MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Continuous Generation of Volumetric Patients Images During Radiotherapy Treatment Using Triggered KV Images and An External Surrogate

M Lafreniere1*, N Mahadeo2 , J Rottmann3 , C Williams4 , (1) Harvard Medical School - DFCI, Cambridge, MA, (2) ,,,(3) Brigham and Woman's Hospital, Boston, MA, (4) Brigham and Women's Hospital, Boston, MA

Presentations

(Tuesday, 7/31/2018) 7:30 AM - 9:30 AM

Room: Room 205

Purpose: In patients undergoing radiotherapy treatment of the chest and abdomen, respiratory motion changes cause inter-fraction and intra-fraction variations in patient anatomy, resulting in discrepancies between the planned and the delivered dose. Continuous knowledge of the volumetric anatomical information during treatment would allow a precise estimation of the delivered dose and would also enable the adaption of therapy to account for any in-treatment deviations. However, current clinical techniques do not provide sufficient information to fully reconstruct 3D anatomy in real time. We developed a technique that combines information obtained during treatment on a standard clinical accelerator with a patient-specific motion model to generate continuous 3D volumetric imaging that can serve as a basis for delivered dose calculation and adaptive radiotherapy.

Methods: A patient-specific motion model is constructed by performing principal component analysis (PCA) on deformation vector fields produced from deformably registering the phases of the patient's pre-treatment 4DCT scan. During treatment, kV planar images are acquired with a maximum rate of three images per second, and the position of an external respiratory surrogate block is monitored continuously. Volumetric images are reconstructed at each kV acquisition by iteratively fitting the PCA weights of the motion model to match the acquired kV projection image. A correlation between these PCA weights and surrogate kinetics enables a continuous estimation of a patient's 3D anatomy throughout treatment delivery.

Results: Performance of the algorithm was evaluated using a digital XCAT phantom, which was programmed to move in accordance with tumor motions measured from patient treatments. The correlation model allowed generation of volumetric images from the surrogate trace. Tumor positions were reconstructed with an average root mean square error of 0.67 mm between generated images and ground truth.

Conclusion: Using a correlation model between an external surrogate and kV imaging enables continuous generation of volumetric patient anatomy.

Funding Support, Disclosures, and Conflict of Interest: This work was supported by a research grant from Varian Medical Systems.

Keywords

Image Guidance, Image-guided Therapy, Reconstruction

Taxonomy

TH- RT Interfraction motion management : Development (new technology and techniques)

Contact Email