P6485Biventricular imaging markers to predict outcome in non-compaction cardiomyopathy: a machine learning study

Aims: Left ventricular non-compaction cardiomyopathy (LVNC) is a genetic heart disease, which main clinical manifestations are heart failure, arrhythmias and embolic events. Its prevalence is increasing due to improvements in cardiac imaging methods, such as echocardiography (echo) and cardiac magnetic resonance imaging (CMRI). However, imaging parameters related to prognosis are poorly known. The goal of this study was to analyze a large set of echo and CMRI parameters using machine learning techniques, in order to find an association between these imaging markers and clinical outcome in a long-term follow-up of LVNC patients. Methods: One hundred and eight LVNC patients (38.3 +15.5 years, 48% male) diagnosed by echo and CMRI criteria were recruited for this study and were followed for 5.83 + 3.9 years, during which combined “hard events” (death, cardiac transplantation, heart failure hospitalization, aborted sudden cardiac death, complex ventricular arrhythmias, and embolisms) were registered. CMRI and echo parameters were extracted by expert users and subsequently analyzed via a supervised machine learning methodology using “cross-validation”.