Optimization and decision-making of novel laser-induced thermal therapy for deep-lying tumor based on multi-objective genetic algorithm and three-way decisions method

Abstract

In this paper, we develop novel laser-induced thermal therapy (LITT) schemes, multiple and single rotating laser irradiation, for deep-lying breast tumor to maximize tumor killing while minimizing normal tissue damage and preventing the surface overheating without the aid of external media such as nanoparticles and photosensitizers. Combined with numerical simulation and multi-objective genetic algorithm (MOGA), the incident angle, radius, intensity and exposure time of the multiple laser irradiation, as well as the rotational period and exposure time of the single rotating laser irradiation are optimized based on the surface response model, and a series of Pareto-optimal solutions are obtained. The three-way decisions method is utilized to effectively manage the Pareto-optimal solutions and select the ideal solution with a comprehensive consideration of the risk preferences (aversion, proneness and neutralness) of the surgeon. Compared with the traditional LITT, the ideal solution selected by the operator with risk neutralness attitude makes the surface temperature maintain at an acceptable level, and the killing volume of normal tissue is reduced by 80.4% for the multiple laser irradiation and 99.74% for the single rotating laser irradiation in the case of complete tumor killing. Results obtained can be employed as a guideline for the selection of LITT regimens for deep-lying breast tumor under consideration of the surgeon’s risk preferences.