Rheumatoid arthritis (RA) is a heterogeneous disease with variable symptoms, prognosis, and treatment response, necessitating refined patient classification. We applied multimodal deep learning and clustering to identify distinct RA phenotypes using baseline clinical data from 1,387 patients in the Leiden Rheumatology clinic. Four Joint Involvement Patterns (JIP) emerged: foot-predominant arthritis, seropositive oligoarticular disease, seronegative hand arthritis, and polyarthritis. Findings were validated in clinical trial data (n = 307) and an independent secondary care cohort (n = 515). Clusters showed high stability and significant differences in remission rates (P = 0.007) and methotrexate failure (P < 0.001). JIP-hand patients had superior outcomes (particularly in ACPA-positive patients) versus JIP-foot (HR:0.37, P < 0.001) and JIP-poly (HR:0.33, P = 0.005), independent of baseline disease activity and clinical markers. Synovial histology analysis (n = 194) revealed distinct inflammatory patterns across clusters, hinting at different underlying biological mechanisms. These validated RA phenotypes based on joint involvement patterns may enable targeted research into disease mechanisms and personalized treatment strategies.

Objective: To facilitate patient disease subset and risk factor identification by constructing a pipeline which is generalizable, provides easily interpretable results, and allows replication by overcoming electronic health records (EHRs) batch effects. Material and Methods: We used 1872 billing codes in EHRs of 102 880 patients from 12 healthcare systems. Using tools borrowed from single-cell omics, we mitigated center-specific batch effects and performed clustering to identify patients with highly similar medical history patterns across the various centers. Our visualization method (PheSpec) depicts the phenotypic profile of clusters, applies a novel filtering of noninformative codes (Ranked Scope Pervasion), and indicates the most distinguishing features. Results: We observed 114 clinically meaningful profiles, for example, linking prostate hyperplasia with cancer and diabetes with cardiovascular problems and grouping pediatric developmental disorders. Our framework identified disease subsets, exemplified by 6 "other headache"clusters, where phenotypic profiles suggested different underlying mechanisms: migraine, convulsion, injury, eye problems, joint pain, and pituitary gland disorders. Phenotypic patterns replicated well, with high correlations of ≥0.75 to an average of 6 (2-8) of the 12 different cohorts, demonstrating the consistency with which our method discovers disease history profiles. Discussion: Costly clinical research ventures should be based on solid hypotheses. We repurpose methods from single-cell omics to build these hypotheses from observational EHR data, distilling useful information from complex data. Conclusion: We establish a generalizable pipeline for the identification and replication of clinically meaningful (sub)phenotypes from widely available high-dimensional billing codes. This approach overcomes datatype problems and produces comprehensive visualizations of validation-ready phenotypes.