A Proof-of-Concept Study of Multitask Learning for Cranial Synthetic CT Generation Across Heterogeneous MRI Field Strengths
For clinicians and researchers needing reliable CT synthesis from MRI across different scanners, this work addresses a practical bottleneck in clinical translation, though it is a proof-of-concept study.
The study tackles the problem of generalizing cranial CT synthesis from MRI across heterogeneous field strengths and protocols. The proposed deep learning framework achieves improved performance and generalization compared to conventional approaches on multi-site datasets.
Accurate synthesis of computed tomography (CT) images from magnetic resonance imaging (MRI) is clinically valuable for cranial applications such as attenuation correction, radiotherapy planning, and image-guided interventions. However, heterogeneity across MRI field strengths and acquisition protocols limits the generalizability of existing methods. In this study, we formulate cranial CT synthesis as a modular, structurally coupled problem and propose a deep learning framework to improve robustness across heterogeneous MRI conditions. The model is designed to adapt to variations in field strength and imaging protocols while preserving anatomical consistency. Experiments on multi-site datasets demonstrate improved performance and generalization compared with conventional approaches. The proposed method enables reliable CT synthesis across heterogeneous MRI settings, supporting broader clinical translation.