Synthetic CT Imaging for Pelvic and Abdominal MR-Only Radiotherapy: Clinical Validation on a 1.5T MR-Linac
Abstract
Purpose
To evaluate two prototype deep learning–based synthetic CT (sCT) generation models for pelvic and abdominal anatomies using T1-weighted and T2-weighted MR sequences. This study aims to validate their potential to enable MR-only radiotherapy (RT) workflows on a 1.5T MR-Linac system, eliminating the need for CT imaging while reducing patient burden and workflow complexity.
Methods
Thirty-five patients with balanced distributions across sex, tumor site, and MRI sequence were included. sCTs were generated using two deep learning models (Image+, MVision AI). Geometric accuracy and Hounsfield Unit (HU) fidelity were quantitatively evaluated through comparison with deformed CTs (dCT) registered to MRI geometry. Dosimetric validation was performed by comparing dose distributions using γ-index pass rates between sCT-based plans and two reference standards: dCT and clinical bulk-density CT (bCT).
Results
Mean Dice similarity coefficients indicated good geometric agreement: 0.73 ± 0.18 (air cavities), 0.86 ± 0.06 (adipose tissue), 0.85 ± 0.04 (soft tissue), and 0.75 ± 0.07 (bone). Whole-body HU analysis yielded a Mean Error of 0.06 ± 5.66 HU and a Mean Absolute Error of 35.79 ± 5.84 HU. Dosimetric evaluation showed averaged dose differences of 0.31% ± 0.98% (bCT reference) and 0.16% ± 1.13% (dCT reference) across anatomical structures. Clinical dose agreement within 2% was achieved in 98.8% of pelvic cases and 93.3% of abdominal cases. Individual dose differences ranged from −2.1% to +4.78% for pelvic regions and from −4.15% to +5.86% for abdominal regions.
Conclusion
MRI-based sCTs generated from T1- or T2-weighted MR-Linac sequences demonstrated strong geometric and dosimetric agreement with dCT, particularly for pelvic anatomy, with performance independent of MR sequence type. These findings support the clinical viability of sCTs as an alternative to conventional CT and establish the feasibility of comprehensive MR-only RT workflows from simulation through treatment delivery while eliminating ionizing radiation exposure from simulation imaging.