A digital twin framework for adaptive treatment planning in radiotherapy.
Authors
Affiliations (5)
Affiliations (5)
- Department of Radiology Oncology, Emory University, 1365 Clifton RD NE, Atlanta, Georgia, 30322-1007, United States.
- Department of Radiation Oncology and Winship Cancer Institute, Emory University, 1365 Clifton Rd NE Building C, Atlanta, Georgia, 30322, United States.
- Department of Medicine, Medical College of Georgia, Augusta, GA, 30912, Augusta, Georgia, 30912-0004, United States.
- Computer Science and Informatics, Emory University, Clifton Rd, Atlanta, Georgia, 30322-1007, United States.
- Department of Radiology Oncology, Emory University, Clifton Rd, Atlanta, Georgia, 30322-1007, United States.
Abstract
This work aims to develop a digital twin (DT) framework for fast online adaptive proton therapy planning in prostate stereotactic body radiation therapy (SBRT) with dominant intraprostatic lesion (DIL) boost, achieving clinical-equivalent plan quality with significantly reduced reoptimization time compared to traditional clinical workflows.
Approach: The proposed DT framework integrates deep learning-based multi-atlas deformable image registration, daily patient anatomy updates, and knowledge-based plan quality evaluation to enable predictive and adaptive radiotherapy. Leveraging a database of 43 prior prostate SBRT cases, the framework forecasts potential interfractional anatomical variations for a new patient and pre-generates multiple probabilistic treatment plans. Upon acquiring daily cone-beam CT (CBCT) for the new patient, the framework facilitates rapid online plan reoptimization. Plan quality is assessed using the ProKnow scoring system, evaluating dose coverage to the DIL and clinical target volume (CTV), as well as sparing of organs at risk (OARs).
Main Results: The DT framework achieved an average reoptimization time of 5.52 ± 2.94 min, producing optimal DT-based plans with a mean plan quality score of 164.01 ± 8.03. These scores matched or exceeded those of the clinical plans, which required substantially longer reoptimization times (17.66 ± 7.93 min) to achieve comparable plan quality (161.23 ± 9.50). DT-based plans provided DIL V100 coverage of 99.27% ± 0.64% and CTV V100 of 99.95% ± 0.07%, with reduced OAR doses, including bladder V20.8Gy of 10.61 ± 2.64 cc, rectum V23Gy of 0.61±0.32 cc, and urethra D10 of 89.45%±1.07%, which were comparable to the clinical quality standard.
Significance: The proposed DT framework facilitates rapid, clinically comparable adaptive proton therapy planning, reducing reoptimization time while preserving or enhancing clinical plan quality. By addressing interfractional anatomical variations efficiently, it enhances treatment precision, reduces OAR toxicity, and supports online personalized radiotherapy, offering a transformative approach for prostate SBRT with DIL boost.