High-Frequency Irreversible Electroporation: Optimum Parameter Prediction via Machine-Learning

Abstract

The adoption of high-frequency irreversible electroporation in various medical treatments is becoming increasingly prevalent. There is currently a special focus on its applications in oncology, offering new perspectives in terms of treatable tumor types and treatment effectiveness. A multitude of parameters can influence the efficiency and effectiveness of high-frequency irreversible electroporation procedures, with the selection of suitable electrodes and possible prediction of ablated area as interesting examples. In this paper, we demonstrate that machine-learning strategies, specifically neural networks, provide an appropriate approach for optimizing the choice of some electrode characteristics, and predicting the ablation area, this being quite useful in high-frequency electroporation applications in oncology. This possibility, in turn, may lead to superior results in high-frequency irreversible electroporation, and to a significant reduction of the time required for achieving them.