The purpose of this study is to develop a novel and systematic method to identify deformable image registration (DIR) errors and related dosimetric consequences in image-guided radiotherapy. The accuracy of the DIR and dose reconstruction process is central to determining whether or not the dose delivered to the patient is in agreement with the planned dose distribution. As we begin to reduce planning margins, based on the assurance afforded by daily imaging, and the introduction of real-time targeting devices (e.g. electromagnetic beacons), the impact of tumor (and surrounding organ) deformation may become a limiting factor in accurate targeting of the tumor. Under such circumstances, and regardless of whether on-line or off- line, adaptive corrections are applied, the accuracy of the image registration and dose reconstruction process becomes critical in evaluating the actual dose delivered to the tumor and surrounding healthy tissues. The long-term objective of this application is to ensure that each patient's treatment plan is properly adapted to account for DIR displacement and dose-related errors, to improve targeting accuracy and provide optimal sparing of healthy tissues. To accomplish the goals of this proposal, we will: (1a) Develop a novel elasticity-based model, founded on the concepts of unbalanced forces and energies, to quantify displacement vector field (DVF) errors in deformable image registration, and verify the results against measurements in a deformable phantom;(1b) Apply the method to a large number image datasets of prostate and lung cancer patients previously treated using daily CBCT imaging;(2) Perform dose reconstruction using tri-linear dose interpolation and Monte Carlo-based energy mapping and quantify the resulting dosimetric errors on the patient image datasets;(3) Develop methods to compensate for dose errors from DVF-based displacement errors;(3a) Develop a dose reconstruction system using an optimization process to minimize errors in the dose by incorporating feedback based on the quantified DVF-based dose errors;(3b) Quantify the dosimetric errors as a function of planning margin and develop a margin recipe to account for DVF-related dose errors based on registrations of daily cone-beam CT (CBCT) images with simulation CT images for a large group of prostate and lung cancer patients treated retrospectively.
We will investigate methods for quantifying deformable image registration (DIR) errors and related dosimetric consequences. DIR and dose reconstruction are principle processes in adaptive radiotherapy and are essential requirements for computation of the actual dose delivered to the patient. A feedback system will be developed to minimize the quantified errors and we will formulate margin recipes to account for them in treatment planning.
|Li, Haisen S; Zhong, Hualiang; Kim, Jinkoo et al. (2014) Direct dose mapping versus energy/mass transfer mapping for 4D dose accumulation: fundamental differences and dosimetric consequences. Phys Med Biol 59:173-88|
|Zhong, Hualiang; Chetty, Indrin (2014) A note on modeling of tumor regression for estimation of radiobiological parameters. Med Phys 41:081702|
|Li, Shunshan; Glide-Hurst, Carri; Lu, Mei et al. (2013) Voxel-based statistical analysis of uncertainties associated with deformable image registration. Phys Med Biol 58:6481-94|
|Kim, Jinkoo; Kumar, Sanath; Liu, Chang et al. (2013) A novel approach for establishing benchmark CBCT/CT deformable image registrations in prostate cancer radiotherapy. Phys Med Biol 58:8077-97|
|Zhong, Hualiang; Jin, Jian-Yue; Ajlouni, Munther et al. (2011) Measurement of regional compliance using 4DCT images for assessment of radiation treatment. Med Phys 38:1567-78|