Knee Magnetic Resonance Imaging Research Articles

Development of convolutional neural network model for diagnosing meniscus tear using magnetic resonance image

BackgroundDeep learning (DL) is an advanced machine learning approach used in diverse areas, such as image analysis, bioinformatics, and natural language processing. A convolutional neural network (CNN) is a representative DL model that is advantageous for image recognition and classification. In this study, we aimed to develop a CNN to detect meniscal tears and classify tear types using coronal and sagittal magnetic resonance (MR) images of each patient.MethodsWe retrospectively collected 599 cases (medial meniscus tear = 384, lateral meniscus tear = 167, and medial and lateral meniscus tear = 48) of knee MR images from patients with meniscal tears and 449 cases of knee MR images from patients without meniscal tears. To develop the DL model for evaluating the presence of meniscal tears, all the collected knee MR images of 1048 cases were used. To develop the DL model for evaluating the type of meniscal tear, 538 cases with meniscal tears (horizontal tear = 268, complex tear = 147, radial tear = 48, and longitudinal tear = 75) and 449 cases without meniscal tears were used. Additionally, a CNN algorithm was used. To measure the model’s performance, 70% of the included data were randomly assigned to the training set, and the remaining 30% were assigned to the test set.ResultsThe area under the curves (AUCs) of our model were 0.889, 0.817, and 0.924 for medial meniscal tears, lateral meniscal tears, and medial and lateral meniscal tears, respectively. The AUCs of the horizontal, complex, radial, and longitudinal tears were 0.761, 0.850, 0.601, and 0.858, respectively.ConclusionOur study showed that the CNN model has the potential to be used in diagnosing the presence of meniscal tears and differentiating the types of meniscal tears.

Reproducibility of MRI in the diagnosis of meniscal ramp lesions: an inter-observer study.

Meniscal ramp lesions have gained much prominence in recent years due to a significant increase in their diagnosis and their important biomechanical involvement in the knee. A new proposed classification of these lesions has recently been published. To evaluate the reproducibility of the new classification of meniscal ramp lesions recently published using magnetic resonance imaging (MRI). A total of 249 post-traumatic knee MRI studies were evaluated by three musculoskeletal radiologists independently. Patients with an anterior cruciate ligament (ACL) tear on MRI in addition to a recent history of trauma to the knee for <12 months were included in the study, for a total of 95. We carried out an inter-observer concordance study to analyze whether the new classification is reproducible to detect meniscal ramp lesions and to classify them into their different types. Among our study cohort of 95 patients, we found 47 (49.5%) ramp lesions. In the inter-observer study, we obtained a good concordance (k = 0.733) in the detection of these lesions and an excellent one (k = 0.843) when the type of lesion is described. This study demonstrates that the new classification of meniscal ramp lesions has good reproducibility on MRI.

Concurrent validity and reliability of a semi-automated approach to measuring the magnetic resonance imaging morphology of the knee joint in active youth.

Post-traumatic knee osteoarthritis is attributed to alterations in joint morphology, alignment, and biomechanics triggered by injury. While magnetic resonance (MR) imaging-based measures of joint morphology and alignment are relevant to understanding osteoarthritis risk, time consuming manual data extraction and measurement limit the number of outcomes that can be considered and deter widespread use. This paper describes the development and evaluation of a semi-automated software for measuring tibiofemoral and patellofemoral joint architecture using MR images from youth with and without a previous sport-related knee injury. After prompting users to identify and select key anatomical landmarks, the software can calculate 37 (14 tibiofemoral, 23 patellofemoral) relevant geometric features (morphology and alignment) based on established methods. To assess validity and reliability, 11 common geometric features were calculated from the knee MR images (proton density and proton density fat saturation sequences; 1.5 T) of 76 individuals with a 3-10-year history of youth sport-related knee injury and 76 uninjured controls. Spearman's or Pearson's correlation coefficients (95% CI) and Bland-Altman plots were used to assess the concurrent validity of the semi-automated software (novice rater) versus expert manual measurements, while intra-class correlation coefficients (ICC2,1; 95%CI), standard error of measurement (95%CI), 95% minimal detectable change, and Bland-Altman plots were used to assess the inter-rater reliability of the semi-automated software (novice vs resident radiologist rater). Correlation coefficients ranged between 0.89 (0.84, 0.92; Lateral Trochlear Inclination) and 0.97 (0.96, 0.98; Patellar Tilt Angle). ICC estimates ranged between 0.79 (0.63, 0.88; Lateral Patellar Tilt Angle) and 0.98 (0.95, 0.99; Bisect Offset). Bland-Altman plots did not reveal systematic bias. These measurement properties estimates are equal, if not better than previously reported methods suggesting that this novel semi-automated software is an accurate, reliable, and efficient alternative method for measuring large numbers of geometric features of the tibiofemoral and patellofemoral joints from MR studies.

Optimal Screw Length for Osteochondritis Dissecans Fixation of the Pediatric Distal Femur: A Magnetic Resonance Imaging Comparision Study of Pediatric Epiphyseal Thickness

Background:Magnetic Resonance Imaging (MRI) allows orthopaedic surgeons to assess Osteochondritis Dissecans (OCD) of the distal femur in children. When surgical fixation is indicated, headless compression screws can be used to stabilize the fragment and allow it to heal. During screw placement it is important to make sure the screw is not too long or it may cross the physis, leading to physeal growth arrest.Hypothesis/Purpose:The purpose of the study was to determine the optimal screw length for osteochondritis dissecans fixation of the pediatric distal femur. A secondary objective was to compare the thickness of the epiphysis of the pediatric distal femur at different locations and screw trajectories.Methods:This retrospective, IRB-approved, single-institution study identified 113 pediatric patients (ages 6-17 years old) with a knee MRI from January 1, 2008 to January 1, 2020. A total of 6 measurements were obtained on each pediatric MRI by two residents. The longitudinal axis of the femur was used as a reference trajectory. Three measurements were obtained from medial femoral condyle and lateral femoral condyle, respectively. The lengths measured were from the anteromedial (AM), posteromedial (PM), and distalmedial (DM) on the medial femoral condyle to the physis. On the lateral femoral condyle, the anterolateral (AL), posterolateral (PL), and distallateral (DL) were measured. 2 sample t-testing, chi-square analysis, and Pearson’s correlation coefficient were utilized to find significant associations. Results: Of the 113 pediatric knees identified, 32 did not meet inclusion criteria, yielding 81 patients for final analysis. The average epiphyseal thickness of the bone was 25.5 mm ± 4.2 (CI 24.5 – 26.4, range 14.1 – 33.1). There was a difference in epiphyseal thickness based on screw trajectory () with the average thickness being 23.0, 22.2, 30.5, 29.3, 25.5, and 22.3 mm for the AM, AL, PM, PL, DM and DL respectively. There was a negative correlation between increasing age and epiphyseal cartilage thickness (). There was a positive correlation between increasing age and bone only epiphyseal thickness ().Conclusion:Our study outlines important measurements that orthopedic surgeons must be aware of when operating on OCD lesions in pediatric patients. These results can aid in determining the appropriate screw length that can be implanted into the pediatric distal femur to avoid damaging the distal femur. Our study results show that the optimal screw length to use in pediatric OCD lesion is 18mm in order to avoid physeal disturbance in the majority of pediatric patients.Figure 1.Sagittal MRI of the lateral distal femur showing the projection used to measure the anterolateral (AL), posterolateral(PL), and distallateral (DL) distances to the physisFigure 2.Sagittal MRI of the medial distal femur showing the projection used to measure the anteromedial (AM), posteromedial (PM), and distalmedial (DM) distances to the physis

Age-Related Changes on Mri Of Posterior Cruciate Ligament Morphology in A Skeletally Immature Population

Background:Magnetic resonance imaging (MRI) analysis of the developmental morphology of the anterior cruciate ligament (ACL) in pediatric and adult patients has led to the identification of anatomical risk factors for injury and optimization of surgical reconstruction. Similar work regarding pediatric posterior cruciate ligament (PCL) morphology is limited despite increasing incidence of PCL injury and rising favorability of surgical reconstruction.Purpose:The aim of this study was to describe age-dependent changes in ligamentous and osseous anatomy of the pediatric PCL by comparison of MRI measurements across pre-adolescent and adolescent age groups.Methods:Sixty patients with a knee MRI and wrist radiographs who were seen from 2008 to 2018 at a single tertiary care center were randomly selected from a departmental database and categorized into skeletal age groups (ages 7-9 years, 10-12 years, and 13-15 years) with 20 patients in each group (Table 1). Skeletal age was determined from wrist radiographs and patients with a PCL injury or prior knee surgery were excluded. Twelve dimensions of the PCL substance or its osseous attachment sites were measured on MRI by two independent reviewers. Pairwise comparisons were conducted between the youngest and intermediate, as well as the intermediate and oldest skeletal age groups.Results:There was statistically significant growth across skeletal age groups in femoral physis width (p<0.001), tibial physis width (p<0.001), and medial femoral condyle (MFC) width (p<0.001) and height (p<0.001). Growth in femoral PCL height (p=0.02), femoral PCL footprint to femoral physis height (p=0.004), tibial PCL insertion width (p<0.001), tibial PCL footprint to tibial physis height (p<0.001), and tibial epiphysis height (p=0.003) and width (p<0.001) was only statistically significant between the youngest and intermediate groups. Only MFC height (p=0.03) and width (p=0.001), femoral PCL height (p=0.05), and tibial physis width (p=0.02) were significantly larger in the oldest compared to the intermediate group. Intrasubstance PCL width did not demonstrate significant age-related change (Table 2). Intrarater reliability for all measurements was excellent; interrater reliability was variable.Conclusions:Morphological changes of the pediatric PCL and associated osseous anatomy vary by skeletal age throughout childhood and adolescence. These changes were most pronounced in preadolescents with open physes (ages 7-12 years) compared to adolescents with closing physes (ages 13-15 years), who demonstrated slowed or plateaued growth in most dimensions. These skeletal age-dependent transformations of the PCL and surrounding osseous anatomy may guide surgical planning and inform injury risk prediction and prevention in this active population.Table 1.Cohort characteristics. (N=60)Table 2.MRI measurement summary by skeletal age group.

Quadriceps Tendon Autograft in Pediatric Anterior Cruciate Ligament Reconstruction: Graft Dimensions and Prediction of Size on Preoperative Mri

Background:There has been a growth of interest in quadriceps autograft anterior cruciate ligament (ACL) reconstruction in the pediatric population.Hypothesis/Purpose:The goal of this study was to evaluate children and adolescents who underwent ACL reconstruction using a quadriceps autograft to determine the properties of the harvested graft and to assess the value of demographic, anthropometric, and MRI measurements in predicting the graft size preoperatively.Methods:A retrospective IRB-approved database search was performed from January 2018 through October 2020 for patients undergoing ACL reconstruction. Patients <18 years old at the time of surgery in whom a quadriceps tendon autograft was used were selected. Demographic data and anthropometric measurements were recorded and graft measurements were abstracted from the operative note. Knee magnetic resonance images (MRI) were reviewed to measure the quadriceps tendon thickness on sagittal cuts (Figure 1). Graft length and diameter were then correlated with anthropometric and radiographic data.Results:A total of 169 patients (98 male) were included in the final analysis, with a median age of 15 years (range, 9-17). A tendon length of 65mm or more was harvested in 159 (94%) patients. The mean final graft diameter was 8.4 ± 0.7mm (range, 7 – 11mm) (Table 1). All patients had a graft diameter of 7mm or more, and 139 (82%) had a diameter of 8mm or larger. Preconditioning decreased the graft diameter by a mean of 0.68 ± 0.23mm. Age (P=0.04) and quadriceps thickness on MRI (P<0.001) were significant predictors of the final graft diameter. An MRI sagittal thickness of >6.7mm was 97.4% sensitive for obtaining a graft of ≥8mm in diameter.Conclusion:Our findings suggest that tendon-only quadriceps autograft is a reliable graft source in the pediatric ACL reconstruction, yielding a graft of ≥8mm in diameter in 82% of pediatric patients. Furthermore, preoperative MRI measurements can be reliably used to predict a graft of adequate diameter in children and adolescents undergoing ACL reconstruction, with a sagittal thickness of >6.7mm being highly predictive of a final graft size ≥8mm.Table 1.Univariable and multivariable logistic regression models predicting a final graft diameter of 8mm or more. Significant predictors are indicated in bold. 95%CI: 95% Confidence interval, BMI: body mass index.Figure 1.Preoperative knee sagittal proton density-weighted MRI of a 13-year-old boy with ACL tear. Quadriceps tendon sagittal thickness was measure 1cm above the patella. The three layers of the quadriceps tendon (vastus medialis/lateralis, rectus, intermedius) are also visible in this image.

Can the Disoid Meniscus Remodel After Saucerization?

Background/Purpose:The discoid meniscus is a congenital variant that is reported to be present in 3-5% of the United States population. Treatment involves partial meniscectomy and saucerization to obtain a more anatomic-appearing meniscus. The immediate post-operative cross-sectional morphology is often squared due to the inherent thickness of the discoid meniscus at the resection margin. The purpose of this study is to determine if the discoid meniscus can remodel and regain the triangular cross sectional configuration of a normal meniscus after surgical intervention.Methods:A retrospective chart review was performed to identify all patients undergoing partial meniscectomy for symptomatic discoid meniscus within the past 3 years at our institution. Patients were included if they had a subsequent knee magnetic resonance imaging (MRI) study. Each MRI study was evaluated by two orthopaedic surgeons independently. The extent of remodeling was graded on the cross-sectional morphology of the saucerized discoid meniscus on coronal and sagittal T1-weighted images of the corresponding knee mri based on the following scheme: 1) complete for a triangular cross-section, 2) partial for a trapezoidal cross-section , and 3) none for a square cross-section. Descriptive means and statistical analysis for intraclass correlation coefficient (ICC) were performed using Microsoft Excel.Results:A total of 11 patients and 14 knees were included in this study. Average age at time of surgery was 10.85. Average time to the follow-up MRI of the same knee was 540.50 +/- 341.84 days. 2 knees required additional surgical intervention following their index surgery. Complete remodeling was identified in 78.5% (11/14) knees. 1 knee had no remodeling and 2 knees had partial remodeling. Our ICC was 0.87 (95% CI: [0.66,0.96]).Conclusion:Our results indicate that the discoid meniscus has the potential to remodel into a “normal” triangular cross-sectional morphology after partial meniscectomy and saucerization. In our series, 21.4% (3/14) knees did not experience complete remodeling. This may have been due to over resection of the discoid meniscus at the index surgery owing to the extent of an existing tears. The remodeling all occurred within a two-year time frame though the exact time is unknown. This study indicates no additional intervention is required at the central resection margin to thin out the thick edge that often results from the meniscal biter.

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

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. 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. 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. 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.

Alteration of anterior cruciate ligament orientation in knees with trochlear dysplasia: description of a novel angle on MRI

To assess changes in anterior cruciate ligament (ACL) geometry and inclination in trochlear dysplasia (TD) and analyse their significance. Ninety-nine consecutive knees with TD and 23 normal knee magnetic resonance imaging (MRI) examinations were included as controls (n=122). Varying degrees of TD were classified into four distinct groups (A-D) according to the Dejour classification. MRI images were reviewed independently to measure four ACL angles. Interobserver and intra-observer agreements with statistical significance were determined for TD and various angles. A significant association was found between TD and two measured angles compared with the control group (sagittal ACL and anteromedial ACL angles, p<0.001 for each). The results indicate that TD can predispose to more vertical ACL inclination as measured in the coronal plane on MRI. No association was found with the Blumenstat angle. The present study found significant associations with TD and steeper sagittal ACL, which have been implicated in ACL failure. A novel angle (anteromedial ACL angle) is described which has significant association with TD and is specific for the anteromedial bundle as measured in the coronal plane. Careful consideration of ACL fibre orientation in the coronal plane on MRI is suggested in knees with TD and the use of this newly described angle in assessing ACL reconstruction (ACLR) grafts.

A novel approach to studying early knee osteoarthritis illustrates that bilateral medial tibiofemoral osteoarthritis is a heritable phenotype: an offspring study.

To assess the potential of studying offspring of people with and without knee osteoarthritis to understand the risk factors and heritability for knee osteoarthritis. We selected two groups of Osteoarthritis Initiative (OAI) participants from one clinical site: (1) participants with bilateral radiographic medial tibiofemoral osteoarthritis and (2) those without tibiofemoral osteoarthritis. We then invited biological offspring ≥ 18years old to complete an online survey that inquired about osteoarthritis risk factors and symptoms. Among the survey respondents, we recruited ten offspring of members from each group for a clinic visit with bilateral knee posterior-anterior radiographs and magnetic resonance imaging of the right knee. We established contact with 269/413 (65%) eligible OAI participants. Most (227/269, 84%) had ≥ 1 eligible biological offspring, and 213 (94%) were willing to share information about the new family study with their offspring. Our survey was completed by 188 offspring from 110 OAI participants: mean age of 43.0 (10.4) years, mean body mass index of 23.7 (5.9) kg/m2, 65% female. Offspring obesity (OR = 2.7, 95% CI 1.0-7.3), hypertension (OR = 3.7, 95% CI 1.2-11.3), and Heberden's nodes (OR = 3.6, 95% CI 1.0-13.2) were associated with parental osteoarthritis status; however, adjusted models were not statistically significant. Radiographic tibiofemoral osteoarthritis (16/18 knees vs. 2/20 knees) and meniscal abnormalities (7/9 vs. 2/10 index knees) were more common among offspring with parental osteoarthritis status than not. We established the potential of a novel offspring study design within the OAI, and our results are consistent with bilateral radiographic medial tibiofemoral osteoarthritis being a heritable phenotype of osteoarthritis.

High fidelity deep learning-based MRI reconstruction with instance-wise discriminative feature matching loss.

To improve reconstruction fidelity of fine structures and textures in deep learning- (DL) based reconstructions. A novel patch-based Unsupervised Feature Loss (UFLoss) is proposed and incorporated into the training of DL-based reconstruction frameworks in order to preserve perceptual similarity and high-order statistics. The UFLoss provides instance-level discrimination by mapping similar instances to similar low-dimensional feature vectors and is trained without any human annotation. By adding an additional loss function on the low-dimensional feature space during training, the reconstruction frameworks from under-sampled or corrupted data can reproduce more realistic images that are closer to the original with finer textures, sharper edges, and improved overall image quality. The performance of the proposed UFLoss is demonstrated on unrolled networks for accelerated two- (2D) and three-dimensional (3D) knee MRI reconstruction with retrospective under-sampling. Quantitative metrics including normalized root mean squared error (NRMSE), structural similarity index (SSIM), and our proposed UFLoss were used to evaluate the performance of the proposed method and compare it with others. In vivo experiments indicate that adding the UFLoss encourages sharper edges and more faithful contrasts compared to traditional and learning-based methods with pure loss. More detailed textures can be seen in both 2D and 3D knee MR images. Quantitative results indicate that reconstruction with UFLoss can provide comparable NRMSE and a higher SSIM while achieving a much lower UFLoss value. We present UFLoss, a patch-based unsupervised learned feature loss, which allows the training of DL-based reconstruction to obtain more detailed texture, finer features, and sharper edges with higher overall image quality under DL-based reconstruction frameworks. (Code available at: https://github.com/mikgroup/UFLoss).

DADP: Dynamic abnormality detection and progression for longitudinal knee magnetic resonance images from the Osteoarthritis Initiative.

Osteoarthritis (OA) is the most common disabling joint disease. Magnetic resonance (MR) imaging has been commonly used to assess knee joint degeneration due to its distinct advantage in detecting morphologic cartilage changes. Although several statistical methods over conventional radiography have been developed to perform quantitative cartilage analyses, little work has been done capturing the development and progression of cartilage lesions (or abnormal regions) and how they naturally progress. There are two major challenges, including (i) the lack of building spatial-temporal correspondences and correlations in cartilage thickness and (ii) the spatio-temporal heterogeneity in abnormal regions. The goal of this work is to propose a dynamic abnormality detection and progression (DADP) framework for quantitative cartilage analysis, while addressing the two challenges. First, spatial correspondences are established on flattened 2D cartilage thickness maps extracted from 3D knee MR images both across time within each subject and across all subjects. Second, a dynamic functional mixed effects model (DFMEM) is proposed to quantify abnormality progression across time points and subjects, while accounting for the spatio-temporal heterogeneity. We systematically evaluate our DADP using simulations and real data from the Osteoarthritis Initiative (OAI). Our results show that DADP not only effectively detects subject-specific dynamic abnormal regions, but also provides population-level statistical disease mapping and subgroup analysis.

Improved-Mask R-CNN: Towards an accurate generic MSK MRI instance segmentation platform (data from the Osteoarthritis Initiative).

Objective assessment of osteoarthritis (OA) Magnetic Resonance Imaging (MRI) scans can address the limitations of the current OA assessment approaches. Detecting and extracting bone, cartilage, and joint fluid is a necessary component for the objective assessment of OA, which helps to quantify tissue characteristics such as volume and thickness. Many algorithms, based on Artificial Intelligence (AI), have been proposed over recent years for segmenting bone and soft tissues. Most of these segmentation methods suffer from the class imbalance problem, can't differentiate between the same anatomic structure, or do not support segmenting different rang of tissue sizes. Mask R-CNN is an instance segmentation framework, meaning it segments and distinct each object of interest like different anatomical structures (e.g. bone and cartilage) using a single model. In this study, the Mask R-CNN architecture was deployed to address the need for a segmentation method that is applicable to use for different tissue scales, pathologies, and MRI sequences associated with OA, without having a problem with imbalanced classes. In addition, we modified the Mask R-CNN to improve segmentation accuracy around instance edges. A total of 500 adult knee MRI scans from the publicly available Osteoarthritis Initiative (OAI), and 97 hip MRI scans from adults with symptomatic hip OA, evaluated by two readers, were used for training and validating the network. Three specific modifications to Mask R-CNN yielded the improved-Mask R-CNN (iMaskRCNN): an additional ROIAligned block, an extra decoder block in the segmentation header, and connecting them using a skip connection. The results were evaluated using Hausdorff distance, dice score for bone and cartilage segmentation, and differences in detected volume, dice score, and coefficients of variation (CoV) for effusion segmentation. The iMaskRCNN led to improved bone and cartilage segmentation compared to Mask RCNN as indicated with the increase in dice score from 95% to 98% for the femur, 95-97% for the tibia, 71-80% for the femoral cartilage, and 81-82% for the tibial cartilage. For the effusion detection, the dice score improved with iMaskRCNN 72% versus Mask R-CNN 71%. The CoV values for effusion detection between Reader1 and Mask R-CNN (0.33), Reader1 and iMaskRCNN (0.34), Reader2 and Mask R-CNN (0.22), Reader2 and iMaskRCNN (0.29) are close to CoV between two readers (0.21), indicating a high agreement between the human readers and both Mask R-CNN and iMaskRCNN. Mask R-CNN and iMaskRCNN can reliably and simultaneously extract different scale articular tissues involved in OA, forming the foundation for automated assessment of OA. The iMaskRCNN results show that the modification improved the network performance around the edges.

Impact of a Musculoskeletal "Mini-Residency" Professional Development Program on Knee Magnetic Resonance Imaging Orders by Primary Care Providers.

The US Department of Veterans Affairs has created a portfolio of educational programs to train primary care providers (PCPs) in the evaluation and management of common musculoskeletal (MSK) conditions. Appropriate resource utilization for evaluation of knee pain, including limiting unnecessary magnetic resonance imaging (MRI) studies, is an important theme of these initiatives. The objective of this study was to report the utilization of knee MRI by PCP providers before and after the MSK education program and to determine the appropriateness of these MRI orders. Twenty-six PCPs participated in the MSK Mini-Residency educational program held in Salt Lake City between April 2012 and October 2014. Knee MRI orders submitted by these providers 12 months before and 12 months after their participation were reviewed. Magnetic resonance imaging orders were categorized as "inappropriate," "probably inappropriate," or "possibly appropriate," based on accepted guidelines for knee MRI utilization. Differences in the numbers of precourse and postcourse MRI orders for each of these categories were compared using Student t test. Following our program, MRI orders decreased from 130 (precourse) to 93 (postcourse), a reduction of 28% ( p = 0.04). This reduction was observed entirely within the "inappropriate" and "probably inappropriate" categories; the number of orders categorized as "possibly appropriate" increased, but not significantly. The MSK Mini-Residency training program was a successful educational intervention and was associated with a reduction in inappropriate knee MRI utilization for some participants, while keeping appropriate MRI utilization stable.

Feasibility of Constructing an Automatic Meniscus Injury Detection Model Based on Dual-Mode Magnetic Resonance Imaging (MRI) Radiomics of the Knee Joint.

Objective To explore the feasibility of automatically detecting the degree of meniscus injury by radiomics fusion of dual-mode magnetic resonance imaging (MRI) features of sagittal and coronal planes of the knee joint. Methods This retrospective study included 164 arthroscopically confirmed meniscus injuries in 152 patients admitted to the Department of Orthopaedics of our hospital from July 2018 to March 2021. A total of 1316-dimensional radiomics signatures were extracted from single-mode sagittal and coronal plane images of menisci, respectively. Then, the sagittal and coronal plane features were fused to form a dual-mode joint feature group with a total of 2632-dimensional radiomics signatures. The minimum redundancy maximum relevance (mRMR) algorithm and the least absolute shrinkage and selection operator (LASSO) regression were used to select features and generate optimal radiomics signatures. The single-mode sagittal plane feature model (Model 1), single-mode coronal plane feature model (Model 2), and the combined sagittal and coronal plane feature model (Model 3) performance were tested by receiver operating characteristic (ROC) curves and Delong test. The calibration curve test was used to verify the reliability of radiomics signatures of the three models. Results The average intra- and interobserver intraclass correlation coefficients (ICCs) of the most significant 8-dimensional radiomics signatures of Model 1 and Model 2 were 0.935 (range 0.832-0.998) and 0.928 (range 0.845-0.998), respectively. All the three models had good detection performance; Model 3 had the most significant performance (the areas under the curve (AUCs) of training, and validation sets were 0.947 and 0.923, respectively), which was superior to Model 1 (AUCs of training set and validation set were 0.889 and 0.876, respectively) and Model 2 (AUCs of training set and validation set were 0.831 and 0.851, respectively). The detection probability of training and validation sets in the three models was highly consistent with the actual clinical probability. Conclusions It is feasible to establish a model for automatic detection of meniscus damage by means of radiomics. The detection performance of the dual-mode knee MRI model is better than that of any single-mode model, showing potent feature analysis ability and outstanding detection performance.

EFFECT OF CURCUMA LONGA EXTRACT ON INFLAMMATORY MARKERS AND FACTORS PREDICTING TREATMENT RESPONSE IN KNEE OSTEOARTHRITIS

Purpose: We aimed to determine the effects of Curcuma longa extract (CL) treatment over 12 weeks on serum biomarkers (TNFα, IL6, hsCRP, and MMP3) and synovitis assessed by contrast-enhanced knee magnetic resonance imaging (CE-MRI). We also aimed to describe the factors associated with treatment response.

Medial Meniscal Extrusion Evaluation With Weight-Bearing Ultrasound: Correlation With MR Imaging Findings and Reported Symptoms: Correlation With MR Imaging Findings and Reported Symptoms.

To compare medial meniscal extrusion on weight-bearing ultrasound (US) with supine US and magnetic resonance (MR) imaging correlating with meniscal pathology and reported symptoms. IRB approved study with informed consent. Patients obtaining routine knee MR imaging for suspected knee pathology were prospectively evaluated with supine and weight-bearing US of the medial meniscus. Meniscal extrusion was measured independently by two fellowship-trained musculoskeletal radiologists. Correlation was made to presence or absence of meniscal degeneration or tear on MR imaging, as well as reported symptoms. Statistical significance was calculated via intraclass correlation coefficient (ICC) and analysis of variance (ANOVA). Ninety-nine knees from 95 subjects (50 males, 45 females; mean age 45 ± 15 years) were included. Mean medial meniscal extrusion measured at US for a normal meniscus (n= 36) was 0.8 mm when supine, increasing to 1.6 mm on weight-bearing. Mean meniscal extrusion in subjects with mucoid degeneration (n= 20) and those with meniscal tears (n= 43) was 1.6 mm, increasing to 2.3 mm with weight bearing. Inter-reader reliability showed ICC values of 0.853 to 0.940. There was a significant difference in medial meniscal extrusion comparing subjects with a normal medial meniscus at magnetic resonance imaging (MRI) and subjects with either meniscal degeneration or tear. There was no significant difference in degree of meniscal extrusion between subjects with meniscal degeneration or tear. There was trend of worsening symptoms and increasing functional limitations moving from normal meniscus to meniscal degeneration to meniscal tear. A normal meniscus shows lesser mobility between supine and upright position, than a pathologic meniscus. Both mucoid degeneration and meniscal tear demonstrate extrusion in the supine position, which increases with weight-bearing position.

Patella-patellar tendon angle in relation to the medial patellar plica syndrome, chondromalacia patella, and infrapatellar fat pad syndrome.

The patella-patellar tendon angle (PPTA) assessing the sagittal patellar tilt was reported to be related with anterior knee pain. Herein, clinical effect of PPTA in patients with medial patellar plica (MPP) syndrome, chondromalacia patella, and infrapatellar fat pad (IPFP) syndrome, the most common causes of anterior knee pain, was evaluated. In this retrospective study, 156 patients with anterior knee pain who underwent magnetic resonance imaging (MRI) and arthroscopic surgery that confirmed isolated MPP syndrome, chondromalacia patella, or IPFP syndrome from June 2011 to January 2021 were included in the study group and 118 patients without knee pathology on MRI during the same period were included in the control group. The PPTA was measured on knee MRI and compared between the two groups. A receiver operating characteristic (ROC) analysis was used to evaluate the value of PPTA for predicting the risk of patellofemoral joint disorder. The mean PPTA was significantly smaller in study group (138.1 ± 4.2°) than control group (142.1 ± 4.3°) (p < 0.001). However, there was no significant difference in PPTA among the patients with MPP syndrome, chondromalacia patella, and IPFP syndrome. Furthermore, the ROC analysis revealed that the area under curve, sensitivity, and specificity for predicting the risk of patellofemoral joint disorders were 0.696, 70.3% and 57.6%, respectively, at a PPTA cutoff of 138.3°. Therefore, the smaller PPTA may be associated with MPP syndrome, chondromalacia patella, and IPFP syndrome. Furthermore, PPTA could be a predictive factor for the risk of patellofemoral joint disease in patients with anterior knee pain.

The influence of the morphometric parameters of the intercondylar notch on occurrence of meniscofemoral ligaments.

The purpose of this study was to examine the existence of correlation between the morphometric parameters of the intercondylar notch of the femur and the occurrence of meniscofemoral ligaments (MFLs) and if there is any relationship in the running angle (RA) value between narrowed and normal sized intercondylar notch. Coronal, sagittal and horizontal magnetic resonance (MR) images of 90 patients with specified exclusion criteria were included in this study. The c2 test was used for statistical analysis. In our research either one or both MFLs were identified in 70 (77.8%) of the 90 coronal MR images. In normal sized intercondylar notch, MFLs was seen in 39 (43.3%) cases and on 31 (34.4%) MR images with narrowed intercondylar notch. A significant correlation was established between the occurrence of the MFL and morphometric parameters of the intercondylar notch (p < 0.05). In normal sized intercondylar notch, 12 posterior meniscofemoral ligaments (pMFLs) of type I were detected (RA value 42°), 8 of type II (RA value 33°), 5 of type III (RA value 23°) and two were of indeterminate type, whilst 10 anterior meniscofemoral ligaments (aMFLs) were of type I (RA value 39°), 7 of type II (RA value 31°), 2 of type III (RA value 25°) and the remaining 6 were indeterminate. In narrowed intercondylar notch, 10 ligaments of pMFLs were of type I (RA value 30°), 8 of type II (RA value 25°), 5 of type III (RA value 20°), 10 ligaments of aMFLs were of type I (RA value 35°) and 9 were indeterminate. Statistically significant differences in the value of the running angle of pMFL type I and of type II were evaluated between two groups with different shaped intercondylar notch (p < 0.05). The results shown in our study may be useful in medical clinical practice, reconstructive surgery, interpretation of knee MR images as well as genetic research.

Interscan measurement error in measurements of the meniscus, including the meniscal roots, by fully automatic three-dimensional MRI analysis

The meniscus coverage ratio is an important factor in knee osteoarthritis. The ratio calculated by our software, which automatically segments meniscus and cartilage regions from knee MR images, correlates with cartilage measurements. However, the significance of the changes is questionable due to possible intermeasurement error during sequential observations. The purpose of this study was to show the interscan measurement errors of meniscal measurements, especially the meniscal coverage ratio. Ten healthy volunteers had two 3.0T MRI examinations on the same day. We determined the volume, area, coverage ratio, width, and thickness of the medial (MM) and the lateral (LM) menisci, including their roots, twice per volunteer, using the software with a 3D convolutional neural network algorithm for segmentation. The relative interscan error, expressed as average ± SD, was the absolute value of the difference between the two measurements divided by the average value of all measurements. The relative interscan errors were 1.3 ± 1.1%, 1.7 ± 1.8%, 0.7 ± 0.9%, 1.0 ± 0.9%, 1.9 ± 1.3%, and 0.9 ± 0.9% for MM volume, LM volume, MM area, LM area, MM coverage ratio, and LM coverage ratio, respectively. The relative interscan errors in the middle segment were 0.6 ± 0.3%, 1.3 ± 1.3%, 2.3 ± 1.2%, and 1.6 ± 1.2% for MM width, LM width, MM thickness, and LM thickness, respectively. This study documented the interscan measurement errors in meniscus measurements. These can be indicators to determine the significance of changes observed in longitudinal and interventional studies.