Challenges and opportunities in the development and clinical implementation of artificial intelligence based synthetic computed tomography for magnetic resonance only radiotherapy

• Published in: Radiotherapy and oncology Vol. 198; p. 110387
• Main Authors: Villegas, Fernanda, Dal Bello, Riccardo, Alvarez-Andres, Emilie, Dhont, Jennifer, Janssen, Tomas, Milan, Lisa, Robert, Charlotte, Salagean, Ghizela-Ana-Maria, Tejedor, Natalia, Trnková, Petra, Fusella, Marco, Placidi, Lorenzo, Cusumano, Davide
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.09.2024
• Subjects: Artificial Intelligence; Clinical implementation; Deep Learning; Hematology, Oncology, and Palliative Medicine; Humans; Magnetic Resonance Imaging - methods; MR-only planning; MR-only radiotherapy; Neoplasms - diagnostic imaging; Neoplasms - radiotherapy; Radiotherapy Planning, Computer-Assisted - methods; Radiotherapy, Image-Guided - methods; Synthetic CT; Tomography, X-Ray Computed - methods; OAR; IGRT; RT; PACS; DRR; MRI; DL; MR-Linac; CPU; GPU; HU; CBCT; Deep learning; ESTRO; QA; DDR; DSC; SNR; TPS; ME; MR-only planning; DICOM; Clinical implementation; pCT; PSNR; SSIM; IGRT-Linac; MR; sCT; AI; MU; FOV; MR-only radiotherapy; B0; MAE; CT; Synthetic CT; RF; GAN; PSQA; DVH; SBRT; HFMEA; Artificial intelligence; planning CT; Graphics Processing Unit; Digital Imaging and Communications in Medicine; Hounsfield Unit; A linear accelerator combined with kV-kV or CBCT imaging capability; Signal to Noise Ratio; Mean Error; Peak Signal-to-Noise Ratio; European Society for Radiotherapy and Oncology; Magnetic Resonance Imaging; Dose Distribution Reconstruction; Central Processing Unit; Generative Adversarial Network; Quality Assurance; Monitor Unit; Organ-at-Risk; Field Of View; Structural Similarity Index Measure; Mean Absolute Error; Cone-Beam Computed Tomography; Magnetic Resonance; Healthcare Failure Mode and Effect Analysis; Computed Tomography; Image-Guided Radiation Therapy; Magnetic Resonance Linear Accelerator; Main Magnetic Field Strength; Dose Volume Histogram; Stereotactic Body Radiation Therapy; Radiofrequency; Treatment Planning System; Digitally Reconstructed Radiography; Dice Similarity Coefficient; Picture Archiving and Communication System; Radiation Therapy; Patient Specific Quality Assurance
• ISSN: 0167-8140; 1879-0887; 1879-0887
• DOI: 10.1016/j.radonc.2024.110387

•Synthetic CT generation is a key point in the MR-only radiotherapy.•ESTRO Physics Workshop 2022 group discusses the clinical sCT integration solutions.•Need for a consensus on sCT commissioning and quality assurance.•Report on sCT development, commissioning, and key elements for safe implementation. Synthetic computed tomography (sCT) generated from magnetic resonance imaging (MRI) can serve as a substitute for planning CT in radiation therapy (RT), thereby removing registration uncertainties associated with multi-modality imaging pairing, reducing costs and patient radiation exposure. CE/FDA-approved sCT solutions are nowadays available for pelvis, brain, and head and neck, while more complex deep learning (DL) algorithms are under investigation for other anatomic sites. The main challenge in achieving a widespread clinical implementation of sCT lies in the absence of consensus on sCT commissioning and quality assurance (QA), resulting in variation of sCT approaches across different hospitals. To address this issue, a group of experts gathered at the ESTRO Physics Workshop 2022 to discuss the integration of sCT solutions into clinics and report the process and its outcomes. This position paper focuses on aspects of sCT development and commissioning, outlining key elements crucial for the safe implementation of an MRI-only RT workflow.