Energy-Guided Diffusion Model for CBCT-to-CT Synthesis

TL;DR

This paper introduces an energy-guided diffusion model (EGDiff) that synthesizes high-quality CT images from CBCT scans, improving image accuracy and quality for adaptive radiation therapy.

Contribution

The novel energy-guided diffusion model enhances CBCT-to-CT synthesis, outperforming existing methods in accuracy and visual quality on a chest tumor dataset.

Findings

Achieved average absolute error of 26.87 HU

Structural similarity index of 0.850

Peak signal-to-noise ratio of 19.83 dB

Abstract

Cone Beam CT (CBCT) plays a crucial role in Adaptive Radiation Therapy (ART) by accurately providing radiation treatment when organ anatomy changes occur. However, CBCT images suffer from scatter noise and artifacts, making relying solely on CBCT for precise dose calculation and accurate tissue localization challenging. Therefore, there is a need to improve CBCT image quality and Hounsfield Unit (HU) accuracy while preserving anatomical structures. To enhance the role and application value of CBCT in ART, we propose an energy-guided diffusion model (EGDiff) and conduct experiments on a chest tumor dataset to generate synthetic CT (sCT) from CBCT. The experimental results demonstrate impressive performance with an average absolute error of 26.87$\pm$6.14 HU, a structural similarity index measurement of 0.850$\pm$0.03, a peak signal-to-noise ratio of the sCT of 19.83$\pm$1.39 dB, and a normalized cross-correlation of the sCT of 0.874$\pm$0.04. These results indicate that our method outperforms state-of-the-art unsupervised synthesis methods in accuracy and visual quality, producing superior sCT images.