Biological and Biomedical Engineering
Hossein’s passion in physics and biology and curiosity in machine learning has brought him to medical physics. He is trying to show that pre-planning in high dose rate brachytherapy is possible, and ultimately reduce the time that the patients are under anesthesia. Apart from physics and coding, he enjoys rugby, kickboxing and listening to podcasts.
Penalty Weight Optimization in High Dose Rate Brachytherapy
Treatment plan optimization problem in high dose rate brachytherapy is formulated as a constrained optimization problem. First the dose constraints and penalty weights are determined by the clinicians, then the optimization problem is solved by linear programing. The dose constraints are usually fixed for each patient depending on the treated tumor site and the treatment planning guidelines followed. However, the clinicians select different penalty weights, leading to different optimization problems and finally adopt the one that results in the most desirable dose distribution. To remove the clinicians influence on plan quality, reinforcement learning is explored.
Catheter Position Optimization in High Dose Rate Brachytherapy
In interstitial high dose rate brachytherapy, a highly radioactive source, usually 192Ir, is temporarily placed inside or in close proximity of the tumor via thin hollow implanted catheters which are connected a machine called an afterloader. The afterloader contains a single radioactive source at the end of a wire. The source is pushed into each of the catheters, one by one under computer control and guided to the tumor site. The computer controls where along the catheter the source should pause to deliver its radiation (dwell positions) and how long it dwells at each position (dwell time). After the desired dose is delivered, the source is pulled back to the afterloader and the catheters are removed. Since the dwell times are optimized, the position of catheters has a major impact on the treatment plan quality. Efforts in optimizing the catheter positions have not been explored as extensively as the other aspects of the treatment planning workflow. This gap in knowledge motivates us to further explore this problem.
Bayesian Optimization in Treatment Planning of High Dose Rate Brachytherapy Inproceedings
In: MEDICAL PHYSICS, pp. E200–E200, WILEY 111 RIVER ST, HOBOKEN 07030-5774, NJ USA 2022.
Mitacs Globalink Research Award award