Background: Disease-modifying antirheumatic drugs (bDMARDs) have shown efficacy in treating Rheumatoid Arthritis (RA). Predicting treatment outcomes for RA is crucial as approximately 30% of patients do not respond to bDMARDs and only half achieve a sustained response. This study aims to leverage machine learning to predict both initial response at 6 months and sustained response at 12 months using baseline clinical data. Methods: Baseline clinical data were collected from 154 RA patients treated at the University Hospital in Erlangen, Germany. Five machine learning models were compared: Extreme Gradient Boosting (XGBoost), Adaptive Boosting (AdaBoost), K-nearest neighbors (KNN), Support Vector Machines (SVM), and Random Forest. Nested cross-validation was employed to ensure robustness and avoid overfitting, integrating hyperparameter tuning within its process. Results: XGBoost achieved the highest accuracy for predicting initial response (AUC-ROC of 0.91), while AdaBoost was the most effective for sustained response (AUC-ROC of 0.84). Key predictors included the Disease Activity Score-28 using erythrocyte sedimentation rate (DAS28-ESR), with higher scores at baseline associated with lower response chances at 6 and 12 months. Shapley additive explanations (SHAP) identified the most important baseline features and visualized their directional effects on treatment response and sustained response. Conclusions: These findings can enhance RA treatment plans and support clinical decision-making, ultimately improving patient outcomes by predicting response before starting medication.

The purpose of this study was to assess the accuracy and performance of a new handheld ultrasound (HHUS) machine in comparison to a conventional cart-based sonographic machine in patients with inflammatory arthritis (IA). IA patients with at least one tender and swollen joint count were enrolled. US was performed on the clinically affected joints using a cart-based sonographic device (Samsung HS40) and a HHUS device (Butterfly iQ). One blinded reader scored all images for the presence of erosions, bony enlargement, synovial hypertrophy, joint effusion, bursitis, tenosynovitis, and enthesitis. Synovitis was graded (B mode and power Doppler (PD)) by the 4-level EULAR-OMERACT scale. To avoid bias by the blinded reader, we included 67 joints of two healthy volunteers in the evaluation. We calculated the overall concordance and the concordance by type of joint and pathological finding. We also measured the time required for the US examination per joint with both devices. Thirty-two patients (20 with RA, 10 with PsA, and one each with gout and SLE-associated arthritis) were included, and 186 joints were examined. The overall raw concordance in B mode was 97% (κappa 0.90, 95% CI (0.89, 0.94)). In B mode, no significant differences were found in relation to type of joint or pathological finding examined. The PD mode of the HHUS device did not detect any PD signal, whereas the cart-based device detected a PD signal in 61 joints (33%). The portable device did not offer any time savings compared to the cart-based device (47.0 versus 46.3 s). The HHUS device was accurate in the assessment of structural damage and inflammation in patients with IA, but only in the B mode. Significant improvements are still needed for HHUS to reliably demonstrate blood flow detection in PD mode.

Rheumatoid arthritis (RA) is a chronic autoimmune disease marked by joint inflammation and progressive disability. While biological disease-modifying antirheumatic drugs (bDMARDs) have significantly improved disease control, predicting individual treatment response remains clinically challenging. This study presents a machine learning approach to predict 12-month disease activity, measured by DAS28-CRP, in RA patients beginning bDMARD therapy. We trained and evaluated eight regression models, including Ridge, Lasso, Support Vector Regression, and XGBoost, using baseline clinical features from 154 RA patients treated at University Hospital Erlangen. A rigorous nested cross-validation strategy was applied for internal model selection and validation. Importantly, model generalizability was assessed using an independent external dataset from the Austrian BioReg registry, which includes a more diverse, real-world RA patient population from across multiple clinical sites. The Ridge regression model achieved the best internal performance (MAE: 0.633, R²: 0.542) and showed strong external validity when applied to unseen BioReg data (MAE: 0.678, R²: 0.491). These results indicate robust cross-cohort generalization. By predicting continuous DAS28-CRP scores instead of binary remission labels, our approach supports flexible, individualized treatment planning based on local or evolving clinical thresholds. This work demonstrates the feasibility and clinical value of externally validated, data-driven tools for precision treatment planning in RA.

Purpose: Hand function is considered being critical in rheumatoid (RA) and psoriatic arthritis (PsA) patients; however, comprehensive assessment of hand function is rarely done in patients with inflammatory arthritis. Methods: In this cross-sectional study, RA and PsA patients and healthy controls (HC) were assessed for hand function using isometric grip strength (dynamometer), dynamic grip strength (vigorimeter), and fine-motor skills (Moberg Picking-Up Test). Between-group differences and the associations with disease activity (Disease Activity Score 28 (DAS28), RA, Disease Activity in Psoriasis Arthritis (DAPSA), PsA) were analyzed using linear mixed-effects models, with age as a covariate and participants’ ID as a random intercept term. All analyses were stratified by sex and additionally employed for a subgroup without concurrent hand inflammation (no tender and swollen joints, no ultrasound signs of inflammation). Results: A total of 225 participants were included (76 PsA, 73 RA, and 76 HC). RA and PsA patients showed a similar reduction in isometric and dynamic grip strength and fine-motor skills compared with HC. Impairment of hand function was particularly pronounced in female RA and PsA patients, whereas grip strength was better maintained in males. A one-unit higher DAS28 or DAPSA was associated with lower dynamic grip strength for all RA and PsA patients, and one-unit higher DAPSA was additionally associated with lower fine-motor skills in female PsA patients. Female RA and PsA patients showed impairment in all measures of hand function even without concurrent hand inflammation, whereas among males, only PsA patients showed fine-motor skill impairment. Conclusions: Hand function is impaired to a similar extent in RA and PsA patients, more pronounced in female patients, and also found in the absence of concurrent hand inflammation.