Detection and Grading of Radiographic Hand Osteoarthritis Using an Automated Machine Learning Platform.

OBJECTIVE: Automated machine learning (autoML) platforms allow health care professionals to play an active role in the development of machine learning (ML) algorithms according to scientific or clinical needs. The aim of this study was to develop and evaluate such a model for automated detection and grading of distal hand osteoarthritis (OA). METHODS: A total of 13,690 hand radiographs from 2,863 patients within the Swiss Cohort of Quality Management (SCQM) and an external control data set of 346 non-SCQM patients were collected and scored for distal interphalangeal OA (DIP-OA) using the modified Kellgren/Lawrence (K/L) score. Giotto (Learn to Forecast [L2F]) was used as an autoML platform for training two convolutional neural networks for DIP joint extraction and subsequent classification according to the K/L scores. A total of 48,892 DIP joints were extracted and then used to train the classification model. Heatmaps were generated independently of the platform. User experience of a web application as a provisional user interface was investigated by rheumatologists and radiologists. RESULTS: The sensitivity and specificity of this model for detecting DIP-OA were 79% and 86%, respectively. The accuracy for grading the correct K/L score was 75%, with a κ score of 0.76. The accuracy per DIP-OA class differed, with 86% for no OA (defined as K/L scores 0 and 1), 71% for a K/L score of 2, 46% for a K/L score of 3, and 67% for a K/L score of 4. Similar values were obtained in an independent external test set. Qualitative and quantitative user experience testing of the web application revealed a moderate to high demand for automated DIP-OA scoring among rheumatologists. Conversely, radiologists expressed a low demand, except for the use of heatmaps. CONCLUSION: AutoML platforms are an opportunity to develop clinical end-to-end ML algorithms. Here, automated radiographic DIP-OA detection is both feasible and usable, whereas grading among individual K/L scores (eg, for clinical trials) remains challenging.

Brief Report: Influence of Disease Activity in Rheumatoid Arthritis on Radiographic Progression of Concomitant Interphalangeal Joint Osteoarthritis.

OBJECTIVE: Distal interphalangeal (DIP) joints are commonly considered to be unaffected by rheumatoid arthritis (RA). Despite synovitis and bone marrow edema being associated with radiographic progression in hand osteoarthritis (OA) and hand RA, radiographic courses differ substantially. This study was undertaken to analyze incidence and progression of radiographically evident DIP joint OA in RA patients, in relation to RA activity and patient characteristics. METHODS: In sequential radiographs of 1,988 RA patients in the Swiss Clinical Quality Management in Rheumatic Diseases registry, we evaluated and scored 15,904 DIP joints. Scoring was based on the presence of central erosions and subchondral sclerosis and on the severity of osteophytes and joint space narrowing, according to the modified Kellgren/Lawrence (K/L) grade. The presence of DIP joint OA was defined as ≥1 joint with a K/L grade of ≥2, and progression was defined as an increase in a summed K/L grade. Adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated. RESULTS: The median follow-up time was 4.5 years (interquartile range 3.1-7.0), and the mean ± SD age was 56.1 ± 11.1 years. DIP joint OA was present in 60% of patients at baseline. Higher mean age (OR 1.09 [95% CI 1.08-1.10]), female sex (OR 1.37 [95% CI 1.08-1.74]), and higher mean body mass index (OR 1.03 [95% CI 1.00-1.06]) were associated with the presence of DIP joint OA, but neither the presence of anti-citrullinated protein antibodies (ACPAs) (OR 0.72 [95% CI 0.50-1.03]) nor the presence of rheumatoid factor (OR 1.01 [95% CI 0.74-1.38]) were associated with it. Disease Activity Score using the erythrocyte sedimentation rate and metacarpophalangeal (MCP) joint erosions were not associated with DIP joint OA progression. RA disease duration had no relevant effect size associated with DIP joint OA progression (OR 0.97 [95% CI 0.96-0.99]). CONCLUSION: Known risk factors for DIP joint OA were replicated in patients with RA. The observation that RA activity, the presence of ACPA, and MCP joint erosions were not associated with the prevalence or progression of DIP joint OA indicates that there are distinct roles of inflammation in the pathogenesis of RA and DIP joint OA.

Association of nutritional risk and adverse medical outcomes across different medical inpatient populations.

OBJECTIVE: The aim of this study was to examine the prevalence of nutritional risk and its association with multiple adverse clinical outcomes in a large cohort of acutely ill medical inpatients from a Swiss tertiary care hospital. METHODS: We prospectively followed consecutive adult medical inpatients for 30 d. Multivariate regression models were used to investigate the association of the initial Nutritional Risk Score (NRS 2002) with mortality, impairment in activities of daily living (Barthel Index <95 points), hospital length of stay, hospital readmission rates, and quality of life (QoL; adapted from EQ5 D); all parameters were measured at 30 d. RESULTS: Of 3186 patients (mean age 71 y, 44.7% women), 887 (27.8%) were at risk for malnutrition with an NRS ≥3 points. We found strong associations (odds ratio/hazard ratio [OR/HR], 95% confidence interval [CI]) between nutritional risk and mortality (OR/HR, 7.82; 95% CI, 6.04-10.12), impaired Barthel Index (OR/HR, 2.56; 95% CI, 2.12-3.09), time to hospital discharge (OR/HR, 0.48; 95% CI, 0.43-0.52), hospital readmission (OR/HR, 1.46; 95% CI, 1.08-1.97), and all five dimensions of QoL measures. Associations remained significant after adjustment for sociodemographic characteristics, comorbidities, and medical diagnoses. Results were robust in subgroup analysis with evidence of effect modification (P for interaction < 0.05) based on age and main diagnosis groups. CONCLUSION: Nutritional risk is significant in acutely ill medical inpatients and is associated with increased medical resource use, adverse clinical outcomes, and impairments in functional ability and QoL. Randomized trials are needed to evaluate evidence-based preventive and treatment strategies focusing on nutritional factors to improve outcomes in these high-risk patients.