Validating clinical feasibility of MRCAT and deep learning-based synthetic CT images for cervical cancer patient

This study validated the clinical feasibility of MRCAT (MR for Calculating ATtenuation) and deep learning-based synthetic CT (sCT) images for implementing an MR-only workflow in radiation therapy planning for cervical cancer patients. MRCAT and sCT were compared and analyzed against actual CT images in 50 cervical cancer patients, evaluating the accuracy of Hounsfield unit (HU) values, dose distribution, and dose-volume histogram (DVH) parameters. The results confirmed that both MRCAT and sCT showed dose differences of less than 1% in the planning target volume (PTV) and organs at risk (OAR), demonstrating clinically acceptable levels. In particular, sCT showed superior HU accuracy in bone regions. This study performed contouring and dose calculation using Coreline Soft's AVIEW RT-ACS system. These findings support the feasibility of MR-only radiotherapy workflows, demonstrating that synthetic CT can provide reliable dose calculation and anatomical representation comparable to conventional CT-based planning. In this study, AVIEW RT-ACS system was utilized for contouring and dose calculation, supporting consistent and reproducible radiotherapy planning.