¹University of Lübeck, Department of Rheumatology and Clinical Immunology, Lübeck
²Rheuma-Liga Schleswig-Holstein e.V., Kiel
³Christian-Albrechts University Kiel, Institute of Experimental Medicine, Kiel
⁴Rheumapraxis Kiel, Kiel
⁵St. Franziskus MVZ Am Hang, Harrislee
⁶Gemeinschaftspraxis Dres. med. Bolouri & Schwab, Kiel

Text

Introduction: Epidemiological data indicate an elevated prevalence of inflammatory rheumatic diseases (IRD) in Germany. In particular, the most common IRD, rheumatoid arthritis (RA), showed an increase, currently between 0.42% and 1.85% [1]. Therefore, new models of early consultation and new diagnostic and therapeutic approaches are needed since outcome and quality of life are associated with early intervention within 6 weeks after symptom onset [2]. However, the limited number of specialists in rheumatology predicts waiting times for a first rheumatology visit, often over 3 months [2]. Particularly in remote non-urban areas such as Schleswig-Holstein (SH), there are differences in the availability of rheumatologic care between urban and rural areas. Autonomic systems that can objectively detect inflammatory joint pain as the first rheumatologic manifestation could improve the early initial diagnosis of RA in this care setting [3]. This study aimed to evaluate the real-world use of fully automated robotic arthrosonography (ARTHUR) with AI analysis (DIANA) to detect RA activity.

Methods: Between October 7 and October 12, 2024, the Rheuma-Liga Schleswig-Holstein e.V. initiated a mobile contact point (“Rheuma-Truck”) in different communities of SH, especially on the west coast, using ARTHUR to detect arthrosonographic signs of inflammatory arthritis according to the OMERACT-EULAR synovitis criteria [4]. Data were collected from the participants (age, sex, morning stiffness, joint pain and swelling score, and pain on a numeric analogue scale). C-reactive protein was measured using an AFIAS POC system. Health Assessment Questionnaires (HAQ-II) and Disease Activity Score-28 (DAS28) were administered. According to the DIANA AI analysis, participants with positive results were referred directly to a rheumatology specialist. Participants with elevated CRP and/or severe symptoms were referred to a rheumatology specialist after automated arthrosonography, even if the result was negative.

Results: A total of 248 participants were screened with ARTHUR. Of this total, 70 individuals were excluded due to missing data, and one ARTHUR measurement failed. One hundred seventy-two participants were divided into two groups: persons without a history of IRD (Non-IRD, n=150, age: 65.5±13.1 yr, 74.7% female, DAS28: 3.7±1.1, HAQ II: 1.2±0.9, CRP 7. 8±7.7 mg/I) and participants with IRD (IRD, RA n=18, polymyalgia rheumatica n=6; systemic sclerosis n=1, systemic lupus erythematosus n=1, granulomatosis with polyangiitis n=1, psoriatic arthritis n=1; age 66.0±13.1 yr, 92.6% female, DAS28 4.1±1.1, HAQ II 1.4±1.2, CRP 7.8±7.3 mg/I). In the NON-IRD group, 18 individuals were found positive for inflammatory signs using ARTHUR. Of these, two participants were considered false positives by a rheumatology specialist. Only one person with a negative result by ARTHUR was detected with false negatives (ROC-AUC: 0.9953, sensitivity 98.53, specificity 100.00%, P<0.0001, number needed to diagnose: 1.057). Only patients with RA were included in the IRD group. Here, five patients with active disease were detected by ARTHUR, with no false positives. In this group, there were no false negatives by a rheumatologist (ROC-AUC>: 1,000, sensitivity 100%, specificity 100%, P=0.0027).

Conclusion: Fully automatic robotic arthrosonography with AI analysis predicted high sensitivity and specificity for detecting inflammatory arthritis in patients with initial symptoms and RA patients suspected of relapse.

Literatur

[1] Albrecht K, Binder S, Minden K, Poddubnyy D, Regierer AC, Strangfeld A, Callhoff J. Systematisches Review zur Schätzung der Prävalenz entzündlich rheumatischer Erkrankungen in Deutschland [Systematic review to estimate the prevalence of inflammatory rheumatic diseases in Germany. German version]. Z Rheumatol. 2023 Nov;82(9):727-38. German. DOI: 10.1007/s00393-022-01305-2
[2] Braun J, Albrecht K, Callhoff J, Haase I, Krause A, Lakomek HJ, Meyer-Olson D, Schmale-Grede R, Wagner U, Zeidler J, Zinke S, Voormann A, Specker C; die Kommission Versorgung der DGRh. Rheumatologische Versorgung in Deutschland : Memorandum der Deutschen Gesellschaft für Rheumatologie und Klinische Immunologie 2024 [Rheumatological care in Germany : Memorandum of the German Society for Rheumatology and Clinical Immunology 2024]. Z Rheumatol. 2024 Aug;83(Suppl 2):249-84. German. DOI: 10.1007/s00393-024-01539-2
[3] Aplin Frederiksen B, Ammitzbøll Danielsen M, Berner Hammer H, Schultz Overgaard B, Weber A, Terslev L, Rajeeth Savarimuthu T, Just S. Automated Ultrasound System ARTHUR with AI Analysis DIANA Matches Expert Rheumatologist in Hand Joint Assessment of Rheumatoid Arthritis Patients. Arthritis Rheumatol. 2024; 76 (suppl 9). Available from: https://acrabstracts.org/abstract/automated-ultrasound-system-arthur-with-ai-analysis-diana-matches-expert-rheumatologist-in-hand-joint-assessment-of-rheumatoid-arthritis-patients/
[4] D'Agostino MA, Terslev L, Aegerter P, Backhaus M, Balint P, Bruyn GA, Filippucci E, Grassi W, Iagnocco A, Jousse-Joulin S, Kane D, Naredo E, Schmidt W, Szkudlarek M, Conaghan PG, Wakefield RJ. Scoring ultrasound synovitis in rheumatoid arthritis: a EULAR-OMERACT ultrasound taskforce-Part 1: definition and development of a standardised, consensus-based scoring system. RMD Open. 2017 Jul 11;3(1):e000428. DOI: 10.1136/rmdopen-2016-000428