Abstract
OBJECTIVE: Microwave hyperthermia is a clinically validated adjunctive therapy in oncology, employing antenna applicators to selectively raise tumor tissue temperature to 40-44 °C. For deep-seated tumors, especially those in anatomically complex areas like the head and neck (H&N) region, phased array antennas are typically employed. Determining optimal antenna feeding coefficients is crucial to maximize the specific absorption rate (SAR) within the tumor and minimize hotspots in healthy tissues. Conventionally, this optimization relies on meta-heuristic global algorithms such as particle swarm optimization (PSO). METHODS: In this study, we consider a deterministic alternative to PSO in microwave hyperthermia SAR-based optimization, which is based on the Alternating Projections Algorithm (APA). This method iteratively projects the electric field distribution onto a set of constraints to shape the power deposition within a predefined mask, enforcing SAR focusing within the tumor while actively suppressing deposition in healthy tissues. To address the challenge of selecting appropriate power levels, we introduce an adaptive power threshold search mechanism using a properly defined quality parameter, which quantifies the excess of deposited power in healthy tissues. RESULTS: The proposed method is validated on both a simplified numerical testbed and a realistic anatomical phantom. Results demonstrate that the proposed method achieves heating quality comparable to PSO in terms of tumor targeting, while significantly improving hotspot suppression. CONCLUSIONS: The proposed APA framework offers a fast and effective deterministic alternative to meta-heuristic methods, enabling SAR-based optimization in microwave hyperthermia with improved tumor targeting and enhanced suppression of hotspots in healthy tissue.