Assessment of an Iterative Framework for Brain PET Simulation in Epilepsy

Abstract

Monte Carlo–based PET simulation frameworks are widely used in methodological research, and their underlying modeling and optimization strategies must be validated to ensure clinical relevance. This work evaluates an iterative PET simulation framework for brain imaging in epilepsy using patient-specific simulations derived from clinical PET and MRI data of 13 subjects. Anatomically derived attenuation maps and an iterative activity map optimization strategy anchored to real PET images were employed. Clinical fidelity was evaluated using uptake ratios relevant to epilepsy assesment, together with other similarity and image quality metrics. Simulated PET images reproduced clinically relevant uptake ratios with a global mean relative error of approximately 4% and showed stable convergence. Noise and contrast characteristics were preserved, supporting the proposed framework as a reliable methodology for quantitative epilepsy PET studies