a potent resource for rehabilitation, MRI synchronized with NMES presents a valuable validation tool. Our aim was to show how minimal NMES-induced muscle contraction characterization, as evaluated through phase-contrast MRI, differs between senior and young volunteers. MATERIALS AND METHODS: Simultaneous NMES of the quadriceps muscle and phase-contrast imaging were applied at 3 T to 11 senior (75 ± 3 years) and 12 young volunteers (29 ± 7 years). A current sufficient to induce muscle twitch without knee extension was applied to both groups. RESULTS: Strain vectors were extracted from the velocity fields and strain datasets were compared with non-parametric tests and descriptive statistics. Strain values were noticeably different between both groups at both current intensities and significant differences were observed for similar current level. DISCUSSION: In conclusion, NMES-synchronized MRI could be successfully applied in senior volunteers with strain results clearly different from the younger volunteers. Also, differences within the senior group were detected both in the magnitude of strain and in the position of maximum strain pixels.

Comparison of 3D X-ray tomography with computed tomography in patients with distal extremity fractures.

OBJECTIVE: To compare fracture detection, image quality, and radiation dose in patients with distal extremity fractures using 3D tomography and computed tomography (CT). MATERIALS AND METHODS: IRB approval was obtained including informed consent for this prospective study from June to December 2016. Patients diagnosed with an acute fracture at CT were consecutively scanned on the same day using 3D tomography. Anatomical location (effected bone and location within the bone) and morphological characteristics of fractures (avulsion, articular involvement, mono- vs. multifragmented, displacement), visibility of bone/soft tissue structures, and image quality were assessed independently by two blinded readers on a 5-point Likert scale. Dose-length-product (DLP; mGy*cm) was compared between both modalities. Descriptive statistics, Wilcoxon signed rank test (P < 0.05), Student's t test (P < 0.05), and Cohen's kappa (κ) for interreader reliability were calculated. RESULTS: In 46 patients (28 males; 18 females; mean age, 53 ± 20 years) with 28 hand/wrist and 18 foot/ankle examinations, 86 out of 92 fractures were diagnosed with 3D tomography compared with CT. No false-positive finding occurred at 3D tomography. The six missed fractures on 3D tomography were five avulsion fractures of the carpals/metacarpals or tarsals/metatarsals, respectively, and one nondisplaced fracture of the capitate. Interreader agreement of anatomical location and morphological characteristics was substantial to almost perfect for upper (κ = 0.80-0.96) and lower (κ = 0.70-0.97) extremity fractures. Visibility of bone and soft tissue structures and image quality were slightly inferior using 3D tomography compared with CT (upper extremity P < 0.001-0.038 and lower extremity P < 0.001-0.035). DLP of a comparable scan coverage was significantly lower for 3D tomography (P < 0.001) for both upper (3D mean, 19.4 ± 5.9 mGy*cm; estimated CT mean, 336.5 ± 52.2 mGy*cm) and lower extremities (3D mean, 24.1 ± 11.1 mGy*cm; estimated CT mean, 182.9 ± 6.5 mGy*cm). Even the highest DLP with 3D tomography was < 30% of the mean estimated CT dose of a comparable area of coverage. CONCLUSION: Fracture assessment of peripheral extremities is reliable utilizing a low-dose 3D tomography X-ray system, with slightly reduced image quality.

Prevalence of abnormal findings in 230 knees of asymptomatic adults using 3.0 T MRI.

OBJECTIVE: To identify abnormalities in asymptomatic sedentary individuals using 3.0 Tesla high-resolution MRI. MATERIALS AND METHODS: The cohort comprised of 230 knees of 115 uninjured sedentary adults (51 males, 64 females; median age: 44 years). All participants had bilateral knee 3.0 T MRIs. Two senior musculoskeletal radiologists graded all intraarticular knee structures using validated scoring systems. Participants completed Knee Injury and Osteoarthritis Outcome Score questionnaires at the time of the MRI scan. RESULTS: MRI showed abnormalities in the majority (97%) of knees. Thirty percent knees had meniscal tears: horizontal (23%), complex (3%), vertical (2%), radial (2%) and bucket handle (1%). Cartilage and bone marrow abnormalities were prevalent at the patellofemoral joint (57% knees and 48% knees, respectively). Moderate and severe cartilage lesions were common, in 19% and 31% knees, respectively, while moderate and severe bone marrow oedema in 19% and 31% knees, respectively. Moderate-intensity lesion in tendons was found in 21% knees and high-grade tendonitis in 6% knees-the patellar (11% and 2%, respectively) and quadriceps (7% and 2%, respectively) tendons being most affected. Three percent partial ligamentous ruptures were found, especially of the anterior cruciate ligament (2%). CONCLUSION: Nearly all knees of asymptomatic adults showed abnormalities in at least one knee structure on MRI. Meniscal tears, cartilage and bone marrow lesions of the patellofemoral joint were the most common pathological findings. Bucket handle and complex meniscal tears were reported for the first time in asymptomatic knees.

Is the immediate effect of marathon running on novice runners' knee joints sustained within 6 months after the run? A follow-up 3.0 T MRI study.

OBJECTIVE: To evaluate changes in the knee joints of asymptomatic first-time marathon runners, using 3.0 T MRI, 6 months after finishing marathon training and run. MATERIALS AND METHODS: Six months after their participation in a baseline study regarding their knee joints, 44 asymptomatic novice marathoners (17 males, 27 females, mean age 46 years old) agreed to participate in a repeat MRI investigation: 37 completed both a standardized 4-month-long training programme and the marathon (marathon runners); and 7 dropped out during training (pre-race dropouts). The participants already underwent bilateral 3.0 T MRIs: 6 months before and 2 weeks after their first marathon, the London Marathon 2017. This study was a follow-up assessment of their knee joints. Each knee structure was assessed using validated scoring/grading systems at all time points. RESULTS: Two weeks after the marathon, 3 pre-marathon bone marrow lesions and 2 cartilage lesions showed decrease in radiological score on MRI, and the improvement was sustained at the 6-month follow-up. New improvements were observed on MRI at follow-up: 5 pre-existing bone marrow lesions and 3 cartilage lesions that remained unchanged immediately after the marathon reduced in their extent 6 months later. No further lesions appeared at follow-up, and the 2-week post-marathon lesions showed signs of reversibility: 10 of 18 bone marrow oedema-like signals and 3 of 21 cartilage lesions decreased on MRI. CONCLUSION: The knees of novice runners achieved sustained improvement, for at least 6 months post-marathon, in the condition of their bone marrow and articular cartilage.

Normal Anatomy and Traumatic Injury of the Midtarsal (Chopart) Joint Complex: An Imaging Primer.

The midtarsal (Chopart) joint complex consists of the talonavicular and calcaneocuboid joints and their stabilizing ligaments. Detailed assessment of this complex at MRI can be challenging owing to frequent anatomic variation and the small size of the structures involved. Nevertheless, a wide spectrum of pathologic conditions affect the joint complex, and its imaging evaluation deserves more thorough consideration. This review focuses on MRI evaluation of normal ligamentous anatomy and common variations about the Chopart joint, presenting practical imaging tips and potential diagnostic pitfalls. Imaging findings across a spectrum of traumatic Chopart joint injuries are also reviewed, from midtarsal sprains to Chopart fracture-dislocations. Midtarsal sprains-commonly associated with ankle inversion injuries-are emphasized, along with their often predictable radiographic and MRI injury patterns. Online DICOM image stacks are available for this article. ©RSNA, 2018.

Artificial Intelligence in Musculoskeletal Imaging: Review of Current Literature, Challenges, and Trends.

Artificial intelligence (AI) has gained major attention with a rapid increase in the number of published articles, mostly recently. This review provides a general understanding of how AI can or will be useful to the musculoskeletal radiologist. After a brief technical background on AI, machine learning, and deep learning, we illustrate, through examples from the musculoskeletal literature, potential AI applications in the various steps of the radiologist's workflow, from managing the request to communication of results. The implementation of AI solutions does not go without challenges and limitations. These are also discussed, as well as the trends and perspectives.

Meniscal pathologies on MRI correlate with increased bone tracer uptake in SPECT/CT.

OBJECTIVES: To assess the relationship of subchondral bone tracer uptake (BTU) on SPECT/CT and meniscal pathologies on MRI in patients with painful knees. METHODS: Twenty-five patients who had MRI and SPECT/CT within 3 months without knee surgery or grade ≥3 cartilage lesions were prospectively included. Maximum values of each subchondral femorotibial area were quantified and a ratio was calculated in relation to a femoral shaft reference region, which represented the BTU background activity. Meniscal lesions were graded (intact/degeneration/tear) and meniscal extrusion (no/yes) was assessed using MRI by two musculoskeletal radiologists blinded to the SPECT/CT findings. One-tailed Spearman correlations served for statistics (p < 0.05). RESULTS: Knees with meniscal degeneration or tear showed a significantly higher BTU in the medial femorotibial compartment (p = 0.045) when compared to intact menisci. Meniscal degeneration was associated with an increased BTU in the lateral femorotibial compartment; however, this was not statistically significant (p = 0.143). Patients with an extruded meniscus showed significantly higher BTU compared to a non-extruded meniscus (p < 0.020). CONCLUSIONS: Medial femorotibial BTU in SPECT/CT was associated with meniscal pathologies. Highest BTU was found in patients with meniscal tears. SPECT/CT appears to be a useful imaging modality to identify patients with overloading or early osteoarthritis. KEY POINTS: • Meniscal degeneration and tears correlate significantly with increased BTU using SPECT/CT. • Medial meniscus extrusion is associated with an increased BTU in SPECT/CT. • SPECT/CT allows detection of overloading and early osteoarthritis.

Imaging of Chopart (Midtarsal) Joint Complex: Normal Anatomy and Posttraumatic Findings.

OBJECTIVE: The objective of this article is to review the normal anatomy and posttraumatic findings of the Chopart joint complex. Key imaging features of the normal ligaments and patterns of ligamentous and osseous injuries are discussed. CONCLUSION: Traumatic midtarsal injuries, particularly midtarsal sprain, are often overlooked clinically and on imaging but are relatively common and typically are associated with inversion ankle injuries. Radiologists should be familiar with Chopart joint anatomy and the imaging features of midtarsal injuries because early diagnosis may help optimize clinical management.

Acute Fracture of the Anterior Process of Calcaneus: Does It Herald a More Advanced Injury to Chopart Joint?

OBJECTIVE: Injuries at the calcaneocuboid and talonavicular joint have been described as two distinct, unrelated entities in the radiology literature. Our purpose was to assess the coexistence of these injuries using radiography and MRI and to correlate our findings with radiologic and clinical diagnoses. MATERIALS AND METHODS: Twenty-one patients with injury at the anterior calcaneal process on radiographs or MR images were retrospectively assessed for concomitant injury at the talonavicular joint. Radiologic and clinical diagnoses and treatment were documented. McNemar and kappa statistics were calculated; p values < 0.05 were considered statistically significant. RESULTS: Radiographic and MRI rates of detection of injuries across the Chopart joint were statistically different. Calcaneocuboid avulsion fractures were evident on 48% of radiographs and 100% of MR images (p = 0.001). Talonavicular joint injuries were evident on 38% of radiographs and 76% of MR images (p = 0.008). Concomitant injury at both joints was evident on 14% of radiographs and 76% of MR images (p < 0.0001). Interrater agreement was 0.488-0.637 and 0.286-0.364 for talonavicular and 0.144-0.538 and 0.976-1 for calcaneocuboid injuries on radiography and MRI, respectively. Sixty percent of calcaneocuboid fractures were prospectively missed on radiography (none on MRI), whereas 38% and 25% of talonavicular findings were missed on radiography and MRI, respectively. Sixty percent of injuries were clinically misdiagnosed as ankle sprains. Chopart joint injury was never mentioned in prospective clinical or imaging diagnoses. CONCLUSION: Calcaneocuboid and talonavicular injuries commonly coexist. Radiographs underestimate severity of injury; MR images show more subtle abnormalities. Lack of mention of Chopart joint injury clinically and on imaging reports underlies the need for greater familiarity with this entity.

JOURNAL CLUB: MRI Evaluation of Midtarsal (Chopart) Sprain in the Setting of Acute Ankle Injury.

OBJECTIVE: This study determined the frequency and MRI appearance of osseous and ligamentous injuries in midtarsal (Chopart) sprains and their association with ankle sprains after acute ankle injuries. Prospective diagnosis of and interobserver agreement regarding midtarsal injury among musculoskeletal radiologists were also assessed. SUBJECTS AND METHODS: Two cohorts with ankle MRI were identified via a digital PACS search: patients who had undergone MRI within 8 weeks after ankle injury and control subjects who had not sustained ankle trauma. Studies were retrospectively reviewed in consensus as well as independently, assessing ligamentous and osseous injury to the Chopart joint (calcaneocuboid and talonavicular joints) and associated lateral collateral and deltoid ligamentous injury. Interobserver agreement was calculated, and prospective radiology reports were reviewed to determine the musculoskeletal radiologist's familiarity with Chopart joint injury. RESULTS: MR images of control subjects (n = 16) and patients with ankle injury (n = 47) were reviewed. The normal dorsal calcaneocuboid and calcaneocuboid component of bifurcate ligaments were variably visualized; the remaining normal ligaments were always seen. Eleven patients (23%) had midtarsal ligamentous and osseous injury consistent with midtarsal sprain (eight acute or subacute, one probable, and two old). Six (75%) of eight acute or subacute cases had coexisting lateral collateral ligament injury. Eighty-nine percent of osseous injuries were reported prospectively, but 83% of ligamentous injuries were missed. Substantial interobserver agreement was achieved regarding diagnosis of midtarsal sprain. CONCLUSION: Midtarsal sprains are commonly associated with acute ankle injury and with ankle sprains. Presently, midtarsal sprains may be underrecognized by radiologists; thus, greater familiarity with the MRI spectrum of ligamentous and osseous injuries at the Chopart joint is important for accurate diagnosis and clinical management.