Applying High School Concepts In Medicine: Comparison of Metaheuristic Algorithms for Treatment Planning In Brachytherapy
Abstract
Purpose
Many of the mathematical principles used in medical physics such as geometric reasoning, exponential decay, iterative improvement, and basic probability originate from concepts introduced in high school. This study illustrates how these foundational concepts can be applied to a clinically relevant problem by examining population-based optimization strategies for dwell-time planning in brachytherapy.
Methods
Two metaheuristic algorithms, Genetic Algorithms (GA) and Hippopotamus Optimization (HO), were implemented for optimization of cylinder and tandem-and-ovoid brachytherapy. Dwell-time optimization is a simpler radiotherapy optimization problem, with the added advantage of calculating dose by incorporating the concept of exponential decay, also a part of high-school curriculum. GA, based on natural selection, is useful for students to understand the evolutionary process and the convergence to an optimal solution. HO, inspired by cooperative foraging dynamics teaches about the balance between local and global exploration for a solution, while engaging in group behavior. During the search strategy, target-coverage and organ-at-risk constraints were used so that the concepts of competing objectives and trade-offs could be learned. Multiple optimization runs were performed to evaluate convergence behavior, robustness, and computational efficiency.
Results
Both GA and HO produced clinically acceptable plans across both applicator types. For cylinder cases with only target-dose objectives, the algorithms achieved comparable coverage, converging within seconds. In the more complex tandem-and-ovoid geometry, GA converged more rapidly, though both methods achieved acceptable solutions without getting trapped in local minima.
Conclusion
This work demonstrates that advanced treatment-planning optimization can be understood and implemented using ideas accessible at the high-school level. By framing metaheuristic optimization through familiar constructs such as probability, iterative improvement, genetics and teamwork, we show that foundational mathematics and biology knowledge can be leveraged to address sophisticated clinical problems in brachytherapy.