Single and dual-energy computed tomography guided radiation therapy
Skip to main content
Open Access Publications from the University of California


UCLA Electronic Theses and Dissertations bannerUCLA

Single and dual-energy computed tomography guided radiation therapy

  • Author(s): Lyu, Qihui
  • Advisor(s): Sheng, Ke
  • et al.

Purpose: Conformal dose and precise imaging are key to radiation therapy. Here we introduce a series of integrated optimization frameworks to improve computed tomography (CT) image quality, refine dual-energy CT (DECT) material decomposition, and advance treatment planning and delivery methods.Methods: We formulate our optimization framework as a least-square fidelity term and a regularization term. The regularization term was designed specifically for each application, including accounting for the piecewise smoothness of the CT image, sparsity in the DECT decomposition image, and mechanical constraints in radiotherapy. The flexible optimization framework allows novel treatments that unleash unnecessary constraints in the current Volumetric Modulated Arc Therapy (VMAT) or Intensity Modulated Radiation Therapy (IMRT), including allowing dual-layer Multi-Layer Collimator (DLMLC), non-isocentric treatment, dynamic collimator, non-coplanar beams, and FLASH radiotherapy. The added degrees of freedom significantly expand the searching space, and these large-scale optimization problems were solved with a Fast Iterative Shrinkage-Thresholding Algorithm (FISTA). Results: On a Catphan study, our integrated CT reconstruction distinguishes a higher number of line pairs and preserves low contrast objects compared with conventional Filtered Back Projection (FBP) and total variation (TV) reconstruction. Our framework improved the decomposition accuracy from 63.9% to 99.8% when applying for material decomposition compared with the classic direct inversion method. Compared with single-layer MLC (SLMLC) VMAT, DLMLC VMAT reduced R50 by 10% for radiotherapy treatment. Compared with isocentric 100cm-source-to-isocenter distance (SID-100cm) 4πIMRT, the non-isocentric SID-50cm 4πIMRT reduced R50 and integral dose by 5.3% and 9.6%. Compared with static collimator VMAT (SCVMAT), dynamic collimator VMAT (DCVMAT) reduced the max and mean organs-at-risk (OAR) dose by 4.49% and 2.53% of the prescription dose. Compared with coplanar VMAT, 4πVMAT reduced R50 by 19.7%. Compared with clinical VMAT, our FLASH delivery method reduced the max and mean OAR physical doses by 4.8Gy and 6.3Gy in addition to potential biological gains. Conclusions: The integrated optimization framework improves CT image quality, DECT decomposition accuracy, and radiotherapy dose conformality.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View