Advanced neural networks for classification of MRI in psoriatic arthritis, seronegative, and seropositive rheumatoid arthritis.

OBJECTIVES: To evaluate whether neural networks can distinguish between seropositive RA, seronegative RA, and PsA based on inflammatory patterns from hand MRIs and to test how psoriasis patients with subclinical inflammation fit into such patterns. METHODS: ResNet neural networks were utilized to compare seropositive RA vs PsA, seronegative RA vs PsA, and seropositive vs seronegative RA with respect to hand MRI data. Results from T1 coronal, T2 coronal, T1 coronal and axial fat-suppressed contrast-enhanced (CE), and T2 fat-suppressed axial sequences were used. The performance of such trained networks was analysed by the area under the receiver operating characteristics curve (AUROC) with and without presentation of demographic and clinical parameters. Additionally, the trained networks were applied to psoriasis patients without clinical arthritis. RESULTS: MRI scans from 649 patients (135 seronegative RA, 190 seropositive RA, 177 PsA, 147 psoriasis) were fed into ResNet neural networks. The AUROC was 75% for seropositive RA vs PsA, 74% for seronegative RA vs PsA, and 67% for seropositive vs seronegative RA. All MRI sequences were relevant for classification, however, when deleting contrast agent-based sequences the loss of performance was only marginal. The addition of demographic and clinical data to the networks did not provide significant improvements for classification. Psoriasis patients were mostly assigned to PsA by the neural networks, suggesting that a PsA-like MRI pattern may be present early in the course of psoriatic disease. CONCLUSION: Neural networks can be successfully trained to distinguish MRI inflammation related to seropositive RA, seronegative RA, and PsA.

Deep learning methods allow fully automated segmentation of metacarpal bones to quantify volumetric bone mineral density.

Arthritis patients develop hand bone loss, which leads to destruction and functional impairment of the affected joints. High resolution peripheral quantitative computed tomography (HR-pQCT) allows the quantification of volumetric bone mineral density (vBMD) and bone microstructure in vivo with an isotropic voxel size of 82 micrometres. However, image-processing to obtain bone characteristics is a time-consuming process as it requires semi-automatic segmentation of the bone. In this work, a fully automatic vBMD measurement pipeline for the metacarpal (MC) bone using deep learning methods is introduced. Based on a dataset of HR-pQCT volumes with MC measurements for 541 patients with arthritis, a segmentation network is trained. The best network achieves an intersection over union as high as 0.94 and a Dice similarity coefficient of 0.97 while taking only 33 s to process a whole patient yielding a speedup between 2.5 and 4.0 for the whole workflow. Strong correlation between the vBMD measurements of the expert and of the automatic pipeline are achieved for the average bone density with 0.999 (Pearson) and 0.996 (Spearman's rank) with [Formula: see text] for all correlations. A qualitative assessment of the network predictions and the manual annotations yields a 65.9% probability that the expert favors the network predictions. Further, the steps to integrate the pipeline into the clinical workflow are shown. In order to make these workflow improvements available to others, we openly share the code of this work.

A Virtual Reality-Based App to Educate Health Care Professionals and Medical Students About Inflammatory Arthritis: Feasibility Study.

BACKGROUND: Inflammatory arthritides (IA) such as rheumatoid arthritis or psoriatic arthritis are disorders that can be difficult to comprehend for health professionals and students in terms of the heterogeneity of clinical symptoms and pathologies. New didactic approaches using innovative technologies such as virtual reality (VR) apps could be helpful to demonstrate disease manifestations as well as joint pathologies in a more comprehensive manner. However, the potential of using a VR education concept in IA has not yet been evaluated. OBJECTIVE: We evaluated the feasibility of a VR app to educate health care professionals and medical students about IA. METHODS: We developed a VR app using data from IA patients as well as 2D and 3D-visualized pathological joints from X-ray and computed tomography-generated images. This VR app (Rheumality) allows the user to interact with representative arthritic joint and bone pathologies of patients with IA. In a consensus meeting, an online questionnaire was designed to collect basic demographic data (age, sex); profession of the participants; and their feedback on the general impression, knowledge gain, and potential areas of application of the VR app. The VR app was subsequently tested and evaluated by health care professionals (physicians, researchers, and other professionals) and medical students at predefined events (two annual rheumatology conferences and academic teaching seminars at two sites in Germany). To explore associations between categorical variables, the χ2 or Fisher test was used as appropriate. Two-sided P values ≤.05 were regarded as significant. RESULTS: A total of 125 individuals participated in this study. Among them, 56% of the participants identified as female, 43% identified as male, and 1% identified as nonbinary; 59% of the participants were 18-30 years of age, 18% were 31-40 years old, 10% were 41-50 years old, 8% were 51-60 years old, and 5% were 61-70 years old. The participants (N=125) rated the VR app as excellent, with a mean rating of 9.0 (SD 1.2) out of 10, and many participants would recommend use of the app, with a mean recommendation score of 3.2 (SD 1.1) out of 4. A large majority (120/125, 96.0%) stated that the presentation of pathological bone formation improves understanding of the disease. We did not find any association between participant characteristics and evaluation of the VR experience or recommendation scores. CONCLUSIONS: The data show that IA-targeting innovative teaching approaches based on VR technology are feasible.

Accuracy, patient-perceived usability, and acceptance of two symptom checkers (Ada and Rheport) in rheumatology: interim results from a randomized controlled crossover trial.

BACKGROUND: Timely diagnosis and treatment are essential in the effective management of inflammatory rheumatic diseases (IRDs). Symptom checkers (SCs) promise to accelerate diagnosis, reduce misdiagnoses, and guide patients more effectively through the health care system. Although SCs are increasingly used, there exists little supporting evidence. OBJECTIVE: To assess the diagnostic accuracy, patient-perceived usability, and acceptance of two SCs: (1) Ada and (2) Rheport. METHODS: Patients newly presenting to a German secondary rheumatology outpatient clinic were randomly assigned in a 1:1 ratio to complete Ada or Rheport and consecutively the respective other SCs in a prospective non-blinded controlled randomized crossover trial. The primary outcome was the accuracy of the SCs regarding the diagnosis of an IRD compared to the physicians' diagnosis as the gold standard. The secondary outcomes were patient-perceived usability, acceptance, and time to complete the SC. RESULTS: In this interim analysis, the first 164 patients who completed the study were analyzed. 32.9% (54/164) of the study subjects were diagnosed with an IRD. Rheport showed a sensitivity of 53.7% and a specificity of 51.8% for IRDs. Ada's top 1 (D1) and top 5 disease suggestions (D5) showed a sensitivity of 42.6% and 53.7% and a specificity of 63.6% and 54.5% concerning IRDs, respectively. The correct diagnosis of the IRD patients was within the Ada D1 and D5 suggestions in 16.7% (9/54) and 25.9% (14/54), respectively. The median System Usability Scale (SUS) score of Ada and Rheport was 75.0/100 and 77.5/100, respectively. The median completion time for both Ada and Rheport was 7.0 and 8.5 min, respectively. Sixty-four percent and 67.1% would recommend using Ada and Rheport to friends and other patients, respectively. CONCLUSIONS: While SCs are well accepted among patients, their diagnostic accuracy is limited to date. TRIAL REGISTRATION: DRKS.de, DRKS00017642 . Registered on 23 July 2019.

Advanced machine learning for predicting individual risk of flares in rheumatoid arthritis patients tapering biologic drugs.

BACKGROUND: Biological disease-modifying anti-rheumatic drugs (bDMARDs) can be tapered in some rheumatoid arthritis (RA) patients in sustained remission. The purpose of this study was to assess the feasibility of building a model to estimate the individual flare probability in RA patients tapering bDMARDs using machine learning methods. METHODS: Longitudinal clinical data of RA patients on bDMARDs from a randomized controlled trial of treatment withdrawal (RETRO) were used to build a predictive model to estimate the probability of a flare. Four basic machine learning models were trained, and their predictions were additionally combined to train an ensemble learning method, a stacking meta-classifier model to predict the individual flare probability within 14 weeks after each visit. Prediction performance was estimated using nested cross-validation as the area under the receiver operating curve (AUROC). Predictor importance was estimated using the permutation importance approach. RESULTS: Data of 135 visits from 41 patients were included. A model selection approach based on nested cross-validation was implemented to find the most suitable modeling formalism for the flare prediction task as well as the optimal model hyper-parameters. Moreover, an approach based on stacking different classifiers was successfully applied to create a powerful and flexible prediction model with the final measured AUROC of 0.81 (95%CI 0.73-0.89). The percent dose change of bDMARDs, clinical disease activity (DAS-28 ESR), disease duration, and inflammatory markers were the most important predictors of a flare. CONCLUSION: Machine learning methods were deemed feasible to predict flares after tapering bDMARDs in RA patients in sustained remission.

Machine Learning Electronic Health Record Identification of Patients with Rheumatoid Arthritis: Algorithm Pipeline Development and Validation Study.

BACKGROUND: Financial codes are often used to extract diagnoses from electronic health records. This approach is prone to false positives. Alternatively, queries are constructed, but these are highly center and language specific. A tantalizing alternative is the automatic identification of patients by employing machine learning on format-free text entries. OBJECTIVE: The aim of this study was to develop an easily implementable workflow that builds a machine learning algorithm capable of accurately identifying patients with rheumatoid arthritis from format-free text fields in electronic health records. METHODS: Two electronic health record data sets were employed: Leiden (n=3000) and Erlangen (n=4771). Using a portion of the Leiden data (n=2000), we compared 6 different machine learning methods and a naïve word-matching algorithm using 10-fold cross-validation. Performances were compared using the area under the receiver operating characteristic curve (AUROC) and the area under the precision recall curve (AUPRC), and F1 score was used as the primary criterion for selecting the best method to build a classifying algorithm. We selected the optimal threshold of positive predictive value for case identification based on the output of the best method in the training data. This validation workflow was subsequently applied to a portion of the Erlangen data (n=4293). For testing, the best performing methods were applied to remaining data (Leiden n=1000; Erlangen n=478) for an unbiased evaluation. RESULTS: For the Leiden data set, the word-matching algorithm demonstrated mixed performance (AUROC 0.90; AUPRC 0.33; F1 score 0.55), and 4 methods significantly outperformed word-matching, with support vector machines performing best (AUROC 0.98; AUPRC 0.88; F1 score 0.83). Applying this support vector machine classifier to the test data resulted in a similarly high performance (F1 score 0.81; positive predictive value [PPV] 0.94), and with this method, we could identify 2873 patients with rheumatoid arthritis in less than 7 seconds out of the complete collection of 23,300 patients in the Leiden electronic health record system. For the Erlangen data set, gradient boosting performed best (AUROC 0.94; AUPRC 0.85; F1 score 0.82) in the training set, and applied to the test data, resulted once again in good results (F1 score 0.67; PPV 0.97). CONCLUSIONS: We demonstrate that machine learning methods can extract the records of patients with rheumatoid arthritis from electronic health record data with high precision, allowing research on very large populations for limited costs. Our approach is language and center independent and could be applied to any type of diagnosis. We have developed our pipeline into a universally applicable and easy-to-implement workflow to equip centers with their own high-performing algorithm. This allows the creation of observational studies of unprecedented size covering different countries for low cost from already available data in electronic health record systems.

Bone Mass, Bone Microstructure and Biomechanics in Patients with Hand Osteoarthritis.

The impact of primary hand osteoarthritis (HOA) on bone mass, microstructure, and biomechanics in the affected skeletal regions is largely unknown. HOA patients and healthy controls (HCs) underwent high-resolution peripheral quantitative computed tomography (HR-pQCT). We measured total, trabecular, and cortical volumetric bone mineral densities (vBMDs), microstructural attributes, and performed micro-finite element analysis for bone strength. Failure load and scaled multivariate outcome matrices from distal radius and second metacarpal (MCP2) head measurements were analyzed using multiple linear regression adjusting for age, sex, and functional status and reported as adjusted Z-score differences for total and direct effects. A total of 105 subjects were included (76 HC: 46 women, 30 men; 29 HOA: 23 women, six men). After adjustment, HOA was associated with significant changes in the multivariate outcome matrix from the MCP2 head (p < .001) (explained by an increase in cortical vBMD (Δz = 1.07, p = .02) and reduction in the trabecular vBMD (Δz = -0.07, p = .09). Distal radius analysis did not show an overall effect of HOA; however, there was a gender-study group interaction (p = .044) explained by reduced trabecular vBMD in males (Δz = -1.23, p = .02). HOA was associated with lower failure load (-514 N; 95%CI, -1018 to -9; p = 0.05) apparent in males after adjustment for functional status. HOA is associated with reduced trabecular and increased cortical vBMD in the MCP2 head and a reduction in radial trabecular vBMD and bone strength in males. Further investigations of gender-specific changes of bone architecture in HOA are warranted. © 2020 The Authors. Journal of Bone and Mineral Research published by American Society for Bone and Mineral Research.

Effect of disease-modifying anti-rheumatic drugs on bone structure and strength in psoriatic arthritis patients.

OBJECTIVES: To address whether the use of methotrexate (MTX) and biological disease-modifying anti-rheumatic drugs (bDMARDs) impacts bone structure and biomechanical properties in patients with psoriatic arthritis (PsA). METHODS: This is a cross-sectional study in PsA patients receiving no DMARDs, MTX, or bDMARDs. Volumetric bone mineral densities (vBMDs), microstructural parameters, and biomechanical properties (stiffness/failure load) were determined by high-resolution peripheral quantitative CT and micro-finite element analysis in the respective groups. Bone parameters were compared between PsA patients with no DMARDs and those receiving any DMARDs, MTX, or bDMARDs, respectively. RESULTS: One hundred sixty-five PsA patients were analyzed, 79 received no DMARDs, 86 received DMARDs, of them 52 bDMARDs (TNF, IL-17- or IL-12/23 inhibitors) and 34 MTX. Groups were balanced for age, sex, comorbidities, functional index, and bone-active therapy, while disease duration was longest in the bDMARD group (7.8 ± 7.4 years), followed by the MTX group (4.6 ± 7.4) and the no-DMARD group (2.9 ± 5.2). No difference in bone parameters was found between the no-DMARD group and the MTX group. In contrast, the bDMARD group revealed significantly higher total (p = 0.001) and trabecular vBMD (p = 0.005) as well as failure load (p = 0.012) and stiffness (p = 0.012). In regression models, age and bDMARDs influenced total vBMD, while age, sex, and bDMARDs influenced failure load and stiffness. CONCLUSION: Despite longer disease duration, bDMARD-treated PsA patients benefit from higher bone mass and better bone strength than PsA patients receiving MTX or no DMARDs. These data support the concept of better control of PsA-related bone disease by bDMARDs.