Purpose: The objective of this study was to create and assess a Deep Learning-Based Radiomics model using a single sequence MRI that could accurately predict early Femoral Head Osteonecrosis (ONFH). This is the first time such a model was used for the diagnosis of early ONFH. Its simpler than the previously published multi-sequence MRI radiomics based method, and it implements Deep learning to improve on radiomics. It has the potential to be highly beneficial in the early stages of diagnosis and treatment planning.Methods: MRI scans from 150 patients in total (80 healthy, 70 necrotic) were used, and split into training and testing sets in a 7:3 ratio. Handcrafted as well as deep learning features were retrieved from Tesla 2 weighted (T2W1) MRI slices. After a rigorous selection process, these features were used to construct three models: a Radiomics-based (Rad-model), a Deep Learning-based (DL-model), and a Deep Learning-based Radiomics (DLR-model). The performance of these models in predicting early ONFH was evaluated by comparing them using the receiver operating characteristic (ROC) and decision curve analysis (DCA).Results: 1,197 handcrafted radiomics and 512 DL features were extracted then processed; after the final selection: 15 features were used for the Rad-model, 12 features for the DL-model, and only 9 features were selected for the DLR-model. The most effective algorithm that was used in all of the models was Logistic regression (LR). The Rad-model depicted good results outperforming the DL-model; AUC = 0.944 (95%CI, 0.862–1.000) and AUC = 0.930 (95%CI, 0.838–1.000) respectively. The DLR-model showed superior results to both Rad-model and the DL-model; AUC = 0.968 (95%CI, 0.909–1.000); and a sensitivity of 0.95 and specificity of 0.920. The DCA showed that DLR had a greater net clinical benefit in detecting early ONFH.Conclusion: Using a single sequence MRI scan, our work constructed and verified a Deep Learning-Based Radiomics Model for early ONFH diagnosis. This strategy outperformed a Deep learning technique based on Resnet18 and a model based on Radiomics. This straightforward method can offer essential diagnostic data promptly and enhance early therapy strategizing for individuals with ONFH, all while utilizing just one MRI sequence and a more standardized and objective interpretation of MRI images.

Purpose: Distinguishing between Osteonecrosis of the femoral head (ONFH) and Osteoarthritis (OA) can be subjective and vary between users with different backgrounds and expertise. This study aimed to construct and evaluate several Radiomics-based machine learning models using MRI to differentiate between those two disorders and compare their efficacies to those of medical experts.Methods: 140 MRI scans were retrospectively collected from the electronic medical records. They were split into training and testing sets in a 7:3 ratio. Handcrafted radiomics features were harvested following the careful manual segmentation of the regions of interest (ROI). After thoroughly selecting these features, various machine learning models have been constructed. The evaluation was carried out using receiver operating characteristic (ROC) curves. Then NaiveBayes (NB) was selected to establish our final Radiomics-model as it performed the best. Three users with different expertise and backgrounds diagnosed and labeled the dataset into either OA or ONFH. Their results have been compared to our Radiomics-model.Results: The amount of handcrafted radiomics features was 1197 before processing; after the final selection, only 12 key features were retained and used. User 1 had an AUC of 0.632 (95% CI 0.4801-0.7843), User 2 recorded an AUC of 0.565 (95% CI 0.4102-0.7196); while User 3 was on top with an AUC of 0.880 (95% CI 0.7753-0.9843). On the other hand, the Radiomics model attained an AUC of 0.971 (95% CI 0.9298-1.0000); showing greater efficacy than all other users. It also demonstrated a sensitivity of 0.937 and a specificity of 0.885. DCA (Decision Curve Analysis displayed that the radiomics-model had a greater clinical benefit in differentiating OA and ONFH.Conclusion: We have successfully constructed and evaluated an interpretable radiomics-based machine learning model that could distinguish between OA and ONFH. This method has the ability to aid both junior and senior medical professionals to precisely diagnose and take prompt treatment measures.

Knee osteoarthritis prognostic tools often target structural progression or surgery and require imaging or biomarker inputs that are not routinely available. Using Osteoarthritis Initiative data, we developed a fully clinical nomogram to estimate both the probability of long-term pain non-resilience (clinically important worsening) and, by complement, maintenance of acceptable pain in radiographic knee osteoarthritis. We included participants with radiographic knee osteoarthritis and complete worst-knee WOMAC pain scores at baseline, 24 and 48 months; non-resilience was defined as a ≥9-point increase on the 0–100 WOMAC pain scale over 4 years. A six-predictor Firth logistic regression model (age, body mass index, Kellgren–Lawrence grade, baseline pain, 0–24-month pain change and Center for Epidemiologic Studies Depression Scale score) was fitted and translated into a point-based nomogram. Among 2365 eligible participants, 527 (22.3%) were non-resilient. The model showed good performance, with optimism-corrected AUC 0.74 and Brier score 0.15, and decision-curve analysis indicated positive net benefit versus treat-none across 1–15% thresholds and small gains versus treat-all. Early pain worsening and higher depressive symptoms were the strongest predictors of non-resilience. This six-variable, clinic-ready nomogram provides a simple, well-calibrated tool for prognostic counseling and risk stratification in radiographic knee osteoarthritis and requires external validation before wider clinical use.

Background: Knee dislocation is an uncommon but limb-threatening injury that often arises during athletic activities and can result in neurovascular compromise, multi-ligament disruption, and long-term disability. However, the Global Burden of Disease (GBD) framework does not include a specific sports-injury category, so the contribution of athletic-type mechanisms to the global burden of knee dislocation in young adults remains unclear. We, therefore, used GBD 2021 data to estimate the young-adult component of the global burden of knee dislocation from 1990 to 2021, describe long-term trends, and compare a proxy for athletic-type mechanisms with other external causes. Methods: We performed a secondary analysis of Global Burden of Disease 2021 estimates for knee dislocation, extracting annual incidence and prevalence for adults aged 20–40 years between 1990 and 2021 by sex. A proxy for athletic-type mechanisms was defined as cases assigned to the external-cause categories “falls” and “other exposure to mechanical forces” and was compared with transport injuries and all other external causes. We calculated population age-standardized rates per 100,000 populations with 95% uncertainty intervals and summarized temporal patterns using average annual percentage change derived from log-linear regression models. Results: Absolute incidence counts increased for all causes (1023103→1084122; +5.96%) and the athletic proxy (577923→675111; +16.82%), increasing the proxy’s share of 20–40 incidence from 56.49% to 62.27%. Population-weighted incidence rates decreased for all causes (61.18→46.12 per 100000; −24.62%, APC −0.94%) and the proxy (34.56→28.72; −16.90%, APC −0.74%). Prevalence rates also decreased (all causes 11.20→8.59; proxy 6.12→5.08), while prevalence counts increased (all causes 187256→202052; proxy 102281→119503). In 2021, proxy rates were higher in male individuals than in female individuals (incidence 36.49 vs. 20.76; prevalence 6.06 vs. 4.08 per 100,000). Within the proxy, falls increased from 76.58% to 81.12% of incident cases. Conclusion: Young adults carry a large global burden of knee dislocation, consistent with athletic trauma. Despite increasing counts, population-weighted rates fell steadily from 1990 to 2021; male individuals remain at higher risks, and falls dominate the proxy mechanism. The proxy approach offers a reproducible way to monitor athletic-type knee dislocations where sports-injury labels are absent.

Deep learning models for Kellgren–Lawrence (KL) grading often report optimistic performance due to data leakage and fail to generalize across institutions because of domain shift. To address this reproducibility crisis, we introduce KL‐FuseNet, a multitask architecture fusing global (ConvNeXt‐Base) and local (ResNet‐50) features to predict ordinal grades, label distributions, and binary severity (KL≥2). Using strict patient‐wise stratified splits on an internal osteoarthritis initiative dataset ( n = 8260) and an independent Chinese cohort ( n = 2295), we compared zero‐shot transfer against selective fine‐tuning. KL‐FuseNet achieved robust internal agreement (quadratic Cohen's kappa [QWK]: 0.881; accuracy: 70.3%). While external zero‐shot deployment revealed a domain gap, with accuracy dropping to 66.1%, our selective fine‐tuning protocol significantly bridged this divide, boosting external accuracy to 80.0% and QWK to 0.950, with an AUC of 0.984 for clinically significant osteoarthritis (KL≥2). These results demonstrate that while KL‐FuseNet achieves state‐of‐the‐art performance under rigorous evaluation, domain‐aware adaptation is essential for clinical utility. This study establishes a reproducible pathway for deploying automated grading models across heterogeneous medical centers.

Introduction: Knee osteoarthritis is heterogeneous in radiographic severity, pain, and multimorbidity, and racial disparities in pain and knee replacement are documented. We derived knee osteoarthritis phenotypes from electronic health record and radiograph data and assessed their associations with race and incident total knee arthroplasty. Methods: We conducted a retrospective cohort study in the Emory Knee Radiograph dataset of adults aged ≥40 years with knee osteoarthritis (n = 33,553). Among participants with pain scores within ±90 days of the index radiograph (n = 20,030), we applied k-means clustering to features including age, radiographic severity, pain measures, and comorbidity indicators to define phenotypes. Among participants without pre-index strict total knee arthroplasty, we used Cox proportional hazards models to evaluate time to incident post-index strict total knee arthroplasty, with follow-up censored at last observed contact and capped at the procedure-code ascertainment horizon. Results: Clustering identified three phenotypes: structural/metabolic disease characterized by radiographic severity and multimorbidity; younger trauma-associated mild disease; and pain-dominant disease with pain disproportionate to radiographic severity. Black individuals were overrepresented in the pain-dominant phenotype and underrepresented in the structural/metabolic phenotype, whereas White individuals showed the opposite pattern. In time-to-event models, higher radiographic severity and male sex were associated with faster time to total knee arthroplasty, whereas the younger trauma-associated and pain-dominant phenotypes were associated with lower hazard of total knee arthroplasty than the structural/metabolic phenotype. These findings were preserved in known-race-only and categorical Kellgren–Lawrence grade sensitivity analyses. Discussion: Routinely available electronic health record and radiographic data can identify clinically meaningful knee osteoarthritis phenotypes that differ by race and are independently associated with subsequent total knee arthroplasty. These findings support further validation of phenotype-aware approaches to studying disparities and guiding osteoarthritis care.