MRI-Only Simulation for Prostate MRI-Guided SBRT.
Authors
Affiliations (3)
Affiliations (3)
- Department of Radiation Oncology, Brigham and Women's Hospital, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02115, USA; Mannheim Institute for Intelligent Systems in Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim 68167, Germany; Department of Physics and Astronomy, Heidelberg University, 69117 Heidelberg, Germany; Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, 69117 Heidelberg, Germany. Electronic address: [email protected].
- Department of Radiation Oncology, Brigham and Women's Hospital, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02115, USA.
- Mannheim Institute for Intelligent Systems in Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim 68167, Germany; Department of Physics and Astronomy, Heidelberg University, 69117 Heidelberg, Germany; Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, 69117 Heidelberg, Germany; Central Institute for Computer Engineering (ZITI), Heidelberg University, 69117 Heidelberg, Germany; CZS Heidelberg Initiative for Model-Based AI, Heidelberg University, 69117 Heidelberg, Germany.
Abstract
The SMART (Stereotactic MR-Guided Adaptive Radiation Therapy) protocol for prostate SBRT has demonstrated favorable clinical outcomes using a 3T-MRI for delineation and a 0.35 T MR-Linac for adaptation, addressing challenges associated with low-field MRI delineation. However, this workflow necessitates three distinct simulation scans (CT, 3T-MRI, and 0.35T-MRI), resulting in logistical complexity and inefficiency. To address these limitations, we have developed and validated an end-toend MR-only workflow to eliminate the need for a planning CT scan (pCT), thereby streamlining treatment delivery and reducing patient burden. A workflow was developed wherein ten prostate cancer patients were simulated on a 3T Siemens MAGNETOM Vida MRI scanner. Consistent with the SMART protocol, high-resolution T2-weighted BLADE/DWI sequences were used for definitive urethra/dominant intraprostatic lesion (DIPL) delineation, and additionally a T1weighted DIXON VIBE sequence was acquired to generate a synthetic CT (sCT) using an FDA-approved deep-learning-based algorithm. This single 3T-simulation dataset was used to plan and subsequently deliver treatment on the ViewRay Systems MRIdian 0.35T MR-Linac. To validate this CT-free pathway, pCT-based plans were retrospectively recalculated on the sCTs. Dosimetric agreement was assessed using DVH analysis and 3D gamma index analysis. Clinical implementation of the full workflow was successful. Dosimetric validation demonstrated high fidelity between sCT and pCT calculations. Computed γ-indices were 96.80±1.33% (2% dose deviation (DD), 2 mm dose-to-agreement (DTA), 10% threshold, local dose normalization) and 99.98±0.03% (3%DD,2mm DTA, 10%, global). Mean absolute differences in PTV D95% were small (0.06±0.14 Gy). Integrating high-field 3T MR simulation with 0.35 T MR-Linac delivery via an sCT pathway is clinically feasible and dosimetrically robust. MRI-only simulation reduces patient burden and improves workflow efficiency, while maintaining the high-quality 3T delineation and MR-guidance that defines the SMART approach.