A New Optimization Model for Branchytheraphy Treatment Planning

A New Optimization Model for Branchytheraphy Treatment Planning & Action

Cyrus Angelo A. Selga

(MS Graduated: 2nd Sem 2009-2010)

ERDT Conference Papers,  February (2009)

Abstract

According to statistics from the Department of Health, prostate cancer is the 2nd leading cancer case among Filipino men. An effective treatment strategy is radiation therapy, wherein radiation is delivered to the prostate while attempting to minimize exposure to the nearby organs. Radiation therapy can be delivered either through an external beam or internally using radionuclide sources, the latter technique being called brachytherapy. An important step in brachytherapy is the treatment planning stage wherein the source positions are optimized in order to achieve a quality treatment plan. Quality, however, involves several dimensions and previous research efforts used the a priori decision-making approach, particularly the weighted-sum method. This approach, unfortunately, does not allow the decision-maker to gain an insight into the tradeoff relationships among the different treatment objectives.

In this paper, an a posteriori multi-objective optimization framework is proposed. The model is solved using the Normal-Boundary Intersection technique, which has been extended to accommodate both the a posteriori approach and its integration with the a priori approach. Software implementation used a free and open-source solver from the Computational Infrastructure for Operations Research. Results from actual patient data provided by the Urology Center of the Philippines show that multiple quality treatment plans which use less resources compared to manual planning can be obtained. Further research in algorithm runtime is required in order to generate the whole set of treatment plans within the time frame of manual planning.