Preview

PDF, English - main document Download (6MB) | Terms of use

Abstract

Megavoltage imaging offers unique clinical applications due to providing a beam’s-eye-view of the actual radiation delivery. However, poor electronic portal imaging device (EPID) performance presently limits the clinical utility of megavoltage imaging. This thesis describes the clinical translation, implementation, and trial of a novel multilayer imager (MLI) designed to address current EPID shortcomings, as well as the development of an application using the imager to track tumor location during treatment.

The prototype MLI was constructed, with standard imaging metrics demonstrating a 5.7x increase in detective quantum efficiency, as well as substantially improved contrast- and signal-to-noise ratios compared to standard EPID. Pre-clinical tests were performed on an anthropomorphic phantom to verify improved performance despite anatomical variations. Subsequently, we conducted a clinical trial of six patients receiving radiation for liver metastases. A beam’s-eye-view tumor tracking algorithm was utilized to assess MLI performance compared to a standard single layer imager. Tumor tracking using MLI was found to be significantly more accurate and efficient at successfully tracking on more frames. Further analysis revealed correlation between noise reduction and improved tracking performance. Given the MLI’s superior performance, for clinical beam’s-eye applications we recommend noise reduction strategies such as employing multiple detection layers in the EPID.