Abstract 4117883: Long noncoding RNAs and machine learning to improve cardiovascular outcomes of COVID-19

• DOI: 10.1161/circ.150.suppl_1.4117883
• Publication Date: 2024-11-13T13:17:35Z
• Authors: Kanita Karaduzovi...; Muhamed Adilović; Lu Zhang; Andrew Lumley; Pranay Shah; Muhammad Shoaib; Venkata Satagopam; Prashant Srivastava; Costanza Emanueli; Simona Greco; Teresa Padro; Pedro Domingo; Mitja Lustrek; Markus Scholz; Maciej Rosolowski; Marko Jordan; Bettina Benczik; Bence Agg; Peter Ferdinandy; Andrew Baker; Guy Fagherazzi; Markus Ollert; Joanna Michel; Gabriel Sanchez; Hueseyin Firat; Timo Brandenburger; Fabio Martelli; Lina Badimon; Yvan Devaux
• Keywords: 2019-20 coronavirus outbreak

Introduction/Background: Cardiovascular symptoms appear in a high proportion of patients the few months following severe SARS-CoV-2 infection. Non-invasive methods to predict disease severity could help personalizing healthcare and reducing occurrence these symptoms. Research Questions/Hypothesis: We hypothesized that blood long noncoding RNAs (lncRNAs) machine learning (ML) COVID-19 severity. Goals/Aims: To develop model based on lncRNAs ML for predicting Methods/Approach: Expression data 2906 were obtained by targeted sequencing plasma samples collected at baseline from four independent cohorts, totaling 564 patients. Patients aged 18+ recruited 2020 2023 PrediCOVID cohort (n=162; Luxembourg), COVID19_OMICS-COVIRNA (n=100, Italy), TOCOVID (n=233, Spain), MiRCOVID (n=69, Germany). The study complied with Declaration Helsinki. Cohorts approved ethics committees signed an informed consent. Results/Data: After curation pre-processing, 463 complete datasets included further analysis, representing 101 (in-hospital death or ICU admission) 362 stable (no hospital admission but not ICU). Feature selection Boruta, random forest-based method, identified age five (LINC01088-201, FGDP-AS1, LINC01088-209, AKAP13, novel lncRNA) associated severity, which used build predictive models using six algorithms. A naïve Bayes predicted AUC 0.875 [0.868-0.881] accuracy 0.783 [0.775-0.791]. Conclusion: developed including This improve patients’ management cardiovascular outcomes.