Defining the presence of radiographic knee osteoarthritis: a comparison between the Kellgren and Lawrence system and OARSI atlas criteria

Artificial intelligence in musculoskeletal imaging: realistic clinical applications in the next decade

This article will provide a perspective review of the most extensively investigated deep learning (DL) applications for musculoskeletal disease detection that have the best potential to translate into routine clinical practice over the next decade. Deep learning methods for detecting fractures, estimating pediatric bone age, calculating bone measurements such as lower extremity alignment and Cobb angle, and grading osteoarthritis on radiographs have been shown to have high diagnostic performance with many of these applications now commercially available for use in clinical practice. Many studies have also documented the feasibility of using DL methods for detecting joint pathology and characterizing bone tumors on magnetic resonance imaging (MRI). However, musculoskeletal disease detection on MRI is difficult as it requires multi-task, multi-class detection of complex abnormalities on multiple image slices with different tissue contrasts. The generalizability of DL methods for musculoskeletal disease detection on MRI is also challenging due to fluctuations in image quality caused by the wide variety of scanners and pulse sequences used in routine MRI protocols. The diagnostic performance of current DL methods for musculoskeletal disease detection must be further evaluated in well-designed prospective studies using large image datasets acquired at different institutions with different imaging parameters and imaging hardware before they can be fully implemented in clinical practice. Future studies must also investigate the true clinical benefits of current DL methods and determine whether they could enhance quality, reduce error rates, improve workflow, and decrease radiologist fatigue and burnout with all of this weighed against the costs.

What is New in Osteoarthritis Imaging?

Osteoarthritis (OA) is the leading joint disorder globally, affecting a significant proportion of the population. Recent studies have changed our understanding of OA, viewing it as a complex pathology of the whole joint with a multifaceted etiology, encompassing genetic, biological, and biomechanical elements. This review highlights the role of imaging in diagnosing and monitoring OA. Today's role of radiography is discussed, while also elaborating on the advances in computed tomography and magnetic resonance imaging, discussing semiquantitative methods, quantitative morphologic and compositional techniques, and giving an outlook on the potential role of artificial intelligence in OA research.

ShapeMed-Knee: A Dataset and Neural Shape Model Benchmark for Modeling 3D Femurs

Analyzing anatomic shapes of tissues and organs is pivotal for accurate disease diagnostics and clinical decision-making. One prominent disease that depends on anatomic shape analysis is osteoarthritis, which affects 30 million Americans. To advance osteoarthritis diagnostics and prognostics, we introduce ShapeMed-Knee, a 3D shape dataset with 9,376 high-resolution, medical-imaging-based 3D shapes of both femur bone and cartilage. Besides data, ShapeMed-Knee includes two benchmarks for assessing reconstruction accuracy and five clinical prediction tasks that assess the utility of learned shape representations. Leveraging ShapeMed-Knee, we develop and evaluate a novel hybrid explicit-implicit neural shape model which achieves up to 40% better reconstruction accuracy than a statistical shape model and implicit neural shape model. Our hybrid models achieve state-of-the-art performance for preserving cartilage biomarkers; they're also the first models to successfully predict localized structural features of osteoarthritis, outperforming shape models and convolutional neural networks applied to raw magnetic resonance images and segmentations. The ShapeMed-Knee dataset provides medical evaluations to reconstruct multiple anatomic surfaces and embed meaningful disease-specific information. ShapeMed-Knee reduces barriers to applying 3D modeling in medicine, and our benchmarks highlight that advancements in 3D modeling can enhance the diagnosis and risk stratification for complex diseases. The dataset, code, and benchmarks will be made freely accessible.

Deep learning applications in osteoarthritis imaging

Deep learning (DL) is one of the most exciting new areas in medical imaging. This article will provide a review of current applications of DL in osteoarthritis (OA) imaging, including methods used for cartilage lesion detection, OA diagnosis, cartilage segmentation, and OA risk assessment. DL techniques have been shown to have similar diagnostic performance as human readers for detecting and grading cartilage lesions within the knee on MRI. A variety of DL methods have been developed for detecting and grading the severity of knee OA and various features of knee OA on X-rays using standardized classification systems with diagnostic performance similar to human readers. Multiple DL approaches have been described for fully automated segmentation of cartilage and other knee tissues and have achieved higher segmentation accuracy than currently used methods with substantial reductions in segmentation times. Various DL models analyzing baseline X-rays and MRI have been developed for OA risk assessment. These models have shown high diagnostic performance for predicting a wide variety of OA outcomes, including the incidence and progression of radiographic knee OA, the presence and progression of knee pain, and future total knee replacement. The preliminary results of DL applications in OA imaging have been encouraging. However, many DL techniques require further technical refinement to maximize diagnostic performance. Furthermore, the generalizability of DL approaches needs to be further investigated in prospective studies using large image datasets acquired at different institutions with different imaging hardware before they can be implemented in clinical practice and research studies.

Charting Aging Trajectories of Knee Cartilage Thickness for Early Osteoarthritis Risk Prediction: An MRI Study from the Osteoarthritis Initiative Cohort

Knee osteoarthritis (OA), a prevalent joint disease in the U.S., poses challenges in terms of predicting of its early progression. Although high-resolution knee magnetic resonance imaging (MRI) facilitates more precise OA diagnosis, the heterogeneous and multifactorial aspects of OA pathology remain significant obstacles for prognosis. MRI-based scoring systems, while standardizing OA assessment, are both time-consuming and labor-intensive. Current AI technologies facilitate knee OA risk scoring and progression prediction, but these often focus on the symptomatic phase of OA, bypassing initial-stage OA prediction. Moreover, their reliance on complex algorithms can hinder clinical interpretation. To this end, we make this effort to construct a computationally efficient, easily-interpretable, and state-of-the-art approach aiding in the radiographic OA (rOA) auto-classification and prediction of the incidence and progression, by contrasting an individual's cartilage thickness with a similar demographic in the rOA-free cohort. To better visualize, we have developed the toolset for both prediction and local visualization. A movie demonstrating different subtypes of dynamic changes in local centile scores during rOA progression is available at https://tli3.github.io/KneeOA/. Specifically, we constructed age-BMI-dependent reference charts for knee OA cartilage thickness, based on MRI scans from 957 radiographic OA (rOA)-free individuals from the Osteoarthritis Initiative cohort. Then we extracted local and global centiles by contrasting an individual's cartilage thickness to the rOA-free cohort with a similar age and BMI. Using traditional boosting approaches with our centile-based features, we obtain rOA classification of KLG ≤ 1 versus KLG = 2 (AUC = 0.95, F1 = 0.89), KLG ≤ 1 versus KLG ≥ 2 (AUC = 0.90, F1 = 0.82) and prediction of KLG2 progression (AUC = 0.98, F1 = 0.94), rOA incidence (KLG increasing from < 2 to ≥ 2; AUC = 0.81, F1 = 0.69) and rOA initial transition (KLG from 0 to 1; AUC = 0.64, F1 = 0.65) within a future 48-month period. Such performance in classifying KLG ≥ 2 matches that of deep learning methods in recent literature. Furthermore, its clinical interpretation suggests that cartilage changes, such as thickening in lateral femoral and anterior femoral regions and thinning in lateral tibial regions, may serve as indicators for prediction of rOA incidence and early progression. Meanwhile, cartilage thickening in the posterior medial and posterior lateral femoral regions, coupled with a reduction in the central medial femoral region, may signify initial phases of rOA transition.

An Update on Clinical Utility of Musculoskeletal Ultrasonography in Knee Osteoarthritis

In knee osteoarthritis (KOA), timely and accurate assessment of the severity is essential to help orthopedic surgeons determine the most appropriate therapeutic strategies and evaluate disease outcomes and responses for corresponding treatments. In KOA, musculoskeletal ultrasonography (MSUS) could effectively help detect various abnormalities, including synovitis, osteophytes, and cartilage damage. Further, MSUS could be used to monitor the response to different therapies in KOA, to guide local diagnostic and therapeutic procedures. In the future, applications based on continuously evolving US tools could enhance the clinical utility of MSUS in KOA.

Machine learning as an adjunct to expert observation in classification of radiographic knee osteoarthritis: findings from the Hertfordshire Cohort Study

BACKGROUND

Osteoarthritis is the most prevalent type of arthritis. Many approaches exist for characterising radiographic knee OA, including machine learning (ML).

AIMS

To examine Kellgren and Lawrence (K&L) scores from ML and expert observation, minimum joint space and osteophyte in relation to pain and function.

METHODS

Participants from the Hertfordshire Cohort Study, comprising individuals born in Hertfordshire from 1931 to 1939, were analysed. Radiographs were assessed by clinicians and ML (convolutional neural networks) for K&L scoring. Medial minimum joint space and osteophyte area were ascertained using the knee OA computer-aided diagnosis (KOACAD) program. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) was administered. Receiver operating characteristic analysis was implemented for minimum joint space, osteophyte, and observer- and ML-derived K&L scores in relation to pain (WOMAC pain score > 0) and impaired function (WOMAC function score > 0).

RESULTS

359 participants (aged 71-80) were analysed. Among both sexes, discriminative capacity regarding pain and function was fairly high for observer-derived K&L scores [area under curve (AUC): 0.65 (95% CI 0.57, 0.72) to 0.70 (0.63, 0.77)]; results were similar among women for ML-derived K&L scores. Discriminative capacity was moderate among men for minimum joint space in relation to pain [0.60 (0.51, 0.67)] and function [0.62 (0.54, 0.69)]. AUC < 0.60 for other sex-specific associations.

DISCUSSION

Observer-derived K&L scores had higher discriminative capacity regarding pain and function compared to minimum joint space and osteophyte. Among women, discriminative capacity was similar for observer- and ML-derived K&L scores.

CONCLUSION

ML as an adjunct to expert observation for K&L scoring may be beneficial due to the efficiency and objectivity of ML.

Artificial-Intelligence-Aided Radiographic Diagnostic of Knee Osteoarthritis Leads to a Higher Association of Clinical Findings with Diagnostic Ratings

BACKGROUND

Radiographic knee osteoarthritis (OA) severity and clinical severity are often dissociated. Artificial intelligence (AI) aid was shown to increase inter-rater reliability in radiographic OA diagnosis. Thus, AI-aided radiographic diagnoses were compared against AI-unaided diagnoses with regard to their correlations with clinical severity.

METHODS

Seventy-one DICOMs (m/f = 27:42, mean age: 27.86 ± 6.5) (X-ray format) were used for AI analysis (KOALA software, IB Lab GmbH). Subjects were recruited from a physiotherapy trial (MLKOA). At baseline, each subject received (i) a knee X-ray and (ii) an assessment of five main scores (Tegner Scale (TAS); Knee Injury and Osteoarthritis Outcome Score (KOOS); International Physical Activity Questionnaire; Star Excursion Balance Test; Six-Minute Walk Test). Clinical assessments were repeated three times (weeks 6, 12 and 24). Three physicians analyzed the presented X-rays both with and without AI via KL grading. Analyses of the (i) inter-rater reliability (IRR) and (ii) Spearman's Correlation Test for the overall KL score for each individual rater with clinical score were performed.

RESULTS

We found that AI-aided diagnostic ratings had a higher association with the overall KL score and the KOOS. The amount of improvement due to AI depended on the individual rater.

CONCLUSION

AI-guided systems can improve the ratings of knee radiographs and show a stronger association with clinical severity. These results were shown to be influenced by individual readers. Thus, AI training amongst physicians might need to be increased. KL might be insufficient as a single tool for knee OA diagnosis.

Osteoarthritis today: Lost in translation?

Osteoarthritis (OA) is a highly prevalent joint disease that is associated with pain, loss of function, and high direct and indirect economic costs. The current therapeutic options are inadequate, providing only a moderate symptom relief without the possibility of disease modification. While treatment options and personalized medicines are increasing for many complex diseases, OA drug development has been impeded by the advanced state of disease at the time of diagnosis and intervention, heterogeneity in both symptoms and rates of progression, and a lack of validated biomarkers and relevant outcome measures. This review article summarizes the OA landscape, including therapies in development as potential OA treatments, potential biomarkers undergoing evaluation by the US Food and Drug Administration, and a summary of current OA treatment guidelines, with a particular focus on the knee OA.

A fully automatic target detection and quantification strategy based on object detection convolutional neural network YOLOv3 for one-step X-ray image grading

Methods for automatic image analysis are demanded for dealing with the explosively increased imaging data in clinics. Osteoarthritis (OA) is a typical disease diagnosed based on X-ray imaging. Herein, we propose a novel modeling strategy based on YOLO version 3 (YOLOv3) for automatic simultaneous localization of knee joints and quantification of radiographic knee OA. As an advanced deep convolutional neural network (CNN) algorithm for target detection, YOLOv3 enables simultaneous small object detection and quantification due to its unique residual connection and feature map merging. Hence, a unified CNN model is built for the elegant integration of knee joint detection and corresponding OA severity grading using the YOLOv3 framework. We achieve desirable accuracy in knee OA grading using the public and clinical datasets. It provides improvements in the precision, recall, F₁ score and diagnostic accuracy of knee OA as well. Because of the fully automatic target detection and quantification, the time of handling an image is merely 40 ms from inputting the image to getting its label, supporting quick clinic decisions. It, thus, affords convenient and efficient image analysis for daily clinical diagnosis.

Early knee OA definition-what do we know at this stage? An imaging perspective

While criteria for early-stage knee osteoarthritis (OA) in a primary care setting have been proposed, the role of imaging has been limited to radiography using the standard Kellgren-Lawrence classification. Standardized imaging and interpretation are critical with radiographs, yet studies have also shown that even early stages of radiographic OA already demonstrate advanced damage to knee joint tissues such as cartilage, menisci, and bone marrow. Morphological magnetic resonance imaging (MRI) shows degenerative damage earlier than radiographs and definitions for OA using MRI have been published though no accepted definition of early OA based on MRI is currently available. The clinical significance of structural abnormalities has also not been well defined, and the differentiation between normal aging and structural OA development remains a challenge. Compositional MRI of cartilage provides information on biochemical, degenerative changes within the cartilage matrix before cartilage defects occur and when cartilage damage is potentially reversible. Studies have shown that cartilage composition can predict cartilage loss and radiographic OA. However, while this technology is most promising for characterizing early OA it has currently limited clinical application. Better standardization of compositional MRI is required, which is currently work in progress. Finally, there has been renewed interest in computed tomography (CT) for assessing early knee OA as new techniques such as weight bearing and spectral CT are available, which may provide information on joint loading, cartilage, and bone and potentially have a role in better characterizing early OA. In conclusion, while imaging may have a limited role in diagnosing early OA in a primary care setting, there are advanced imaging technologies available, which detect early degeneration and may thus significantly alter management as new therapeutic modalities evolve.

Artificial intelligence-based computer-aided system for knee osteoarthritis assessment increases experienced orthopaedic surgeons' agreement rate and accuracy

PURPOSE

The aims of this study were to (1) analyze the impact of an artificial intelligence (AI)-based computer system on the accuracy and agreement rate of board-certified orthopaedic surgeons (= senior readers) to detect X-ray features indicative of knee OA in comparison to unaided assessment and (2) compare the results to those of senior residents (= junior readers).

METHODS

One hundred and twenty-four unilateral knee X-rays from the OAI study were analyzed regarding Kellgren-Lawrence grade, joint space narrowing (JSN), sclerosis and osteophyte OARSI grade by computerized methods. Images were rated for these parameters by three senior readers using two modalities: plain X-ray (unaided) and X-ray presented alongside reports from a computer-assisted detection system (aided). After exclusion of nine images with incomplete annotation, intraclass correlations between readers were calculated for both modalities among 115 images, and reader performance was compared to ground truth (OAI consensus). Accuracy, sensitivity and specificity were also calculated and the results were compared to those from a previous study on junior readers.

RESULTS

With the aided modality, senior reader agreement rates for KL grade (2.0-fold), sclerosis (1.42-fold), JSN (1.37-fold) and osteophyte OARSI grades (3.33-fold) improved significantly. Reader specificity and accuracy increased significantly for all features when using the aided modality compared to the gold standard. On the other hand, sensitivity only increased for OA diagnosis, whereas it decreased (without statistical significance) for all other features. With aided analysis, senior readers reached similar agreement and accuracy rates as junior readers, with both surpassing AI performance.

CONCLUSION

The introduction of AI-based computer-aided assessment systems can increase the agreement rate and overall accuracy for knee OA diagnosis among board-certified orthopaedic surgeons. Thus, use of this software may improve the standard of care for knee OA detection and diagnosis in the future.

LEVEL OF EVIDENCE

Level II.

Automatic assessment of knee osteoarthritis severity in portable devices based on deep learning

BACKGROUND

For knee osteoarthritis, the commonly used radiology severity criteria Kellgren-Lawrence lead to variability among surgeons. Most existing diagnosis models require preprocessed radiographs and specific equipment.

METHODS

All enrolled patients diagnosed with KOA who met the criteria were obtained from **** Hospital. This study included 2579 images shot from posterior-anterior X-rays of 2,378 patients. We used RefineDet to train and validate this deep learning-based diagnostic model. After developing the model, 823 images of 697 patients were enrolled as the test set. The whole test set was assessed by up to 5 surgeons and this diagnostic model. To evaluate the model's performance we compared the results of the model with the KOA severity diagnoses of surgeons based on K-L scales.

RESULTS

Compared to the diagnoses of surgeons, the model achieved an overall accuracy of 0.977. Its sensitivity (recall) for K-L 0 to 4 was 1.0, 0.972, 0.979, 0.983 and 0.989, respectively; for these diagnoses, the specificity of this model was 0.992, 0.997, 0.994, 0.991 and 0.995. The precision and F1-score were 0.5 and 0.667 for K-L 0, 0.914 and 0.930 for K-L 1, 0.978 and 0.971 for K-L 2, 0.981 and 0.974 for K-L 3, and 0.988 and 0.985 for K-L 4, respectively. All K-L scales perform AUC > 0.90. The quadratic weighted Kappa coefficient between the diagnostic model and surgeons was 0.815 (P < 0.01, 95% CI 0.727-0.903). The performance of the model is comparable to the clinical diagnosis of KOA. This model improved the efficiency and avoided cumbersome image preprocessing.

CONCLUSION

The deep learning-based diagnostic model can be used to assess the severity of KOA in portable devices according to the Kellgren-Lawrence scale. On the premise of improving diagnostic efficiency, the results are highly reliable and reproducible.

An Automatic Method for Assessing Spiking of Tibial Tubercles Associated with Knee Osteoarthritis

Efficient and scalable early diagnostic methods for knee osteoarthritis are desired due to the disease’s prevalence. The current automatic methods for detecting osteoarthritis using plain radiographs struggle to identify the subjects with early-stage disease. Tibial spiking has been hypothesized as a feature of early knee osteoarthritis. Previous research has demonstrated an association between knee osteoarthritis and tibial spiking, but the connection to the early-stage disease has not been investigated. We study tibial spiking as a feature of early knee osteoarthritis. Additionally, we develop a deep learning based model for detecting tibial spiking from plain radiographs. We collected and graded 913 knee radiographs for tibial spiking. We conducted two experiments: experiments A and B. In experiment A, we compared the subjects with and without tibial spiking using Mann-Whitney U-test. Experiment B consisted of developing and validating an interpretative deep learning based method for predicting tibial spiking. The subjects with tibial spiking had more severe Kellgren-Lawrence grade, medial joint space narrowing, and osteophyte score in the lateral tibial compartment. The developed method achieved an accuracy of 0.869. We find tibial spiking a promising feature in knee osteoarthritis diagnosis. Furthermore, the detection can be automatized.

Conventional MRI-derived subchondral trabecular biomarkers and their association with knee cartilage volume loss as early as 1 year: a longitudinal analysis from Osteoarthritis Initiative

OBJECTIVE

To study associations between MRI-derived subchondral trabecular biomarkers obtained from conventional MRI sequences and knee cartilage loss over 12 and 24 months, using the FNIH osteoarthritis (OA) biomarkers consortium.

MATERIALS AND METHODS

Data of the 600 subjects in the FNIH OA biomarkers consortium (a nested case-control study within Osteoarthritis Initiative [OAI]) were extracted from the online database. Baseline knee MRI (intermediate-weighted (IW) sequences) were evaluated to determine conventional MRI-derived trabecular thickness (cTbTh) and bone-to-total ratio (cBV/TV). The measurements for medial and lateral volumes of cartilages using baseline, 12-, and 24-month knee MRI were extracted from the OAI database, and cartilage volume loss over 12 and 24 months of follow-up were determined using Relative Change Index. The association between conventional MRI-based subchondral trabecular biomarkers and cartilage volume loss were studied using logistic regression models, adjusted for relevant confounders including age, sex, body mass index (BMI), vitamin D use, Kellgren Lawrence grade (KLG), and tibiofemoral alignment.

RESULTS

Higher medial cTbTh and cBV/TV at baseline were associated with increased odds of medial tibial cartilage volume loss over 12 months (ORs: 1.01 [1.00-1.02] and 1.24 [1.10-1.39] per 1-SD change) and 24 months (ORs: 1.01 [1.00-1.02] and 1.22 [1.08-1.37], per 1-SD change). No significant association was observed between medial subchondral trabecular biomarkers and lateral tibial or femoral (medial or lateral) cartilage volume loss over the first and second follow-up years.

CONCLUSIONS

Conventional MRI-derived subchondral trabecular biomarkers (higher medial cTbTh and cBV/TV) may be associated with increased medial tibial cartilage volume loss as early as 1 year.

Knee osteoarthritis severity grading using vision transformer

Knee osteoarthritis severity grading from plain radiographs is of great significance in the diagnosis of osteoarthritis (OA). Recently, deep learning had a great impact on improving the Kellgren and Lawrence (KL) grading scheme of Knee osteoarthritis KOA using models that acquire the contextual features spontaneously without the need for any conventional high computational spatial configuration modeling. In this study, we apply the state-of-art Vision Transformer (ViT) for the KL grading of Knee Osteoarthritis and show that a simple transfer learning approach of such model can lead to better results than those achieved by other complex architectures over less number of training data. The study concludes that such a pre-trained ViT, fine-tuned on OAI dataset yield to promising results in KL grading KOA, in which these results are in line with the state-of-art studies.

Colchicine versus Physical Therapy in Knee Osteoarthritis

Background: The treatment of osteoarthritis remains a major challenge due to the unavailability of a disease-modifying medication and the limitations of current therapeutic perspectives, which mainly target the symptoms, not the disease itself. The purpose of our study is to compare the efficacy of colchicine treatment versus physical therapy. Methods: The study included 62 patients diagnosed with knee osteoarthritis (KOA) according to the American College of Rheumatology (ACR) criteria, hospitalized within the time frame of October 2020−March 2022 in the Department of Rehabilitation Medicine and Physical Therapy of the Emergency Clinical County Hospital of Craiova. The participants were randomly divided into two groups. The observation period was 16 weeks long. The first group (31 patients) received colchicine at a dosage of 1 mg/day together with analgesics (acetaminophen < 2 g/day), while the second group (31 patients) received analgesics (acetaminophen < 2 g/day) together with a 16-week plan of physiotherapy. Results: Group II, in which patients underwent physical therapy, demonstrated a statistically significant decrease in both left (p < 0.001) and right (p = 0.012) knee VAS and WOMAC (p = 0.038) scores at 16 weeks, compared to the group treated with colchicine. Regarding the MSUS examination at 16 weeks, there were no significant changes in the structural abnormalities and no improvement in cartilage aspect or thickness. Higher BMI was associated with higher WOMAC score (p = 0.012), but not with higher VAS score (p = 0.062). Cholesterol and triglyceride levels were associated with high WOMAC (p < 0.001; p = 0.021) and high VAS (p = 0.023; p < 0.001) scores. Conclusions: Our study monitored VAS and WOMAC scores in two groups of patients with KOA, showing that physical therapy is more effective than colchicine in reducing symptoms. We found no statistically significant difference in musculoskeletal ultrasound (MSUS) feature improvement during the 16-week study.

Corresponding Changes of Autophagy-Related Genes and Proteins in Different Stages of Knee Osteoarthritis: An Animal Model Study

Objective

To investigate the effect of autophagy expression levels of different weight‐bearing states and different stages of osteoarthritis in animal models, as well as the corresponding mechanisms.

Methods

We used the male Sprague–Dawley (SD) rats (12‐week‐old, SPF) to establish the OA animal models by modified Hulth method, and grouped animal models according to the length of time after surgery and different weight‐bearing areas. RT‐qPCR was carried out for detection of autophagy‐related genes such as Atg7, Atg12, P62, etc. Western blot analysis was used to detect the expression levels of corresponding autophagy‐related proteins such as LC3B, P62, etc. T test was performed for statistical analysis to compare different groups, while the differences were deemed statistically significant with P < 0.05. Transmission electron microscopy was used to observe the autophagosome to demonstrate the level of autophagy expression and the status of the chondrocytes.

Results

The results of the RT‐qPCR testing showed that when the weight‐bearing cartilage of the 4‐week group (relatively mild) was compared with that of the 10‐week group (relatively severe), there were statistically significant differences in all the genes tested, in detail: Atg3 (P < 0.01), Atg7 (P < 0.01), Atg12 (P < 0.01), P62 (P < 0.0001). The expression of autophagy‐related mRNA in the 4‐week group is increased compared with that of the 10‐week group. As for the expression of proteins, Western blotting showed that in the comparison between the 4‐ and the 10‐week groups, statistically significant results include Atg12 (P < 0.01) in the non‐weight‐bearing area, with decreased autophagy in the 10‐week group compared with that of the 4‐week group, while expression of LC3B (P < 0.05) protein was significantly higher in the 4‐week group than in the control in the non‐weight‐bearing area. The expression of LC3B (P < 0.0001) and P62 (P < 0.05) in the 10‐week group were higher than that of the control. Transmission electron microscope showed that autophagy in the weight‐bearing area is stronger than that in the non‐weight‐bearing area, and autophagy in the 4‐week group is stronger than in the 10‐week group for the weight‐bearing area.

Conclusions

The expression of autophagy varies during different stages of osteoarthritis, in which the autophagy is stronger in the early stage of osteoarthritis, and gradually decreases with the progression of the disease. Autophagy in different weight‐bearing areas may also be different.

An update on the knee osteoarthritis severity grading using wide residual learning

BACKGROUND

Knee Osteoarthritis (KOA) is the most common type of Osteoarthritis (OA) and it is diagnosed by physicians using a standard 0 -4 Kellgren Lawrence (KL) grading system which sets the KOA on a spectrum of 5 grades; starting from normal (0) to Severe OA (4).

OBJECTIVES

In this paper, we propose a transfer learning approach of a very deep wide residual learning-based network (WRN-50-2) which is fine-tuned using X-ray plain radiographs from the Osteoarthritis Initiative (OAI) dataset to learn the KL severity grading of KOA.

METHODS

We propose a data augmentation approach of OAI data to avoid data imbalance and reduce overfitting by applying it only to certain KL grades depending on their number of plain radiographs. Then we conduct experiments to test the model based on an independent testing data of original plain radiographs acquired from the OAI dataset.

RESULTS

Experimental results showed good generalization power in predicting the KL grade of knee X-rays with an accuracy of 72% and Precision 74%. Moreover, using Grad-Cam, we also observed that network selected some distinctive features that describe the prediction of a KL grade of a knee radiograph.

CONCLUSION

This study demonstrates that our proposed new model outperforms several other related works, and it can be further improved to be used to help radiologists make more accurate and precise diagnosis of KOA in future clinical practice.

Multicentre Study Using Machine Learning Methods in Clinical Diagnosis of Knee Osteoarthritis

Knee osteoarthritis (OA) is one of the most common musculoskeletal disorders. OA diagnosis is currently conducted by assessing symptoms and evaluating plain radiographs, but this process suffers from the subjectivity of doctors. In this study, we retrospectively compared five commonly used machine learning methods, especially the CNN network, to predict the real-world X-ray imaging data of knee joints from two different hospitals using Kellgren-Lawrence (K-L) grade of knee OA to help doctors choose proper auxiliary tools. Furthermore, we present attention maps of CNN to highlight the radiological features affecting the network decision. Such information makes the decision process transparent for practitioners, which builds better trust towards such automatic methods and, moreover, reduces the workload of clinicians, especially for remote areas without enough medical staff.

Anterior Cruciate Ligament Reconstruction versus Nonoperative Treatment: Better Function and Less Secondary Meniscectomies But No Difference in Knee Osteoarthritis-A Meta-Analysis

OBJECTIVES

The impact of anterior cruciate ligament (ACL) reconstruction on knee osteoarthritis (OA) is still unclear. The aim of the current meta-analysis was to compare surgical treatment versus nonoperative management of ACL tears to assess the impact of these approaches on knee OA development at a 5 and 10 years of follow-up.

DESIGN

A meta-analysis was performed after a systematic literature search (May 2021) was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines. Both randomized and nonrandomized comparative studies with more than 5 years of follow-up were selected. Influence of the treatment was assessed in terms of knee OA development, subjective and objective clinical results, activity level, and risk of further surgeries. Risk of bias and quality of evidence were assessed following the Cochrane guidelines.

RESULTS

Twelve studies matched the inclusion criteria, for a total of 1,004 patients. Level of evidence was rated low to very low. No difference was documented in terms of knee OA development, Tegner score, subjective International Knee Documentation Committee (IKDC), and Lysholm scores. A significant difference favoring the surgical treatment in comparison with a nonsurgical approach was observed in terms of objective IKDC score (P = 0.03) and risk of secondary meniscectomy (P < 0.0001). The level of evidence was considered very low for subjective IKDC, low for knee OA development, objective IKDC, number of secondary meniscectomies, and Lysholm score, and moderate for post-op Tegner score.

CONCLUSIONS

The meta-analysis did not support an advantage of ACL reconstruction in terms of OA prevention in comparison with a nonoperative treatment. Moreover, no differences were reported for subjective results and activity level at 5 and 10 years of follow-up. On the contrary, patients who underwent surgical treatment of their ACL tear presented important clinical findings in terms of better objective knee function and a lower rate of secondary meniscectomies when compared with conservatively managed patents.Protocol Registration: CRD420191156483 (PROSPERO).

Automated Knee Osteoarthritis Assessment Increases Physicians' Agreement Rate and Accuracy: Data from the Osteoarthritis Initiative

Objective. To assess the impact of a computerized system on physicians' accuracy and agreement rate, as compared with unaided diagnosis. Methods. A set of 124 unilateral knee radiographs from the Osteoarthritis Initiative (OAI) study were analyzed by a computerized method with regard to Kellgren-Lawrence (KL) grade, as well as joint space narrowing, osteophytes, and sclerosis Osteoarthritis Research Society International (OARSI) grades. Physicians scored all images, with regard to osteophytes, sclerosis, joint space narrowing OARSI grades and KL grade, in 2 modalities: through a plain radiograph (unaided) and a radiograph presented together with the report from the computer assisted detection system (aided). Intraclass correlation between the physicians was calculated for both modalities. Furthermore, physicians' performance was compared with the grading of the OAI study, and accuracy, sensitivity, and specificity were calculated in both modalities for each of the scored features. Results. Agreement rates for KL grade, sclerosis, and osteophyte OARSI grades, were statistically increased in the aided versus the unaided modality. Readings for joint space narrowing OARSI grade did not show a statistically difference between the 2 modalities. Readers' accuracy and specificity for KL grade >0, KL >1, sclerosis OARSI grade >0, and osteophyte OARSI grade >0 was significantly increased in the aided modality. Reader sensitivity was high in both modalities. Conclusions. These results show that the use of an automated knee OA software increases consistency between physicians when grading radiographic features of OA. The use of the software also increased accuracy measures as compared with the OAI study, mostly through increases in specificity.

Differential Contribution of the Medial and the Lateral Side of the Joint to Symptoms in Knee Osteoarthritis: A Radiographic and Laboratory Analysis in the Nagahama Study

OBJECTIVE

This cross-sectional study aimed to explore the differences of the medial and lateral sides of the knee joint and precise radiographic abnormalities in contribution to the knee pain and clinical outcomes.

DESIGN

Participants 60 years or older who underwent radiographic evaluation were included. Knee radiography was assessed using grading systems of the Osteoarthritis Research Society International (OARSI) atlas. The Japanese Knee Osteoarthritis Measure (JKOM) was evaluated as clinical outcomes. Serum high-sensitivity C-reactive protein (hsCRP) was used to evaluate systemic inflammation. We divided the participants into normal, medial-, lateral-, and medial & lateral-OA types and compared their JKOM using an analysis of covariance. Furthermore, we analyzed the relationship between the knee pain and stiffness of JKOM and the grading of each radiographic feature using a multiple regression model.

RESULTS

Lateral- and medial & lateral-OA groups had a significantly worse symptoms in the total and the pain score, especially in movement subscales, in JKOM score. Lateral-OA groups had higher hsCRP than medial-OA group. Multivariate analysis showed that medial joint space narrowing (JSN), and lateral femoral and tibial osteophytes significantly affected knee pain (adjusted odds ratios: 1.73, 1.28, and 1.55, respectively). The radiographic changes are associated with pain more in JSN in the medial side and osteophytes in the lateral side.

CONCLUSION

Lateral- and medial & lateral-OA groups showed worth symptom. In addition, medial JSN and lateral osteophytes have potent effects on the knee pain.

Should return to pivoting sport be avoided for the secondary prevention of osteoarthritis after anterior cruciate ligament reconstruction? A prospective cohort study with MRI, radiographic and symptomatic outcomes

OBJECTIVE

To evaluate if returning to pivoting sport following anterior cruciate ligament reconstruction (ACLR) is associated with longitudinal structural and symptomatic osteoarthritis outcomes.

DESIGN

Eighty-one adults aged 18-50 years were followed prospectively 1- to 5-years post-ACLR. Return to pivoting sport was assessed at 1-, 3- and 5-years. Longitudinal changes in osteoarthritis features were evaluated from 1- and 5-year magnetic resonance imaging (MRI)s using MRI Osteoarthritis Knee Score (MOAKS). Radiographic osteoarthritis and self-reported knee symptoms, function and quality of life were assessed using the Osteoarthritis Research Society International (OARSI) atlas and Knee injury Osteoarthritis Outcome Score (KOOS), respectively, at 5 years post-ACLR. Generalised linear models (adjusted for baseline characteristics) assessed whether returning to pivoting sport was associated with risk of worsening osteoarthritis features on MRI, radiographic osteoarthritis and KOOS.

RESULTS

Thirty participants returned to pivoting sport 1-year post-ACLR and 50 returned at any time (i.e., 1-, 3- or 5-years). Returning to pivoting sport was not associated with worsening of any MRI osteoarthritis feature (risk ratio (RR) range: 0.59-2.91) or 5-year KOOS (β range: -2.73-3.69). Returning to pivoting sport at 1-year and up to 5-years post-ACLR was associated with a 50% (RR 0.49, 95%CI 0.10-2.37) and 40% (RR 0.60, 95%CI 0.16-2.17) reduced risk of radiographic osteoarthritis, respectively, but these risk reductions were inconclusive due to wide confidence intervals.

CONCLUSION

After ACLR, returning to pivoting sport was not associated with increased risk of worsening knee osteoarthritis features on MRI, radiographic osteoarthritis or knee symptoms. Participation in pivoting sport need not be avoided as part of osteoarthritis secondary prevention strategies.

West Lake staging: A new staging system orchestrated by X-ray and MRI on knee osteoarthritis

Purpose: To investigate the differences on X-ray and MRI among each stage of knee osteoarthritis (KOA) and further propose a new staging system called West Lake (WL) staging. Methods: A cross-sectional study was conducted on patients with KOA. Stage I, II, III, and IV were divided based on stepwise treatment strategy of Knee osteoarthritis (KOA). Joint space widths (JSW) were measured on X-rays, whereas cartilage injuries (CI) and bone marrow lesions (BML) were evaluated on MRI. The differences of them across the groups were calculated by T-test. Receiver operating characteristic (ROC) curves were rendered to obtain the areas under the curves (AUC), Youden index and corresponding cut-off points. Results: Eventually, there were significant differences on JSW, CI, and BML between stage II/III and III/IV, while no significant differences between stage I/II. In stage II/III, the AUC of JSW, CI, BML was 0.99, 0.76, 0.71 and the Youden index was 0.94, 0.38, 0.45, meanwhile the cut-off points were ≤5.1 mm, >1, >2. In stage III/IV, the AUC of JSW, CI, BML was 0.96, 0.79, 0.74 and the Youden index was 0.84, 0.58, 0.38, meanwhile the cut-off points were ≤3.2 mm, >3, >4. Conclusion: The WL staging was described as follows: Stage I, X-ray shows no joint space narrow, normal MRI or MRI shows cartilage degeneration and only 1 or 2 sections are involved in BML. Stage II, X-ray shows joint space narrow, MRI shows cartilage defect but no full-thickness cartilage defect, meanwhile 3 or 4 sections are involved in BML. Stage III, X-ray shows serious joint space narrow even JSW disappeared, MRI shows full-thickness cartilage defect, more than 4 sections are involved in BML.

Towards novel osteoarthritis biomarkers: Multi-criteria evaluation of 46,996 segmented knee MRI data from the Osteoarthritis Initiative

Convolutional neural networks (CNNs) are the state-of-the-art for automated assessment of knee osteoarthritis (KOA) from medical image data. However, these methods lack interpretability, mainly focus on image texture, and cannot completely grasp the analyzed anatomies' shapes. In this study we assess the informative value of quantitative features derived from segmentations in order to assess their potential as an alternative or extension to CNN-based approaches regarding multiple aspects of KOA. Six anatomical structures around the knee (femoral and tibial bones, femoral and tibial cartilages, and both menisci) are segmented in 46,996 MRI scans. Based on these segmentations, quantitative features are computed, i.e., measurements such as cartilage volume, meniscal extrusion and tibial coverage, as well as geometric features based on a statistical shape encoding of the anatomies. The feature quality is assessed by investigating their association to the Kellgren-Lawrence grade (KLG), joint space narrowing (JSN), incident KOA, and total knee replacement (TKR). Using gold standard labels from the Osteoarthritis Initiative database the balanced accuracy (BA), the area under the Receiver Operating Characteristic curve (AUC), and weighted kappa statistics are evaluated. Features based on shape encodings of femur, tibia, and menisci plus the performed measurements showed most potential as KOA biomarkers. Differentiation between non-arthritic and severely arthritic knees yielded BAs of up to 99%, 84% were achieved for diagnosis of early KOA. Weighted kappa values of 0.73, 0.72, and 0.78 were achieved for classification of the grade of medial JSN, lateral JSN, and KLG, respectively. The AUC was 0.61 and 0.76 for prediction of incident KOA and TKR within one year, respectively. Quantitative features from automated segmentations provide novel biomarkers for KLG and JSN classification and show potential for incident KOA and TKR prediction. The validity of these features should be further evaluated, especially as extensions of CNN-based approaches. To foster such developments we make all segmentations publicly available together with this publication.

Automating classification of osteoarthritis according to Kellgren-Lawrence in the knee using deep learning in an unfiltered adult population

Background

Prevalence for knee osteoarthritis is rising in both Sweden and globally due to increased age and obesity in the population. This has subsequently led to an increasing demand for knee arthroplasties. Correct diagnosis and classification of a knee osteoarthritis (OA) are therefore of a great interest in following-up and planning for either conservative or operative management. Most orthopedic surgeons rely on standard weight bearing radiographs of the knee. Improving the reliability and reproducibility of these interpretations could thus be hugely beneficial. Recently, deep learning which is a form of artificial intelligence (AI), has been showing promising results in interpreting radiographic images. In this study, we aim to evaluate how well an AI can classify the severity of knee OA, using entire image series and not excluding common visual disturbances such as an implant, cast and non-degenerative pathologies.

Methods

We selected 6103 radiographic exams of the knee taken at Danderyd University Hospital between the years 2002-2016 and manually categorized them according to the Kellgren & Lawrence grading scale (KL). We then trained a convolutional neural network (CNN) of ResNet architecture using PyTorch. We evaluated the results against a test set of 300 exams that had been reviewed independently by two senior orthopedic surgeons who settled eventual interobserver disagreements through consensus sessions.

Results

The CNN yielded an overall AUC of more than 0.87 for all KL grades except KL grade 2, which yielded an AUC of 0.8 and a mean AUC of 0.92. When merging adjacent KL grades, all but one group showed near perfect results with AUC > 0.95 indicating excellent performance.

Conclusion

We have found that we could teach a CNN to correctly diagnose and classify the severity of knee OA using the KL grading system without cleaning the input data from major visual disturbances such as implants and other pathologies.

Tenderness of the knee is associated with thinning of the articular cartilage evaluated with ultrasonography in a community-based cohort: The Nagahama study

OBJECTIVES

This study aimed to elucidate the association between joint line tenderness (JLT) of the knee and knee joint structural changes evaluated with ultrasonography (US) for the early diagnosis of knee osteoarthritis (KOA).

METHODS

This cross-sectional study included 121 participants (age 71.7 ± 5.8 years, 75 women) from a community-based population. Bilateral structural changes in the knee joint were evaluated with US, and the presence or absence of JLT was evaluated using a pressure algometer. Logistic regression analysis was performed to evaluate the odds ratios (ORs) of US findings for the presence of JLT. Moreover, when the analysis was limited to knees with pre-/early radiographic KOA, the ORs were also calculated using logistic regression analysis.

RESULTS

Among the 242 knees, 38 had medial JLT, which was significantly associated with female sex (OR 11.87) and loss of cartilage thickness of the distal medial femoral condyle (CTh-MFC) (OR 0.12). Among 96 knees with Kellgren-Lawrence grade ≤ 2, 18 knees had medial JLT, which was also significantly associated with loss of CTh-MFC (OR 0.07) and medial osteophytes (OR 2.01).

CONCLUSIONS

JLT is significantly associated with thinning of the femoral cartilage and larger osteophytes in elderly patients, even in those with pre-/early radiographic KOA.

Association between Patellofemoral and medial Tibiofemoral compartment osteoarthritis progression: exploring the effect of body weight using longitudinal data from osteoarthritis initiative (OAI)

OBJECTIVES

To investigate the associations of medial and lateral patellofemoral osteoarthritis (PF-OA) at baseline with symptomatic and radiographic OA outcomes in the medial tibiofemoral compartment (MTFC) over 4 years, according to baseline overweight status.

METHODS

Data and MRI images of 600 subjects in the FNIH-OA biomarkers consortium were used. Symptomatic worsening and radiographic progression of MTFC-OA were defined using Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain scores and MTFC joint space narrowing (JSN) from baseline to 4-year follow-up. Baseline MRIs were read to establish PF-OA diagnosis. The association between baseline regional PF-OA pattern and odds for MTFC-OA progression was evaluated using regression models (adjusted for relevant confounding covariates including body mass index (BMI), age, sex, PF alignment measurements, KL grade, and knee alignment). To evaluate the effect modifying role for overweight status, stratification analysis was performed (BMI ≥ 25 vs. < 25 kg/m²).

RESULTS

At baseline, 340 (56.7%), 255 (42.5%), and 199 (33.2%) subjects had OA in the medial, lateral, and both PF compartments. Baseline medial PF-OA was associated with WOMAC pain score and MTFC JSN progression at 4 years (Adjusted OR:1.56[95%CI:1.09-2.23] and 1.59[1.11-2.28], respectively) but not lateral PF-OA. In stratification analysis, overweight status was found to be an effect modifier for medial PF-OA and WOMAC pain (OR in overweight vs. non-overweight subjects:1.65[1.13-2.42] vs. 0.50[0.12-1.82]) as well as MTFC-JSN progression (1.63[1.12-2.4] vs. 0.75[0.19-2.81]).

CONCLUSIONS

In addition to the known confounding effect of BMI for PF-OA and MTFC-OA, the overweight status may also play an effect modifier role in the association between baseline medial PF-OA and MTFC-OA progression, which is amenable to secondary prevention.

Learning from Highly Confident Samples for Automatic Knee Osteoarthritis Severity Assessment: Data from the Osteoarthritis Initiative

Knee osteoarthritis (OA) is a chronic disease that considerably reduces patients' life quality. Preventive therapies require early detection and lifetime monitor of OA progression. In the clinical environment, the severity of OA is classified by Kellgren and Lawrence (KL) grading system, ranging from KL-0 to KL-4. Recently, deep learning methods were applied to OA severity assessment to improve the accuracy and efficiency. Researchers fine-tuned convolution neural networks (CNN) on the OA dataset and built end-to-end approaches. However, this task is still challenging due to the ambiguity between adjacent grading, especially in early-stage OA. Low confident samples, which are less representative than the typical ones, undermine the training process. Targeting the uncertainty in the OA dataset, we propose a novel learning scheme that dynamically separates the data into two sets according to their reliability. Besides, we design a hybrid loss function to help CNN learn from the two sets accordingly. With the proposed approach, we emphasize the typical samples and control the impacts of low confident cases. Experiments are conducted in a five-fold manner. Our method achieves a mean accuracy of 70.13\% on the five-class OA assessment task, which outperforms all other start-of-art methods. Despite that early-stage OA detection still benefits from the human intervention of lesion region selection, our approach achieves superior performance on the KL-0 vs. KL-2 task. Moreover, we design an experiment to validate large-scale automatic data refining during training. The result verifies the ability of characterizing low confidence samples by our approach. Dataset used in this paper was obtained from the osteoarthritis Initiative.

Low Rates of Radiographic Knee Osteoarthritis 5 Years After ACL Reconstruction or Rehabilitation Alone: The Delaware-Oslo ACL Cohort Study

BACKGROUND

Patients and clinicians often struggle to choose the optimal management strategy for posttraumatic knee osteoarthritis (OA) after an anterior cruciate ligament (ACL) injury. An evaluation of radiographic outcomes after a decision-making and treatment algorithm applicable in clinical practice can help to inform future recommendations and treatment choices.

PURPOSE

To describe and compare 5-year radiographic outcomes and knee pain in individuals who had gone through our decision-making and treatment algorithm and chosen (1) early (<6 months) ACL reconstruction (ACLR) with pre- and postoperative rehabilitation, (2) delayed (>6 months) ACLR with pre- and postoperative rehabilitation, or (3) progressive rehabilitation alone.

STUDY DESIGN

Cohort study; Level of evidence, 2.

METHODS

We included 276 patients with unilateral ACL injury from a prospective cohort study. Patients chose management using a shared decision-making process and treatment algorithm, and 5-year postoperative radiographs of the index and contralateral knees were assessed using the Kellgren and Lawrence (K&L) classification and minimum joint space width measurements. We defined radiographic tibiofemoral OA as K&L grade ≥2 and knee pain as a Knee injury and Osteoarthritis Outcome Score for Pain ≤72. To further explore early radiographic changes, we included alternative cutoffs for radiographic knee OA using K&L grade ≥2/osteophyte (definite osteophyte without joint space narrowing) and K&L grade ≥1.

RESULTS

At 5 years, 64% had undergone early ACLR; 11%, delayed ACLR; and 25%, progressive rehabilitation alone. Radiographic examination was attended by 187 patients (68%). Six percent of the cohort had radiographic tibiofemoral OA (K&L grade ≥2) in the index knee; 4%, in the contralateral knee. Using the alternative cutoffs at K&L grade ≥2/osteophyte and K&L grade ≥1, the corresponding numbers were 20% and 33% in the index knee and 18% and 29% in the contralateral knee. Six percent had a painful index knee. There were no statistically significant differences in any radiographic outcomes or knee pain among the 3 management groups.

CONCLUSION

There were no statistically significant differences in any 5-year radiographic outcomes or knee pain among the 3 management groups. Very few of the patients who participated in our decision-making and treatment algorithm had knee OA or knee pain at 5 years.

Interpretable and parameter optimized ensemble model for knee osteoarthritis assessment using radiographs

Knee osteoarthritis (KOA) is an orthopedic disorder with a substantial impact on mobility and quality of life. An accurate assessment of the KOA levels is imperative in prioritizing meaningful patient care. Quantifying osteoarthritis features such as osteophytes and joint space narrowing (JSN) from low-resolution images (i.e., X-ray images) are mostly subjective. We implement an objective assessment and quantification of KOA to aid practitioners. In particular, we developed an interpretable ensemble of convolutional neural network (CNN) models consisting of three modules. First, we developed a scale-invariant and aspect ratio preserving model to localize Knee joints. Second, we created multiple instances of "hyperparameter optimized" CNN models with diversity and build an ensemble scoring system to assess the severity of KOA according to the Kellgren-Lawrence grading (KL) scale. Third, we provided visual explanations of the predictions by the ensemble model. We tested our models using a collection of 37,996 Knee joints from the Osteoarthritis Initiative (OAI) dataset. Our results show a superior (13-27%) performance improvement compared to the state-of-the-art methods.

Muscle strength and osteoarthritis progression after surgery or exercise for degenerative meniscal tears: Secondary analyses of a randomized trial

OBJECTIVE

To evaluate muscle strength changes following partial meniscectomy or exercise therapy for degenerative meniscal tears and the relationship between baseline muscle strength and osteoarthritis progression.

METHODS

Secondary analysis of a randomized trial (n=140 participants). Isokinetic quadriceps and hamstrings strength (peak torque [N·m/kg] and total work [J/kg]) were assessed at baseline, three-, 12-month, and five-year follow-up. Between-group differences were analyzed using intention-to-treat linear mixed models. The relationship between baseline muscle strength and osteoarthritis progression (Kellgren and Lawrence, ≥1 grade increase) were assessed using logistic regression models.

RESULTS

We found statistically significant between-group differences favoring exercise therapy at three months (quadriceps: -0.30 N·m/kg, 95% CI -0.40, -0.20; hamstrings: -0.10 N·m/kg, 95% CI -0.15, -0.04) and 12 months (quadriceps: -0.13 N·m/kg, 95% CI -0.23, -0.03; hamstrings: -0.08 N·m/kg, 95% CI -0.14, -0.03). At five years, between-group differences were -0.10 N·m/kg (95% CI -0.21 to 0.01) for quadriceps and -0.07 N·m/kg (95% CI -0.13 to -0.01) for hamstrings. Quadriceps muscle weakness at baseline was associated with knee osteoarthritis progression over five years: adjusted odds ratio of 1.40 for every 0.2 N·m/kg decrease (95% CI 1.15 to 1.71). The adjusted odds ratio for hamstrings was 1.14 (95% CI 0.97-1.35) for every 0.1 N·m/kg decrease.

CONCLUSION

Exercise therapy was effective in improving muscle strength at three and 12-month follow-up compared to partial meniscectomy, but the effect was attenuated at five years. Quadriceps muscle weakness at baseline was associated with higher odds of osteoarthritis progression over five years.

The Clinical Radiographic Incidence of Posttraumatic Osteoarthritis 10 Years After Anterior Cruciate Ligament Reconstruction: Data From the MOON Nested Cohort

BACKGROUND

The incidence of posttraumatic osteoarthritis (PTOA) based on clinical radiographic grading criteria at 10 years after anterior cruciate ligament (ACL) reconstruction (ACLR) has not been well-defined in a prospective cohort of young athletic patients.

HYPOTHESIS

Among young athletic patients, there is a high incidence of clinical radiographic PTOA at 10 years after ACLR. Additionally, there is a significant difference in clinical radiographic osteoarthritis (OA) changes (joint space narrowing and osteophyte formation) between ACL-reconstructed and contralateral knees at 10 years.

STUDY DESIGN

Case series; Level of evidence, 4.

METHODS

The first 146 patients in an ongoing nested cohort study of the Multicenter Orthopaedic Outcomes Network (MOON) prospective cohort presented for a minimum 10-year follow-up. Included patients had a sports-related ACL injury, were aged <33 years at the time of ACLR, had no history of ipsilateral or contralateral knee surgery, and did not undergo revision ACLR before follow-up. Bilateral knee metatarsophalangeal view radiographs were obtained and graded according to International Knee Documentation Committee (IKDC), Osteoarthritis Research Society International (OARSI), and modified Kellgren-Lawrence (KL) criteria by 2 blinded reviewers. The incidence and severity of ipsilateral and contralateral radiographic OA were determined among patients without a contralateral ACL injury before 10-year follow-up (N = 133).

RESULTS

Interrater reliability was substantial for the IKDC (Gwet Agreement Coefficient [AC] 1 = 0.71), moderate for the KL (0.48), and almost perfect for the OARSI (0.84) grading systems. Among patients with a contralateral radiographically normal knee, the 10-year incidence of clinical radiographic PTOA after ACLR was 37% as defined by osteophytes and 23% as defined by joint space narrowing. The maximum side-to-side difference in the OARSI osteophyte grade in the medial or lateral compartment was 0 in 65% of patients, 1 in 20%, and ≥2 in 15%. The maximum side-to-side difference in the OARSI joint space narrowing grade was 0 in 77% of patients, 1 in 19%, and ≥2 in 4%.

CONCLUSION

In young active patients, the 10-year incidence of clinical radiographic PTOA after ACLR was 37% as defined by osteophytes and 23% as defined by joint space narrowing. The mean difference in the degree of osteophyte formation (≤1 grade in 85%) and joint space narrowing (≤1 grade in 96%) between the ACL-reconstructed and contralateral knees was small.

REGISTRATION

NCT02717559 (ClinicalTrials.gov identifier).

Can Additional Patient Information Improve the Diagnostic Performance of Deep Learning for the Interpretation of Knee Osteoarthritis Severity

The study compares the diagnostic performance of deep learning (DL) with that of the former radiologist reading of the Kellgren-Lawrence (KL) grade and evaluates whether additional patient data can improve the diagnostic performance of DL. From March 2003 to February 2017, 3000 patients with 4366 knee AP radiographs were randomly selected. DL was trained using knee images and clinical information in two stages. In the first stage, DL was trained only with images and then in the second stage, it was trained with image data and clinical information. In the test set of image data, the areas under the receiver operating characteristic curve (AUC)s of the DL algorithm in diagnosing KL 0 to KL 4 were 0.91 (95% confidence interval (CI), 0.88-0.95), 0.80 (95% CI, 0.76-0.84), 0.69 (95% CI, 0.64-0.73), 0.86 (95% CI, 0.83-0.89), and 0.96 (95% CI, 0.94-0.98), respectively. In the test set with image data and additional patient information, the AUCs of the DL algorithm in diagnosing KL 0 to KL 4 were 0.97 (95% confidence interval (CI), 0.71-0.74), 0.85 (95% CI, 0.80-0.86), 0.75 (95% CI, 0.66-0.73), 0.86 (95% CI, 0.79-0.85), and 0.95 (95% CI, 0.91-0.97), respectively. The diagnostic performance of image data with additional patient information showed a statistically significantly higher AUC than image data alone in diagnosing KL 0, 1, and 2 (p-values were 0.008, 0.020, and 0.027, respectively).The diagnostic performance of DL was comparable to that of the former radiologist reading of the knee osteoarthritis KL grade. Additional patient information improved DL diagnosis in interpreting early knee osteoarthritis.

Mesenchymal stem cells - a promising strategy for treating knee osteoarthritis

AIMS

The purpose of our study was to determine whether mesenchymal stem cells (MSCs) are an effective and safe therapeutic agent for the treatment of knee osteoarthritis (OA), owing to their cartilage regeneration potential.

METHODS

We searched PubMed, Embase, and the Cochrane Library, with keywords including "knee osteoarthritis" and "mesenchymal stem cells", up to June 2019. We selected randomized controlled trials (RCTs) that explored the use of MSCs to treat knee OA. The visual analogue scale (VAS), Western Ontario and McMaster University Osteoarthritis Index (WOMAC), adverse events, and the whole-organ MRI score (WORMS) were used as the primary evaluation tools in the studies. Our meta-analysis included a subgroup analysis of cell dose and cell source.

RESULTS

Seven trials evaluating 256 patients were included in the meta-analysis. MSC treatment significantly improved the VAS (mean difference (MD), -13.24; 95% confidence intervals (CIs) -23.28 to -3.20, p = 0.010) and WOMAC (MD, -7.22; 95% CI -12.97 to -1.47, p = 0.010). The low-dose group with less than 30 million cells showed lower p-values for both the VAS and WOMAC. Adipose and umbilical cord-derived stem cells also had lower p-values for pain scores than those derived from bone marrow.

CONCLUSION

Overall, MSC-based cell therapy is a relatively safe treatment that holds great potential for OA, evidenced by a positive effect on pain and knee function. Using low-dose (25 million) and adipose-derived stem cells is likely to achieve better results, but further research is needed. Cite this article: Bone Joint Res 2020;9(10):719-728.

Vibroarthrographic Signal Spectral Features in 5-Class Knee Joint Classification

Vibroarthrography (VAG) is a non-invasive and potentially widely available method supporting the joint diagnosis process. This research was conducted using VAG signals classified to five different condition classes: three stages of chondromalacia patellae, osteoarthritis, and control group (healthy knee joint). Ten new spectral features were proposed, distinguishing not only neighboring classes, but every class combination. Additionally, Frequency Range Maps were proposed as the frequency feature extraction visualization method. The results were compared to state-of-the-art frequency features using the Bhattacharyya coefficient and the set of ten different classification algorithms. All methods evaluating proposed features indicated the superiority of the new features compared to the state-of-the-art. In terms of Bhattacharyya coefficient, newly proposed features proved to be over 25% better, and the classification accuracy was on average 9% better.

Discriminative Regularized Auto-Encoder for Early Detection of Knee OsteoArthritis: Data from the Osteoarthritis Initiative

OsteoArthritis (OA) is the most common disorder of the musculoskeletal system and the major cause of reduced mobility among seniors. The visual evaluation of OA still suffers from subjectivity. Recently, Computer-Aided Diagnosis (CAD) systems based on learning methods showed potential for improving knee OA diagnostic accuracy. However, learning discriminative properties can be a challenging task, particularly when dealing with complex data such as X-ray images, typically used for knee OA diagnosis. In this paper, we introduce a Discriminative Regularized Auto Encoder (DRAE) that allows to learn both relevant and discriminative properties that improve the classification performance. More specifically, a penalty term, called discriminative loss is combined with the standard Auto-Encoder training criterion. This additional term aims to force the learned representation to contain discriminative information. Our experimental results on data from the public multicenter OsteoArthritis Initiative (OAI) show that the developed method presents potential results for early knee OA detection.

Identification of the most important features of knee osteoarthritis structural progressors using machine learning methods

Objectives:

The aim was to identify the most important features of structural knee osteoarthritis (OA) progressors and classification using machine learning methods.

Methods:

Participants, features and outcomes were from the Osteoarthritis Initiative. Features were from baseline (1107), including articular knee tissues (135) assessed by quantitative magnetic resonance imaging (MRI). OA progressors were ascertained by four outcomes: cartilage volume loss in medial plateau at 48 and 96 months (Prop_CV_48M, 96M), Kellgren–Lawrence (KL) grade ⩾ 2 and medial joint space narrowing (JSN) ⩾ 1 at 48 months. Six feature selection models were used to identify the common features in each outcome. Six classification methods were applied to measure the accuracy of the selected features in classifying the subjects into progressors and non-progressors. Classification of the best features was done using an automatic machine learning interface and the area under the curve (AUC). To prioritize the top five features, sparse partial least square (sPLS) method was used.

Results:

For the classification of the best common features in each outcome, Multi-Layer Perceptron (MLP) achieved the highest AUC in Prop_CV_96M, KL and JSN (0.80, 0.88, 0.95), and Gradient Boosting Machine for Prop_CV_48M (0.70). sPLS showed the baseline top five features to predict knee OA progressors are the joint space width, mean cartilage thickness of the medial tibial plateau and sub-regions and JSN.

Conclusion:

In this comprehensive study using a large number of features (n = 1107) and MRI outcomes in addition to radiological outcomes, we identified the best features and classification methods for knee OA structural progressors. Data revealed baseline X-ray and MRI-based features could predict early OA knee progressors and that MLP is the best classification method.

Can We Predict Severity of Osteoarthritis of Knees and Compartmental Involvement Based on a Set of Predefined Clinical Questions in Patients of Knee Pain?

BACKGROUND

We investigated whether the severity of Osteoarthritis (OA) knees can be predicted based on a set of predefined clinical questions (PCQs) about activities of daily living (ADL). We studied the association of demographic factors and advanced radiographic OA (KL 3 and 4) and the relationship between various physical activities and radiographic involvement of knee joint compartments based on PCQs.

MATERIALS AND METHODS

Demographic data, radiographic grading of knee OA and PCQs score, were obtained prospectively. Patients' responses to PCQs were marked as scores-that were predefined and graded according to the severity of knee pain. Radiographic knee OA grades were dichotomized and patients were classified as either negative (KL grade 1, 2) or positive (KL grade 3, 4). Multivariate logistic regression was performed to obtain the adjusted odds for total PCQs score in relation with positive radiographic OA considering confounders like age, gender and BMI in the model. Log odds score (LOS) were obtained and ROC analysis was performed on scores to obtain the cut-off value for the screening of knee OA in patients of knee pain.

RESULTS

Age and BMI were significantly negatively correlated with PCQs score (r = - 0.473; P < 0.0001 and r = - 0.136; P = 0.046). PCQs scores were significantly lower in females (P = 0.031). Total PCQs score had corresponding OR of 0.901 (P = 0.002) towards knee OA after adjusting for age, gender and BMI. Multivariate model-based LOS resulted in a cut-off of 1.315, which had a sensitivity of 85.5%, specificity of 66.7% and PPV of 92.7%.

CONCLUSION

Severity of knee OA can be predicted based on PCQs. PCQs can predict severity of knee OA and patellofemoral or medial tibiofemoral compartment without radiographs. LOS based on demographics and total PCQs score can be developed as a screening tool for advanced knee OA.

Development of osteoarthritis in patients with degenerative meniscal tears treated with exercise therapy or surgery: a randomized controlled trial

OBJECTIVE

To evaluate progression of individual radiographic features 5 years following exercise therapy or arthroscopic partial meniscectomy as treatment for degenerative meniscal tear.

DESIGN

Randomized controlled trial including 140 adults, aged 35-60 years, with a magnetic resonance image verified degenerative meniscal tear, and 96% without definite radiographic knee osteoarthritis. Participants were randomized to either 12-weeks of supervised exercise therapy or arthroscopic partial meniscectomy. The primary outcome was between-group difference in progression of tibiofemoral joint space narrowing and marginal osteophytes at 5 years, assessed semi-quantitatively by the OARSI atlas. Secondary outcomes included incidence of radiographic knee osteoarthritis and symptomatic knee osteoarthritis, medial tibiofemoral fixed joint space width (quantitatively assessed), and patient-reported outcome measures. Statistical analyses were performed using a full analysis set. Per protocol and as treated analysis were also performed.

RESULTS

The risk ratios (95% CI) for progression of semi-quantitatively assessed joint space narrowing and medial and lateral osteophytes for the surgery group were 0.89 (0.55-1.44), 1.15 (0.79-1.68) and 0.77 (0.42-1.42), respectively, compared to the exercise therapy group. In secondary outcomes (full-set analysis) no statistically significant between-group differences were found.

CONCLUSION

The study was inconclusive with respect to potential differences in progression of individual radiographic features after surgical and non-surgical treatment for degenerative meniscal tear. Further, we found no strong evidence in support of differences in development of incident radiographic knee osteoarthritis or patient-reported outcomes between exercise therapy and arthroscopic partial meniscectomy.

TRIAL REGISTRATION

www.clinicaltrials.gov (NCT01002794).

Prediction of Total Knee Replacement and Diagnosis of Osteoarthritis by Using Deep Learning on Knee Radiographs: Data from the Osteoarthritis Initiative

Background The methods for assessing knee osteoarthritis (OA) do not provide enough comprehensive information to make robust and accurate outcome predictions. Purpose To develop a deep learning (DL) prediction model for risk of OA progression by using knee radiographs in patients who underwent total knee replacement (TKR) and matched control patients who did not undergo TKR. Materials and Methods In this retrospective analysis that used data from the OA Initiative, a DL model on knee radiographs was developed to predict both the likelihood of a patient undergoing TKR within 9 years and Kellgren-Lawrence (KL) grade. Study participants included a case-control matched subcohort between 45 and 79 years. Patients were matched to control patients according to age, sex, ethnicity, and body mass index. The proposed model used a transfer learning approach based on the ResNet34 architecture with sevenfold nested cross-validation. Receiver operating characteristic curve analysis and conditional logistic regression assessed model performance for predicting probability and risk of TKR compared with clinical observations and two binary outcome prediction models on the basis of radiographic readings: KL grade and OA Research Society International (OARSI) grade. Results Evaluated were 728 participants including 324 patients (mean age, 64 years ± 8 [standard deviation]; 222 women) and 324 control patients (mean age, 64 years ± 8; 222 women). The prediction model based on DL achieved an area under the receiver operating characteristic curve (AUC) of 0.87 (95% confidence interval [CI]: 0.85, 0.90), outperforming a baseline prediction model by using KL grade with an AUC of 0.74 (95% CI: 0.71, 0.77; P < .001). The risk for TKR increased with probability that a person will undergo TKR from the DL model (odds ratio [OR], 7.7; 95% CI: 2.3, 25; P < .001), KL grade (OR, 1.92; 95% CI: 1.17, 3.13; P = .009), and OARSI grade (OR, 1.20; 95% CI: 0.41, 3.50; P = .73). Conclusion The proposed deep learning model better predicted risk of total knee replacement in osteoarthritis than did binary outcome models by using standard grading systems. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Richardson in this issue.

Coronal Alignment of the Lower Extremity: A Gender-Based Radio-Graphic Analysis in Indian Patients

BACKGROUND

Variation in the shape of the femur and tibia has been shown to influence hip-kneeankle-axis angle (HKAA) and bears a role in reconstructive surgeries such as total knee arthroplasty (TKA). However, data on the axial alignment of the lower extremity specific to sex largely remains unavailable. Thus, we conducted a study with an aim to measure alignment and calculate differences in both the sexes in Indian population.

MATERIALS AND METHODS

The following radiological parameters were measured from weight-bearing long leg radiographs of 966 limbs of Indian subjects via a morphometric software (Matlab R2009a) (1) Hip-Knee-Ankle angle (HKAA), (2) femoral bowing, (3) tibial bowing, (4) condylar plateau angle (CPA). The knees were classified according to the Kellegren and Lawrence grading and the differences between both the sexes were calculated with appropriate statistical tests.

RESULTS

56.04% of the subjects were female. An increase in the mean age was observed for both the genders with an increase in the severity of OA. Height did not show any significant association with the alignment of the limb. The mean HKAA observed was - 5.88° ± 0.35° in females and - 4.99° ± 0.41° in males. The overall mean femoral bow and tibial bow was - 1.26° ± 0.24°, - 1.60° ± 0.18° in females and - 1.09 ± 0.28, - 1.47° ± 0.21° in males. The mean condylar plateau angle was higher in females - 2.67 ± 0.34 as compared to males - 2.35° ± 0.39°. A greater lateral bow was seen in males at higher grades of OA for femur and at lower grades of OA for tibia.

CONCLUSION

This study provides gender-based differences in the various axial radio-graphic parameters in a long leg radio-graphs in Indian population which might help in a better understanding of the etiopathogenesis of osteoarthritis and also help planning and execution of reconstructive surgeries such as TKA.

Does surgery reduce knee osteoarthritis, meniscal injury and subsequent complications compared with non-surgery after ACL rupture with at least 10 years follow-up? A systematic review and meta-analysis

Objective

We compared long-term follow-up from surgical versus non-surgical treatment of ACL rupture regarding radiographic knee osteoarthritis (OA), secondary surgery, laxity and patient-reported outcome measures (PROMs).

Design

Systematic review and meta-analysis.

Data sources

Embase, MEDLINE, CINAHL and the Cochrane Library databases.

Eligibility criteria for selecting studies

Studies directly comparing the minimally invasive surgical (arthroscopy or miniarthrotomy) and non-surgical treatment of ACL rupture with at least 10 years of follow-up in adult patients were included.

Results

Five studies met the eligibility criteria. A meta-analysis revealed a higher risk of radiographic knee OA and a lower risk of secondary meniscal surgery for patients in the surgical group. The risk of graft rupture/secondary ACL revision and secondary ACL reconstruction was equal in the surgical and non-surgical groups. Knee laxity was lower among patients in the surgical group in four studies. No difference was found in the PROMs (ie, International Knee Documentation Committee, Tegner, Knee Injury and Osteoarthritis Outcome, and Lysholm scores).

Conclusion

The risk of radiographic knee OA was higher, but the risk of secondary meniscal injury was lower 10 years after surgical treatment of ACL rupture. The risk of graft rupture/secondary ACL revision or secondary reconstruction was unrelated to treatment type. The degree of knee laxity was reduced after surgical treatment in comparison with non-surgical treatment, while PROMs were similar. However, due to the methodological challenges highlighted in this systematic review, these findings must be interpreted with caution.

PROSPERO registration number

CRD42019119468

Intraobserver and interobserver reliability of the computed tomography-based radiographic classification of primary elbow osteoarthritis: comparison with plain radiograph-based classification and clinical assessment

OBJECTIVES

To develop a staging system that could better reflect symptoms by the spurs quantification in the fossa and joint space narrowing using computed tomography (CT) for elbow arthritis and to evaluate its reproducibility with multiple readers.

METHODS

This retrospective study evaluated 81 cases of primary elbow osteoarthritis using both plain radiography and CT. Qualitative and quantitative analyses were independently performed by four orthopedic surgeons using previous and newly proposed staging systems. The reproducibility of the new system was analyzed with intraclass correlation coefficients (ICC). Correlations between symptoms and radiologic classification were assessed using Pearson's correlation coefficient (PCC).

RESULTS

The interobserver agreement (1) and intraobserver agreement (2) among the four evaluators was present by ICC. (1) The system of Hastings and Rettig [first observation, 0.544 (95% confidence interval (CI), 0.436-0.649); second observation, 0.582 (95% CI, 0.478-0.682)] and Broberg and Morrey's staging system [first observation, 0.620 (95% CI, 0.521-0.714); second observation, 0.656 (95% CI, 0.562-0.743)] showed substantial and moderate retrospective agreement, whereas the CT-based staging system showed almost perfect agreement [first observation, 0.867 (95% CI, 0.820-0.906); second observation, 0.909 (95% CI, 0.875-0.936)]. (2) The intraobserver agreement was almost perfect in the Brogerg and Morrey's and CT-based staging systems. CT-based staging showed high correlation with visual analogue scale (PCC 0.754, P < 0.001) and Mayo elbow performance score (PCC -0.614, P < 0.001) and moderate correlation with range of motion (PCC -0.458, P < 0.001).

CONCLUSIONS

CT-based staging system was highly reproducible and clinically feasible than previous plain radiograph-based staging systems.

The lateral joint space width can be measured reliably with Telos valgus stress radiography in medial knee osteoarthritis

OBJECTIVE

To examine the reproducibility of valgus stress radiographs with the Telos stress device for assessment of lateral compartment degenerative changes in patients with medial osteoarthritis of the knee.

MATERIALS AND METHODS

A prospective reliability study was performed. Seventy-nine patients (80 knees) were included, and standardized valgus stress radiographs were obtained using the Telos stress device. Osteophytes and joint space narrowing (JSN) were graded using the OARSI (Osteoarthritis Research Society International) classification, and the joint space width (JSW) was measured in millimeters. Reproducibility was determined as intra-and inter-rater reliability and test-retest reliability. Weighted kappa was used to determine the reliability of osteophyte and JSN grading, and the intra-class correlation coefficient for JSW.

RESULTS

Grading of osteophytes had an intra- and inter-rater reliability ranging from 0.40 to 0.83 on the medial side and ranging from 0.39 to 0.87 on the lateral side. Grading of medial JSN had an intra- and inter-rater reliability ranging from 0.62 to 0.84, and grading of lateral JSN had an intra- and inter-rater reliability ranging from 0.32 to 0.65. Intra- and inter-rater reliability of JSW ranged from 0.84 to 0.98 on the medial side, and from 0.59 to 0.89 on the lateral side. Test-retest reliability of JSW of the medial and lateral side ranged from 0.69 to 0.92.

CONCLUSIONS

Standardized valgus stress radiographs taken with the Telos stress device are a reliable supplement in the assessment of medial OA of the knee. Evaluation of the lateral compartment on valgus stress radiographs is most reliable with measurement of the lateral JSW.

Fully automatic knee osteoarthritis severity grading using deep neural networks with a novel ordinal loss

Knee osteoarthritis (OA) is one major cause of activity limitation and physical disability in older adults. Early detection and intervention can help slow down the OA degeneration. Physicians' grading based on visual inspection is subjective, varied across interpreters, and highly relied on their experience. In this paper, we successively apply two deep convolutional neural networks (CNN) to automatically measure the knee OA severity, as assessed by the Kellgren-Lawrence (KL) grading system. Firstly, considering the size of knee joints distributed in X-ray images with small variability, we detect knee joints using a customized one-stage YOLOv2 network. Secondly, we fine-tune the most popular CNN models, including variants of ResNet, VGG, and DenseNet as well as InceptionV3, to classify the detected knee joint images with a novel adjustable ordinal loss. To be specific, motivated by the ordinal nature of the knee KL grading task, we assign higher penalty to misclassification with larger distance between the predicted KL grade and the real KL grade. The baseline X-ray images from the Osteoarthritis Initiative (OAI) dataset are used for evaluation. On the knee joint detection, we achieve mean Jaccard index of 0.858 and recall of 92.2% under the Jaccard index threshold of 0.75. On the knee KL grading task, the fine-tuned VGG-19 model with the proposed ordinal loss obtains the best classification accuracy of 69.7% and mean absolute error (MAE) of 0.344. Both knee joint detection and knee KL grading achieve state-of-the-art performance. The code, dataset, and models are released at https://github.com/PingjunChen/KneeAnalysis.