Th166 - The Experience of the Artificial Intelligence Tools’ Use in Prediction of Long- COVID-19 in Armenia

Lyudmila Niazyan, MPH, MSc – Researcher, Chair of Professional Competences, Yerevan State Medical University after Mkhitar Heratsi; Garnik Avetisyan, Dr – Vice-Rector, Head of Chair, Chair of Radiology, Yerevan State Medical University after Mkhitar Heratsi; Lusine Simonyan, MD – Plasmapheresis specialist, Therapeutist, 1st University Hospital, Yerevan State Medical University after Mkhitar Heratsi

Abstract Text: Long COVID-19 is a condition that lasts for weeks, months and years following the initial SARS-CoV-2 infection with a broad symptomological spectrum. Persisting neuropsychiatric manifestations, including fatigue, dyspnoea, cognitive impairments, and disturbances in concentration, anxiety, and sleep, may endure beyond 6–12 months post-acute infection. The post-COVID condition is hypothesised to emanate from immune dysregulation and the presence of circulating "microclots" comprising amyloid, fibrin, and fibrinogen, with plasmapheresis demonstrating efficacy in their removal.

The advent of computational tools for the prognostication of deteriorative risk employs a specific scoring algorithm (0-4) predicted on clinical variables. Prioritisation of patients for administration of Therapeutic Plasma Exchange (TPE), this model leverages patient-specific data including the latest laboratory test results, age, body mass index (BMI), risk factors, underlying medical conditions, and clinical symptomatology. The algorithm integrates 30 laboratory parameters, encompassing albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, blood cells and a comprehensive inflammatory marker profile.

In the annum 2023, an investigation involving 198 subjects highlighted the model's utility across diverse symptomatologies including respiratory/allergic, neurological, cardiovascular, and musculoskeletal domains, accounting for 20-40% of presentations. For severe manifestations, a therapeutic regimen comprising up to seven TPE sessions over a six-month trajectory was administered. The utilisation of AI-driven prognostic tools in long-COVID-19 management has evidenced favour able outcomes in both clinical efficacy and patient quality of life, as well as had a pivotal role in physician decision-making process.