Synthetic CT artifacts in MR-only brain radiotherapy: Clinical implementation experience

Area of Science:

• Radiotherapy Physics
• Medical Imaging
• Clinical Oncology

Background:

• Magnetic resonance imaging (MRI) offers superior soft tissue contrast for radiotherapy target delineation.
• Synthetic CT (sCT) generation from MRI enables MR-only workflows, reducing reliance on traditional CT simulations.
• Prior research focused on dosimetric equivalence, with less attention to clinical implementation challenges.

Purpose of the Study:

• To report clinical experiences and practical challenges encountered during the implementation of an MR-only brain radiotherapy workflow.
• To identify and categorize issues arising from synthetic CT generation in a clinical setting.

Main Methods:

• Twenty-eight patients underwent MR simulation for sCT generation alongside standard CT simulation.
• A dedicated MR protocol with specific sequences (T1 VIBE Dixon, T2 SPACE, PETRA, TOF-MRA) was used with an immobilization mask and head coil.
• Treatment plans created with planning CT (pCT) were recalculated on the generated sCT datasets.

Main Results:

• Twenty-five out of 28 synthetic CT datasets were successfully generated.
• Nine sCT datasets required re-registration due to inter-sequence patient motion.
• Challenges included workflow issues (motion, coil placement, marker interference) and algorithm artifacts (HU misclassification, degradation in post-operative cases).

Conclusions:

• Clinical implementation highlighted two main challenge categories: workflow-related sCT issues and algorithm-related imaging artifacts.
• Effective mitigation requires case-specific sCT review, eligibility screening, and prospective site commissioning.
• Practical checklists for commissioning and quality assurance are provided to support safe deployment.