Author: Kwang-Ho Cheong, Seungryong Cho, Joonil Hwang, Jae Won Jung, Hoyeon Lee, Raymond Hyunwoo Moon, Inhwan Yeo, Jihyung Yoon 👨🔬
Affiliation: East Carolina University, INOVA Schar Cancer Institute, University of Hong Kong, Hanllym University, KAIST, University of Rochester 🌍
Purpose: Monitoring a target and neighboring organs during beam delivery is crucial for successful radiotherapy (RT). Conventional transit imaging methods lack volumetric reconstruction capabilities, limiting visualization of surrounding tissues and providing poor-quality images, due to collimation. This study introduces a novel 4D On-Beam CT (OBCT) reconstruction method to acquire high-quality volumetric images to visualize target displacement and surrounding tissues during RT by modifying simulation 4D CT.
Materials and Methods: 4D planning CT images for a simplified lung phantom were acquired with a sinusoidal target motion of 3 cm amplitude, where a 3 cm-diameter spherical target was located at the isocenter (pCTiso). To represent a treatment scenario, the target was shifted to the superior direction by 1cm. Corresponding 4D CT (pCT1cm), and transit projection images were obtained with an Electronic Portal Imaging Device (EPIDMEAS) during arc rotation with a fixed MLC position (360 projections/15 phases/360° gantry rotation). We simulated transit images of the pCTiso at corresponding gantry angles to each respiratory phase utilizing a GPU-accelerated Monte Carlo code (EPIDSIM). The differences between the EPIDSIM and EPIDMEAS were and integrated with pCTiso, creating high-quality 4D OBCT. For validation, we compared the 4D OBCT with pCT1cm across all respiratory phases.
Results: The 4D OBCT images showed strong agreement with pCT1cm. The root mean squared error between OBCT and pCT1cm was 230 HU±18 HU, the peak signal-to-noise ratio was 30.40±1.07 dB, and the structural similarity index was 0.9497±0.0070, validating the accuracy of the proposed method for target and neighboring tissue visualization at beam delivery despite using sparse data.
Conclusion: The proposed 4D OBCT method effectively visualized target displacement and surrounding anatomy during RT. In a forthcoming study, we plan to compare 4D OBCT with kV CBCT and MVCT in terms of image quality and dose calculation.