TU‐B‐204B‐01: A Novel Off‐Axis Scanning Method for Improved CBCT FOV and Reconstruction Using Compressed Sensing

Abstract

Purpose: Current on‐board imaging systems commonly used by modern linear accelerators have a limited field of view (FOV) for a cone‐beam CT (CBCT) scan, which is typically less than 50 cm. Consequently, truncation artifacts often occur for large patients. The goal of this work is to investigate a novel method to increase the FOV for current on‐board CBCT systems. Method: When a large patient is scanned with CBCT, any region outside the FOV is only partially sampled within a short range of projection angles, and at any other angles no x‐ray beams may pass through that region. To increase the sampling rate for the region outside the FOV, we have designed a new source trajectory by shifting the center of rotation during a CBCT scan. This resulted in a reduced sampling rate at the central area and increased sampling rate at the edges. The tradeoff led to a more balanced sampling for an increased FOV. An iterative algorithm was also developed in order to reconstruct the CT image under the new sampling scheme using a compressed sensing technique. The method was validated by computer simulations for a Varian Trilogy CBCT system. A phantom created from CT images of a lung patient was used in the simulations. The isocenter movement followed a near‐ellipse trajectory while the x‐ray source/detector rotated around the object. Results: Under the clinical half‐fan settings with FOV of 45 cm, the truncation artifacts were tremendous for the simulated CBCT scanS. With the new scanning method, artifact‐free imageS can be obtained with the FOV as large as 75 cm. Conclusions: A novel CT scanning geometry was designed and tested, which can dramatically increase the FOV without significant compromise of the CBCT image quality.