Author: Bulent Aydogan, Xiaochuan Pan, Erik Pearson, Zheng Zhang 👨🔬
Affiliation: The University of Chicago 🌍
Purpose: Motion management remains one of the significant challenges in external beam radiation therapy (RT). We are developing time-resolved cone-beam computed tomography (trCBCT) for imaging during the delivery of volumetric modulated arc therapy (VMAT). This approach will generate multiple high resolution CBCT scans per rotation from limited arc data reconstructions without presuming periodic motion using HyperSight (Varian Medical Systems) imaging hardware.
Methods: During VMAT treatment, cone-beam projection data are acquired with the larger HyperSight detector panel which avoids data truncation. Using quantitative consistency metrics, projections are sorted into discrete time points. The use of geometric consistency-based sorting allows time points to bracket abrupt motion to capture full range of motion and individual time points to be minimally impacted by motion artifact. Reconstruction is performed with the Directional Total Variation (DTV) algorithm. Projection data are simulated using models based on patient CT scans with dynamic tissue properties.
Results: The DTV algorithm with additional surface boundary constraints provides accurate and high-resolution reconstructions from the limited sampling angle of trCBCT.
Conclusion: Time-resolved CBCT has the potential to provide an effective solution for managing organ motion and verifying treatments through direct measurement of patient motion. It enables more precise adaptive treatments while leveraging existing hardware—eliminating the need for costly new equipment or complex imaging development, ensuring broad, global impact.