Deep learning-based automatic pipeline for quantitative assessment of thigh muscle morphology and fatty infiltration.

PURPOSE: Fast and accurate thigh muscle segmentation from MRI is important for quantitative assessment of thigh muscle morphology and composition. A novel deep learning (DL) based thigh muscle and surrounding tissues segmentation model was developed for fully automatic and reproducible cross-sectional area (CSA) and fat fraction (FF) quantification and tested in patients at 10 years after anterior cruciate ligament reconstructions. METHODS: A DL model combining UNet and DenseNet was trained and tested using manually segmented thighs from 16 patients (32 legs). Segmentation accuracy was evaluated using Dice similarity coefficients (DSC) and average symmetric surface distance (ASSD). A UNet model was trained for comparison. These segmentations were used to obtain CSA and FF quantification. Reproducibility of CSA and FF quantification was tested with scan and rescan of six healthy subjects. RESULTS: The proposed UNet and DenseNet had high agreement with manual segmentation (DSC >0.97, ASSD < 0.24) and improved performance compared with UNet. For hamstrings of the operated knee, the automated pipeline had largest absolute difference of 6.01% for CSA and 0.47% for FF as compared to manual segmentation. In reproducibility analysis, the average difference (absolute) in CSA quantification between scan and rescan was better for the automatic method as compared with manual segmentation (2.27% vs. 3.34%), whereas the average difference (absolute) in FF quantification were similar. CONCLUSIONS: The proposed method exhibits excellent accuracy and reproducibility in CSA and FF quantification compared with manual segmentation and can be used in large-scale patient studies.

3D MRI of the Knee.

Three-dimensional (3D) magnetic resonance imaging (MRI) of the knee is widely used in musculoskeletal (MSK) imaging. Currently, 3D sequences are most commonly used for morphological imaging. Isotropic 3D MRI provides higher out-of-plane resolution than standard two-dimensional (2D) MRI, leading to reduced partial volume averaging artifacts and allowing for multiplanar reconstructions in any plane with any thickness from a single high-resolution isotropic acquisition. Specifically, isotropic 3D fast spin-echo imaging, with options for tissue weighting similar to those used in multiplanar 2D FSE imaging, is of particular interest to MSK radiologists. New applications for 3D spatially encoded sequences are also increasingly available for clinical use. These applications offer advantages over standard 2D techniques for metal artifact reduction, quantitative cartilage imaging, nerve imaging, and bone shape analysis. Emerging fast imaging techniques can be used to overcome the long acquisition times that have limited the adoption of 3D imaging in clinical protocols.

Inter-rater agreement of rotator cuff tendon and muscle magnetic resonance imaging parameters evaluated preoperatively and during the first postoperative year following rotator cuff repair.

BACKGROUND: Magnetic resonance imaging (MRI) is standard of care for rotator cuff evaluation, with clinical interpretation usually limited to qualitative judgments. The reliability of MRI-based measurements and scoring systems has been evaluated only preoperatively or ≥6 months following rotator cuff repair, when repairs are in the later stages of healing. This study describes the MRI assessments and inter-rater agreement of various rotator cuff tendon and muscle parameters evaluated preoperatively and 4 times during the first postoperative year. METHODS: Two musculoskeletal radiologists independently assessed MRI scans of 42 patients preoperatively and 3, 12, 26, and 52 weeks after rotator cuff repair. Using standardized reading rules, readers assessed tendon integrity (5-point Sugaya classification), tear dimensions, muscle fat (5-point Goutallier classification) and atrophy (4-point Warner classification), muscle cross-sectional areas, and myotendinous junction distance. Raw exact agreement proportions, κ statistics, and correlation coefficients were used to quantify inter-rater agreement. RESULTS: Readers showed moderate to substantial above-chance agreement in scoring rotator cuff tendon integrity and supraspinatus muscle atrophy and good to excellent agreement on tear dimensions and muscle cross-sectional areas but only fair to moderate agreement for fatty infiltration and myotendinous junction distance. Only fatty infiltration grades evidenced observer bias. Inter-rater agreement did not appear time dependent. CONCLUSION: By use of defined reading rules in a research setting, MRI evaluations of rotator cuff tendon integrity, tear dimensions, muscle atrophy, and cross-sectional areas have reasonable reliability at all time points in the first postoperative year. However, the presence of clinically significant disagreements, even in such favorable circumstances, indicates the need for improved imaging tools for precise rotator cuff evaluation.

Nomenclature of Subchondral Nonneoplastic Bone Lesions.

OBJECTIVE. The purpose of this article is to summarize the nomenclature of nonneoplastic conditions affecting subchondral bone through a review of the medical literature and expert opinion of the Society of Skeletal Radiology Subchondral Bone Nomenclature Committee. CONCLUSION. This consensus statement summarizes current understanding of the pathophysiologic characteristics and imaging findings of subchondral nonneoplastic bone lesions and proposes nomenclature to improve effective communication across clinical specialties and help avoid diagnostic errors that could affect patient care.

Upper extremity myxofibrosarcoma mimicking an erosive inflammatory arthritis: a case report.

Myxofibrosarcoma is a malignant fibroblastic soft tissue neoplasm containing a variable amount of myxoid stroma that commonly presents as a slow-growing mass in elderly patients. The neoplasm may be superficial or deep to the muscle fascia and characteristically has an infiltrative growth pattern with a dominant or multinodular mass. We describe an unusual case of high-grade myxofibrosarcoma of the wrist and forearm that infiltrated the muscles, tendons, and wrist joint, causing bone erosions. The tumor was mistakenly diagnosed as synovitis and a chronic, erosive, inflammatory process. The diffuse nature, absence of a dominant mass, and radiographic appearance complicated the diagnosis. Although neoplasms of the synovial spaces are rare, this case demonstrates that tumors with a highly infiltrative growth pattern can mimic inflammatory synovitis and that neoplasms should be considered in the differential diagnosis when clinical and laboratory features are discordant with the imaging appearance.

In vitro chondrocyte toxicity following long-term, high-dose exposure to Gd-DTPA and a novel cartilage-targeted MR contrast agent.

OBJECTIVE: To determine the concentrations exhibiting toxicity of a cartilage-targeted magnetic resonance imaging contrast agent compared with gadopentetate dimeglumine (Gd-DT-PA) in chondrocyte cultures. MATERIALS AND METHODS: A long-term Swarm rat chondrosarcoma chondrocyte-like cell line was exposed for 48 h to 1.0-20 mM concentrations of diaminobutyl-linked nitroxide (DAB4-DLN) citrate, 1.0-20 mM Gd-DTPA, 1.0 µM staurosporine (positive control), or left untreated. Cell appearance, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays of metabolic activity, quantitative PicoGreen assays of DNA content, and calcein-AM viability assays were compared. RESULTS: At 1.0-7.5 mM, minimal decrease in cell proliferation was found for both agents. At all doses of both agents, cell culture appearances were similar after 24 h of treatment. At the higher doses, differences in cell culture appearance were found after 48 h of treatment, with dose-dependent declines in chondrocyte populations for both agents. Concentration-dependent declines in DNA content and calcein fluorescence were found after 48 h of treatment, but beginning at a lower dose of DAB4-DLN citrate than Gd-DTPA. Dose-dependent decreases in MTT staining (cell metabolism) were apparent for both agents, but larger effects were evident at a lower dose for DAB-DLN citrate. Poor MTT staining of cells exposed for 48 h to 20 mM DAB4-DLN citrate probably indicates dead or dying cells. CONCLUSION: The minimal effect of the long-term exposure of model chondrocyte cell cultures to DAB4-DLN citrate and Gd-DTPA concentrations up to 7.5 mM (3x typical arthrographic administration) is supporting evidence that these doses are acceptable for MR arthrography. The findings are reassuring given that the experimental exposure to the contrast agents at sustained concentrations was much longer than when used clinically.

State of the Art: MR Imaging after Knee Cartilage Repair Surgery.

Cartilage injuries are common, especially in athletes. Because these injuries frequently affect young patients, and they have the potential to progress to osteoarthritis, treatment to alleviate symptoms and delay joint degeneration is warranted. A number of surgical techniques are available to treat focal chondral defects, including marrow stimulation, osteochondral auto- and allografting, and autologous chondrocyte implantation. Although arthroscopy is considered the standard of reference for the evaluation of cartilage before and after repair, it is invasive with associated morbidity and cannot adequately depict the deep cartilage layer and underlying bone. Magnetic resonance (MR) imaging provides unparalleled noninvasive assessment of the repair site and all other joint tissues. MR observation of cartilage repair tissue is a well-established semiquantitative scoring system for repair tissue that has primarily been used in clinical research studies. The cartilage repair osteoarthritis knee score (CROAKS) optimizes comprehensive morphologic assessment of the knee joint after cartilage repair. Furthermore, quantitative, compositional MR imaging measurements (eg, T2, T2*, T1ρ), delayed gadolinium-enhanced MR imaging of cartilage (dGEMRIC), and sodium imaging are available for biochemical assessment. These quantitative MR imaging techniques help assess collagen content and orientation, water content, and glycosaminoglycan and/or proteoglycan content both in the repair tissue as it matures and in the "native" cartilage. In this review, the authors discuss the principles of state-of-the-art morphologic and compositional MR imaging techniques for imaging of cartilage repair and their application to longitudinal studies.

Evaluation of the Articular Cartilage of the Knee Joint Using an Isotropic Resolution 3D Fast Spin-Echo Sequence With Conventional and Radial Reformatted Images.

OBJECTIVE: The purpose of this study was to determine whether the use of radial reformatted images could improve the diagnostic performance of a 3D fast spin-echo (FSE) sequence for detecting surgically confirmed cartilage lesions within the knee joint. MATERIALS AND METHODS: An MRI examination consisting of five 2D FSE sequences and a sagittal 3D FSE sequence was performed at 3 T on the knee joint of 150 patients who underwent subsequent knee arthroscopy, which included grading of the articular cartilage. Conventional axial, sagittal, and coronal reformatted images and radial reformatted images were created from the 3D FSE source data. Two musculoskeletal radiologists independently used the 2D FSE sequences, the 3D FSE sequence with conventional reformatted images only, and the 3D FSE sequence with both radial and conventional reformatted images at three separate sessions to grade each articular surface of the knee joint. McNemar tests were used to compare diagnostic performance for detecting cartilage lesions using arthroscopy as the reference standard. RESULTS: The 3D FSE sequence with radial and conventional reformatted images had higher sensitivity (p < 0.001) and similar specificity (p = 0.73) to the 2D FSE sequences for detecting cartilage lesions and higher sensitivity (p < 0.001) and specificity (p = 0.002) than the 3D FSE sequence with conventional reformatted images for detecting cartilage lesions. The 3D FSE sequence with conventional reformatted images had similar sensitivity (p = 0.93) and lower specificity (p = 0.005) than did the 2D FSE sequences for detecting cartilage lesions. CONCLUSION: A 3D FSE sequence had improved diagnostic performance compared with 2D FSE sequences for detecting cartilage lesions within the knee joint but only when using both radial and conventional reformatted images for cartilage evaluation.

Pseudocyclops: two cases of ACL graft partial tears mimicking cyclops lesions on MRI.

Arthroscopic reconstruction of the anterior cruciate ligament (ACL) using autografts or allografts is a common surgical procedure, particularly in young athletes. Although the procedure has excellent success rates, complications such as mechanical impingement, graft rupture, and arthrofibrosis can occur, often necessitating additional surgery. Magnetic resonance (MR) imaging has become a valuable tool in evaluating complications after ACL reconstruction. We report two cases of ACL reconstruction complicated by arthroscopically proven partial graft tears. In both cases the torn anterior graft fibers were flipped into the intercondylar notch, mimicking anterior arthrofibrosis, i.e., a "cyclops lesion," on MR imaging. Careful review of the direction of graft fibers on MR imaging in the "pseudocyclops" lesions can help differentiate these partial tears from the fibrosis of a true cyclops. The "pseudocyclops" lesion is a previously undescribed MR imaging sign of partial ACL graft tear. Larger studies are required to determine the sensitivity and specificity of the sign, as well as the clinical importance of these partial graft tears.

MRI diagnosis of muscle denervation from herpes zoster with discordant distribution of the skin rash.

Herpes zoster is a common disorder characterized by a painful rash along a dermatome caused by reactivation of the varicella zoster virus (VZV). Muscle denervation injury from motor involvement is an uncommon phenomenon. Discordant distribution of the skin rash and motor nerve involvement, presenting as a skin rash in one body part and muscle weakness or pain from nerve involvement in another body part is an even more uncommonly reported finding. We present an unusual case of muscle denervation injury resulting from motor involvement of a peripheral nerve by VZV diagnosed by magnetic resonance imaging with cutaneous manifestations in a different dermatomal distribution. To the best of our knowledge, there has been no similar case reported in the English radiology literature. We suggest that whenever a radiologist notices MRI findings suggesting denervation injury and a cause not readily identified, VZV-related denervation injury should be included in the differential diagnosis, especially in an older immunocompromised patient.

The MOCART (magnetic resonance observation of cartilage repair tissue) 2.0 Ankle Score.

OBJECTIVES: The aim of this study was to introduce the MOCART 2.0 ankle score and evaluate its utility and reproducibility for the radiological assessment of cartilage repair tissue in the ankle joint. METHODS: The MOCART 2.0 ankle score evaluates seven individual variables, including "volume fill of (osteo)chondral defect," "Integration into adjacent cartilage and bone," "surface of the repair tissue," "signal intensity of the repair tissue," "bony defect and bony overgrowth," "presence of edema-like-marrow signal," and "presence of subchondral cysts." Overall, a MOCART 2.0 ankle score between 0 and 100 points may be reached. Two independent readers assessed the 3-T MRI examinations of 48 ankles, who had undergone cartilage repair of a talar cartilage defect using the new MOCART 2.0 ankle score. One of the readers performed two readings. Intra- and interrater reliability were assessed using intraclass correlation coefficients (ICCs) for the overall MOCART 2.0 ankle score. RESULTS: Forty-eight ankles (mean age at surgery 30.2 ± 11.2 years) were evaluated. The overall interrater (ICC = 0.75; 95%CI 0.60-0.85), as well as the intrarater (ICC = 0.83; 95%CI 0.72-0.90) reliability of the MOCART 2.0 ankle score was good. For individual variables the interrater reliability ranged from a kappa value of 0.29 (95%CI 0.01-0.57) for "surface of the repair tissue" to 0.83 (95%CI 0.71-0.95) for "presence of subchondral cysts". CONCLUSIONS: The newly introduced MOCART 2.0 ankle score, which encompasses the distinct anatomy of the ankle joint, demonstrates good intra- and interrater reliability. CRITICAL RELEVANCE STATEMENT: The newly introduced MOCART 2.0 ankle score may facilitate the standardized assessment of cartilage repair in the ankle joint and allow an objective comparison of the morphological outcome between alternative treatment options and between different studies. KEY POINTS: This study introduces the MOCART 2.0 ankle score. The MOCART 2.0 ankle score demonstrated good intra- and interrater reliability. Standardized reporting may improve communication between radiologists and other physicians.

Randomized controlled trial of osteoconductive fixation screws for anterior cruciate ligament reconstruction: a comparison of the Calaxo and Milagro screws.

PURPOSE: To compare the outcome of 2 bioabsorbable screws for tibial interference fixation in anterior cruciate ligament reconstruction with reference to rate of absorption, osteoconductive properties, and clinical outcome. METHODS: Patients undergoing primary anterior cruciate ligament reconstruction with hamstring autograft in a single unit were invited to participate in this study. Patients were randomized to receive either the Calaxo screw (Smith & Nephew, Andover, MA) or Milagro screw (DePuy Mitek, Raynham, MA) for tibial fixation. Patients were reviewed with subjective and objective evaluation by use of the International Knee Documentation Committee form, Lysholm score, KT-1000 arthrometry (MEDmetric, San Diego, CA), and clinical examination. Magnetic resonance imaging was performed at 1 year and computed tomography scanning at 1 week and at 6, 12, and 24 months. RESULTS: Sixty patients agreed to participate in the study, with 32 patients randomized to the Calaxo screw and 28 to the Milagro screw for tibial fixation. There was no significant difference in subjective or objective clinical outcome between the 2 groups. At 24 months, 88% of Calaxo screws showed complete screw resorption compared with 0% of Milagro screws (P < .001). Tibial cysts were present in 88% of the Calaxo group and 7% of the Milagro group (P = .001). At 24 months, the mean volume of new bone formation for the Calaxo group was 21% of original screw volume. Ossification of the Milagro screw was unable to be accurately assessed as a result of incomplete screw resorption. CONCLUSIONS: Both screws showed similar favorable objective and subjective outcomes at 2 years. The Calaxo screw resorbed completely over a period of 6 months and was associated with a high incidence of intra-tunnel cyst formation. The Milagro screw increased in volume over a period of 6 months, followed by a gradual resorption, which was still ongoing at 2 years. Both screws were associated with tunnel widening, and neither showed evidence of significant tunnel ossification. We conclude that, despite satisfactory clinical outcomes, the addition of "osteoconductive" materials to bioabsorbable screws is not associated with bone formation at the screw site at 2 years. LEVEL OF EVIDENCE: Level I, randomized controlled trial.

Quantitative MRI for Evaluation of Musculoskeletal Disease: Cartilage and Muscle Composition, Joint Inflammation, and Biomechanics in Osteoarthritis.

ABSTRACT: Magnetic resonance imaging (MRI) is a valuable tool for evaluating musculoskeletal disease as it offers a range of image contrasts that are sensitive to underlying tissue biochemical composition and microstructure. Although MRI has the ability to provide high-resolution, information-rich images suitable for musculoskeletal applications, most MRI utilization remains in qualitative evaluation. Quantitative MRI (qMRI) provides additional value beyond qualitative assessment via objective metrics that can support disease characterization, disease progression monitoring, or therapy response. In this review, musculoskeletal qMRI techniques are summarized with a focus on techniques developed for osteoarthritis evaluation. Cartilage compositional MRI methods are described with a detailed discussion on relaxometric mapping (T 2 , T 2 *, T 1ρ ) without contrast agents. Methods to assess inflammation are described, including perfusion imaging, volume and signal changes, contrast-enhanced T 1 mapping, and semiquantitative scoring systems. Quantitative characterization of structure and function by bone shape modeling and joint kinematics are described. Muscle evaluation by qMRI is discussed, including size (area, volume), relaxometric mapping (T 1 , T 2 , T 1ρ ), fat fraction quantification, diffusion imaging, and metabolic assessment by 31 P-MR and creatine chemical exchange saturation transfer. Other notable technologies to support qMRI in musculoskeletal evaluation are described, including magnetic resonance fingerprinting, ultrashort echo time imaging, ultrahigh-field MRI, and hybrid MRI-positron emission tomography. Challenges for adopting and using qMRI in musculoskeletal evaluation are discussed, including the need for metal artifact suppression and qMRI standardization.

Patellar Dislocation in Adolescent Patients: Influence on Cartilage Properties Based on T1ρ Relaxation Times.

BACKGROUND: Adolescents who experience a patellar dislocation have an elevated risk of patellofemoral posttraumatic osteoarthritis. Magnetic resonance imaging (MRI)-based T1ρ relaxation times were measured for adolescents to evaluate patellofemoral cartilage after patellar dislocation. Long T1ρ relaxation times are an indicator of cartilage degradation. HYPOTHESIS: The primary hypothesis is that patellofemoral cartilage T1ρ relaxation times will be elevated in the acute phase after patellar dislocation. The secondary hypothesis is that T1ρ relaxation times will be higher for knees with multiple rather than single dislocations due to repeated traumatic injury. STUDY DESIGN: Cross-sectional study; Level of evidence, 3. METHODS: In total, 23 adolescents being treated for a recent patellar dislocation, 13 for a first-time dislocation (47 ± 38 days since most recent dislocation) and 10 for multiple dislocations (55 ± 24 days since most recent dislocation), and 10 healthy controls participated in MRI-based T1ρ relaxation time mapping. For multiple regions of the patellofemoral joint, mean T1ρ values were compared between the 3 groups with multiple group comparisons and post hoc tests. T1ρ relaxation times were also correlated against measures of patellofemoral anatomy and alignment for single and multiple dislocations. Statistical significance was set at P < .05. RESULTS: T1ρ relaxation times were significantly longer for injured knees (single and multiple dislocations) than controls at the medial and central patella and central trochlear groove. For the regions on the patella, significant differences between injured and control knees exceeded 15%. No significant differences were identified between single and multiple dislocations. For the initial dislocation group, T1ρ relaxation times within multiple regions of the patellofemoral joint were significantly correlated with lateral patellar alignment or patellar height. CONCLUSION: Elevated patellofemoral cartilage T1ρ relaxation times are consistent with a high risk of long-term patellofemoral osteoarthritis for adolescents who experience patellar dislocations. T1ρ relaxation times were elevated for multiple regions of patellofemoral cartilage. T1ρ relaxation times were expected to increase with additional dislocation episodes, but relaxation times after single and multiple dislocations were similar. After a first dislocation, parameters related to patellar maltracking were correlated with cartilage degradation.

The Corticosteroid Meniscectomy Trial of Extended-Release Triamcinolone Injection After Arthroscopic Partial Meniscectomy: Protocol for a Double-Blind Randomized Controlled Trial.

Background: Meniscal tear in older adults often accompanies knee osteoarthritis and is commonly treated with arthroscopic partial meniscectomy (APM) when patients have persistent pain after a trial of physical therapy. Cross-sectional evidence suggests that synovitis is associated with baseline pain in this patient population, but little is known about the relationship between synovitis and postoperative recovery or progression of knee osteoarthritis. Purpose/Hypothesis: Intra-articular extended-release triamcinolone may reduce inflammation and thereby improve outcomes and slow disease progression. This article presents the rationale behind the Corticosteroid Meniscectomy Trial (CoMeT) and describes its study design and implementation strategies. Study Design: Randomized controlled trial. Methods: CoMeT is a 2-arm, 3-center, randomized placebo-controlled trial designed to establish the clinical efficacy of extended-release triamcinolone administered via intra-articular injection immediately after APM. The primary outcome is change in Knee injury and Osteoarthritis Outcome Score Pain subscore at 3-month follow-up. Synovial biopsy, joint fluid aspirate, and urine and blood sample analyses will examine the associations between various objective measures of baseline inflammation and pre- and postoperative outcome measures and clinical responses to triamcinolone intervention. Quantitative 3-T magnetic resonance imaging will evaluate cartilage and meniscal composition and 3-dimensional bone shape to detect early joint degeneration. Results: We discuss methodologic innovations and challenges. Conclusion: To our knowledge, this is the first randomized double-blind clinical trial that will analyze the effect of extended-release triamcinolone acetonide on pain, magnetic resonance imaging measures of structural change and effusion/synovitis, soluble biomarkers, and synovial tissue transcriptomics after APM.

Deep Learning Diagnosis and Classification of Rotator Cuff Tears on Shoulder MRI.

BACKGROUND: Detection of rotator cuff tears, a common cause of shoulder disability, can be time-consuming and subject to reader variability. Deep learning (DL) has the potential to increase radiologist accuracy and consistency. PURPOSE: The aim of this study was to develop a prototype DL model for detection and classification of rotator cuff tears on shoulder magnetic resonance imaging into no tear, partial-thickness tear, or full-thickness tear. MATERIALS AND METHODS: This Health Insurance Portability and Accountability Act-compliant, institutional review board-approved study included a total of 11,925 noncontrast shoulder magnetic resonance imaging scans from 2 institutions, with 11,405 for development and 520 dedicated for final testing. A DL ensemble algorithm was developed that used 4 series as input from each examination: fluid-sensitive sequences in 3 planes and a sagittal oblique T1-weighted sequence. Radiology reports served as ground truth for training with categories of no tear, partial tear, or full-thickness tear. A multireader study was conducted for the test set ground truth, which was determined by the majority vote of 3 readers per case. The ensemble comprised 4 parallel 3D ResNet50 convolutional neural network architectures trained via transfer learning and then adapted to the targeted domain. The final tear-type prediction was determined as the class with the highest probability, after averaging the class probabilities of the 4 individual models. RESULTS: The AUC overall for supraspinatus, infraspinatus, and subscapularis tendon tears was 0.93, 0.89, and 0.90, respectively. The model performed best for full-thickness supraspinatus, infraspinatus, and subscapularis tears with AUCs of 0.98, 0.99, and 0.95, respectively. Multisequence input demonstrated higher AUCs than single-sequence input for infraspinatus and subscapularis tendon tears, whereas coronal oblique fluid-sensitive and multisequence input showed similar AUCs for supraspinatus tendon tears. Model accuracy for tear types and overall accuracy were similar to that of the clinical readers. CONCLUSIONS: Deep learning diagnosis of rotator cuff tears is feasible with excellent diagnostic performance, particularly for full-thickness tears, with model accuracy similar to subspecialty-trained musculoskeletal radiologists.

MDCT arthrography of the shoulder with datasets of isotropic resolution: indications, technique, and applications.

OBJECTIVE: The purposes of this review are to summarize the indications for MDCT arthrography of the shoulder, highlight the features of MDCT acquisition, and describe the normal and abnormal MDCT arthrographic appearances of the shoulder. CONCLUSION: MDCT arthrography is a valid alternative for shoulder imaging of patients with contraindications to MRI or after failed MRI. MDCT arthrography is accurate for assessment of a variety of shoulder abnormalities and, with further validation, may become the imaging test of choice for evaluation of the postoperative shoulder.

A synthetic cartilage extracellular matrix model: hyaluronan and collagen hydrogel relaxivity, impact of macromolecular concentration on dGEMRIC.

OBJECTIVE: To develop and characterize the MR properties of a synthetic model for cartilage extra-cellular matrix using hydrogels and to determine the concentration dependence of spin-lattice (T1) and spin-spin (T2) relaxation times of hydrogels and their glycosaminoglycan and collagen components in the presence and absence of gadopentetate dimeglumine (Gd-DTPA) for use in dGEMRIC. MATERIALS AND METHODS: T1 and T2 measurements were made at 3 Tesla on a range of gelatin (i.e., collagen) and hyaluronan (i.e., glycosaminoglycan) solutions (6.25-100 g/l), alone, together in a composite, and as dityramine-bridged hydrogels. Relaxivity was calculated as a function of macromolecular concentration. RESULTS: Even at the highest concentrations, gelatin and hyaluronan solutions had T1 and T2 values significantly larger than those reported for cartilage. Only composite hydrogels with gelatin and hyaluronan concentrations naturally found in cartilage resulted in T1 values, but not T2 values, representative of cartilage. Relaxivities were slightly dependent on both hyaluronan concentration (R1 = 0.0027 l g(-1) s(-1); R2 = 0.025 l g(-1) s(-1)) and gelatin concentration (R1 = 0.0032 l g(-1) s(-1); R2 = 0.020 l g(-1) s(-1)) alone and as a composite (R1 = 0.0068 l g(-1) s(-1); R2 = 0.101 l g(-1) s(-1)). Gd-DTPA relaxivities were dependent upon macromolecular concentration and varied by 14-32% (R1 = 4.24 to 5.55 mM(-1) s(-1); R2 = 4.60 to 6.27 mM(-1) s(-1)) over the range of cartilage biochemistry. CONCLUSIONS: Without the contrast agent, hyaluronan and gelatin, alone or in a composite, have a very small impact on the relaxivities of the model system. The impact on R1 was approximately tenfold less than on R2. In contrast, macromolecular concentrations above 50 g/l significantly impacted Gd-DTPA relaxivity and should be accounted for when measuring the glycosaminoglycan content of cartilage in vivo using dGEMRIC.

Automated knee cartilage segmentation for heterogeneous clinical MRI using generative adversarial networks with transfer learning.

Background: This study aimed to build a deep learning model to automatically segment heterogeneous clinical MRI scans by optimizing a pre-trained model built from a homogeneous research dataset with transfer learning. Methods: Conditional generative adversarial networks pretrained on the Osteoarthritis Initiative MR images was transferred to 30 sets of heterogenous MR images collected from clinical routines. Two trained radiologists manually segmented the 30 sets of clinical MR images for model training, validation and test. The model performance was compared to models trained from scratch with different datasets, as well as two radiologists. A 5-fold cross validation was performed. Results: The transfer learning model obtained an overall averaged Dice coefficient of 0.819, an averaged 95 percentile Hausdorff distance of 1.463 mm, and an averaged average symmetric surface distance of 0.350 mm on the 5 random holdout test sets. A 5-fold cross validation had a mean Dice coefficient of 0.801, mean 95 percentile Hausdorff distance of 1.746 mm, and mean average symmetric surface distance of 0.364 mm. It outperformed other models and performed similarly as the radiologists. Conclusions: A transfer learning model was able to automatically segment knee cartilage, with performance comparable to human, using heterogeneous clinical MR images with a small training data size. In addition, the model proved robust when tested through cross validation and on images from a different vendor. We found it feasible to perform fully automated cartilage segmentation of clinical knee MR images, which would facilitate the clinical application of quantitative MRI techniques and other prediction models for improved patient treatment planning.

Elevated Patellofemoral and Tibiofemoral T1ρ Relaxation Times Following a First Time Patellar Dislocation.

OBJECTIVE: The study was performed to evaluate cartilage within the knee following a first-time patellar dislocation, using elevated MRI-based T1ρ relaxation times as an indicator of low proteoglycan concentration. The hypothesis is that MRI-based T1ρ relaxation times for patellofemoral and tibiofemoral cartilage are significantly longer for knees being treated for patellar dislocation than for healthy control knees. DESIGN: Twenty-one subjects being treated for a first-time, unilateral dislocation of the patella and 16 healthy controls participated in MRI-based T1ρ relaxation time mapping. Mean relaxation times were quantified for patellofemoral and tibiofemoral regions for injured knees, the contralateral knees, and healthy controls. T1ρ values for each region were compared between the 3 groups with generalized estimating equations. Linear regressions were also performed to correlate T1ρ relaxation times with time from injury. RESULTS: The knees with a disloction had longer T1ρ relaxation times than the contralateral knees and control group at the medial patella and longer relaxation times than the control group at the lateral tibia (P < 0.05). T1ρ relaxation times at the medial patella also decreased with time from injury (r2 = 0.21, P = 0.037). CONCLUSIONS: Compositional changes to cartilage on the medial patella are related to traumatic impact during a dislocation. Potential exists for cartilage properties at the medial patella to improve with time. Cartilage degradation at the lateral tibia is not directly related to traumatic impact. The current baseline data are a starting point to characterize the pathway from a first-time dislocation to progressive cartilage degradation and osteoarthritis.