DIFFERENT FATES AND ROUTES OF CHRONIC HEART FAILURE, AN ARTIFICIAL INTELLIGENT BASED ANALYSIS OF THE MECKI SCORE DATASET

Abstract Background Individual prognostic assessment and disease evolution pathways in chronic heart failure (HF) are unmet needs. The application of unsupervised learning methodologies could help at: a) identification of HF patient phenotypes; b) identification of the evolution states and progression in each phenotype and c) assessment of risk of adverse event. Methods We retrospectively analyzed the MECKI registry including 7948 HF patients. After preliminary preprocessing, we selected 4876 patients with a minimum 2–year follow–up. We implemented a Topological Data Analysis (TDA), based on 43 variables derived from clinical, biochemical, cardiac ultrasound and exercise evaluations, to identify several patients’ clusters. Thereafter, we used the trajectory analysis to describe the evolution of HF states, which is able to identify bifurcation points, characterized by different follow–up paths, as well as stopping points, i.e. specific end–stages conditions of the disease. Finally, we conducted a survival analysis to generate 5–year time–to–event curves using as study endpoint the composite of cardiovascular death, LVAD or urgent heart transplant. Findings were validated on internal (n=527) and external (n=777) populations. Results Nineteen patient clusters were identified by TDA. Trajectory analysis revealed a path characterized by 3 bifurcation points with 2 different follow–up path each and 4 stopping points. Clusters survival rate varied from 44% to 100% at 2 years and from 20% to 100% at 5 years, respectively. Finally, we compared the event frequency at a 5–year follow–up for each study cohort cluster with those in the validation cohorts (R = 0.94 and R = 0.84, p < 0.001, for internal and external cohort, respectively). Conclusions Each HF phenotype has a specific disease progression and prognosis. The bifurcation points need attention since are followed by different disease routes as the 4 different disease end–stages which are characterized by different prognosis. These findings allow to individualize HF patient evolutions and to tailor assessment.