Quantum for radiotherapy
Problem
Radiotherapy requires precise Fluence Map Optimization (FMO) to maximize tumor irradiation and minimize healthy tissue damage. Conventional FMO faces computational challenges with increasing number of variables and voxel size. The multi-objective nature of FMO yields a Pareto front of optimal solutions, necessitating exploration of novel computational tools like quantum computing for potential improvements.
Solution
This study explored quantum annealing for FMO. Quantum annealers use quantum mechanics to navigate complex optimization landscapes, potentially faster than classical methods. FMO was reformulated as a Quadratic Unconstrained Binary Optimization (QUBO) problem for compatibility with quantum hardware. Simulated annealing, which was executed on classical hardware, served as a benchmark. FMO formulations were varied (beam angles, resolution, organs at risk, voxel count) to assess quantum solution efficiency.
Benefit
Quantum annealing feasibly solved a simplified FMO problem (with one organ at risk and three beam angles). More complex FMOs were solvable with simulated annealing. Now, a significant gap exists between the current capabilities of quantum annealers (limited to simplified FMO formulations) and the stringent requirements for clinical FMO application; therefore, while feasibility was shown, scalability and benefits of quantum solutions need more research.
This work is supported by the Dutch National Growth Fund (NGF) as part of the Quantum Delta NL programme.
