Cone-Beam CT Image Quality Enhancement Using A Latent Diffusion Model Trained with Simulated CBCT Artifacts

TL;DR

This paper introduces a novel CBCT image enhancement method using a conditional latent diffusion model trained with simulated artifacts, improving image quality while preserving anatomical accuracy.

Contribution

It presents an overcorrection-free, self-supervised learning approach with a latent diffusion model that enhances CBCT images efficiently and accurately, even with limited training data.

Findings

Structural changes in images were less than 0.1% of pixels compared to conventional methods.

Correlation coefficient between generated and reference CT values was 0.916.

The proposed method achieved faster processing and better performance than existing diffusion frameworks.

Abstract

Cone-beam computed tomography (CBCT) images are problematic in clinical medicine because of their low contrast and high artifact content compared with conventional CT images. Although there are some studies to improve image quality, in regions subject to organ deformation, the anatomical structure may change after such image quality improvement. In this study, we propose an overcorrection-free CBCT image quality enhancement method based on a conditional latent diffusion model using pseudo-CBCT images. Pseudo-CBCT images are created from CT images using a simple method that simulates CBCT artifacts and are spatially consistent with the CT images. By performing self-supervised learning with these spatially consistent paired images, we can improve image quality while maintaining anatomical structures. Furthermore, extending the framework of the conditional diffusion model to latent space improves the efficiency of image processing. Our model was trained on pelvic CT-pseudo-CBCT paired data and was applied to both pseudo-CBCT and real CBCT data. The experimental results using data of 75 cases show that with our proposed method, the structural changes were less than 1/1000th (in terms of the number of pixels) of those of a conventional method involving learning with real images, and the correlation coefficient between the CT value distributions of the generated and reference images was 0.916, approaching the same level as conventional methods. We also confirmed that the proposed framework achieves faster processing and superior improvement performance compared with the framework of a conditional diffusion model, even under constrained training settings.