Cardiac pathologies, such as sudden cardiac death (SCD), constitute one of the leading causes of mortality worldwide. Consequently, assessing the underlying mechanisms of these conditions becomes essential. The risk stratification of ventricular cardiac arrhythmias, often preceded by SCD, is specifically the focus of this study. In this context, this work presents a pilot computational approach that aims to optimize virtual electrophysiological investigations of cardiac arrhythmias. Based on non-invasively obtained data, simulations are conducted on 2D geometries of a patient with dilated cardiomyopathy and on reduced segments thereof, located proximal to fibrotic regions, in order to reproduce cases of unidirectional block. The results obtained indicate that from reduced meshes, which exhibit significantly lower computational complexity than entire meshes, it is possible to effectively identify patients prone to arrhythmiasy. When compared to the study on the complete tissue, the adopted approach achieves a 5-fold increase in computational efficiency. Thus, it emerges as an important preliminary tool for the evaluation of cardiac arrhythmias.