Quantum Annealing for Task-Driven Optimization in SPECT

We show how diverse SPECT design choices—spanning acquisition protocols and hardware configuration—can be reframed as Binary Quadratic Models compatible with D-Wave's quantum and hybrid solvers. The approach is physics-aware: task utilities (e.g., detectability/CNR proxies) are combined with quadratic constraints for cardinality, angular spacing, spatial uniformity, and manufacturability, yielding portable QUBO/Ising objectives without bespoke algorithms per task. Using D-Wave Hybrid, we observe reliable solutions for medium-scale, frustrated landscapes where classical heuristics require extensive tuning, while very dense couplings are handled by hybrid decomposition rather than raw QPU embedding. A reusable pipeline maps measured or simulated figures-of-merit to BQM coefficients, solves on Hybrid, and validates selections with analytic projectors and Monte-Carlo tools. Preliminary studies indicate improved coverage uniformity and task utility under realistic constraints, suggesting quantum-assisted optimization as a practical co-design lever for SPECT acquisition and instrumentation when search spaces are large and interactions are nontrivial.