Background: Diagnosing systemic lupus erythematosus (SLE) is often challenging and delayed, contributing to increased disease burden. The SLE Risk Probability Index (SLEPRI) is a machine-learning-based tool developed to support accurate and timely SLE classification [1].

Objectives: To assess the performance of SLEPRI in classifying patients with SLE and stratify them according to the risk of disease flares.

Methods: This retrospective study included consecutively registered patients with physician-diagnosed SLE between 1990 and 2000 at Universitätsklinikum Erlangen, along with non-SLE controls diagnosed with other rheumatic diseases according to respective diagnostic criteria and who were evaluated during the same period and in the same institution. Baseline clinical and serological parameters were analyzed to determine fulfillment of the 2019 EULAR/ACR classification criteria and to calculate the SLE Risk Probability Index (SLEPRI) across its 14 domains [1]. Patients with a SLEPRI score >7 were classified as having SLE. Disease flares within three years of diagnosis, defined by the Safety of Estrogen in Lupus Erythematosus-SLE Disease Activity Index Flare Index (SELENA-SLEDAI Flare Index), were recorded.

Results: A total of 450 patients with diagnosis of rheumatic diseases (65.3% SLE, 11.1% Ankylosing Spondylitis, 10.9% Rheumatoid arthritis, 8.9% Psoriatic arthritis, 1.8% Adult-Onset Still’s disease, 1.1% systemic sclerosis, 0.2% Behcet disease, 0.2% Gout and 0.4% Inflammatory bowel disease-related arthritis) were included. Among the SLE patients, 249 (85.0%) were females, with a median (IQR) age of diagnosis of 31 (23, 45) years. 246 (83.7%) of SLE patients fulfilled the 2019 EULAR/ACR classification criteria and 255 (86.7%) were classified as SLE according to the SLERPI, with a mean (SD) score of 12.0 (3.5) points. The most frequent SLEPRI clinical features were arthritis (n=146, 49.7%), subacute cutaneous lupus erythematosus/discoid lupus erythematosus (n=95, 32.3%) and malar rash/maculopapular rash (n=90, 30.6%). SLERPI showed high discriminating capacity for SLE against competing rheumatological diseases, with a sensitivity and specificity of 86.7% and 97.4%, respectively. Among 258 SLE patients with complete medical records within three years of diagnosis, 115 (44.6%) and 82 (31.8%) patients experienced disease flare (any) and severe disease flare, respectively. Patients who experienced any disease flare within three years had significantly higher baseline SLEPRI scores at diagnosis compared with those without flares (median [IQR]: 9.0 [7.5, 12.5] vs 11 [9.0, 14.5], p = 0.010).

Conclusions: SLEPRI effectively distinguished SLE patients from those with other rheumatic diseases in an independent cohort. Moreover, higher baseline SLEPRI scores were significantly associated with an increased risk of disease flares within three years of diagnosis.

REFERENCES: [1] Adamichou C, Genitsaridi I, Nikolopoulos D, Nikoloudaki M, Repa A, Bortoluzzi A, Fanouriakis A, Sidiropoulos P, Boumpas DT, Bertsias GK. Lupus or not? SLE Risk Probability Index (SLERPI): a simple, clinician-friendly machine learning-based model to assist the diagnosis of systemic lupus erythematosus. Ann Rheum Dis. 2021 Jun;80(6):758-766. doi: 10.1136/annrheumdis-2020-219069. Epub 2021 Feb 10. PMID: 33568388; PMCID: PMC8142436.

Acknowledgments: NIL.

Disclosure of Interests: Andrea Zoli: None declared, Shirley C.W. Chan: None declared, Seda Nur Aydogdu: None declared, Maria Antonietta D’ Agostino NOVARTIS, Paraskevi Chasani: None declared, Christina Bergmann: None declared, Georg Schett: None declared, Panagiotis Garantziotis: None declared.