End-to-End Deep Learning CT Image Reconstruction for Metal Artifact Reduction

Abstract

Metal artifacts are common in CT-guided interventions due to the presence of metallic instruments. These artifacts often obscure clinically relevant structures, which can complicate the intervention. In this work, we present a deep learning CT reconstruction called iCTU-Net for the reduction of metal artifacts. The network emulates the filtering and back projection steps of the classical filtered back projection (FBP). A U-Net is used as post-processing to refine the back projected image. The reconstruction is trained end-to-end, i.e., the inputs of the iCTU-Net are sinograms and the outputs are reconstructed images. The network does not require a predefined back projection operator or the exact X-ray beam geometry. Supervised training is performed on simulated interventional data of the abdomen. For projection data exhibiting severe artifacts, the iCTU-Net achieved reconstructions with SSIM = 0.970±0.009 and PSNR = 40.7±1.6. The best reference method, an image based post-processing network, only achieved SSIM = 0.944±0.024 and PSNR = 39.8±1.9. Since the whole reconstruction process is learned, the network was able to fully utilize the raw data, which benefited from the removal of metal artifacts. The proposed method was the only studied method that could eliminate the metal streak artifacts.

A nonlinear scaling‐based normalized metal artifact reduction to reduce low‐frequency artifacts in energy‐integrating and photon‐counting CT

“…LIMAR is equivalent to NMAR, just without the tissue-normalization step. 18 Primarily in recent deep learning-based MAR methods, LIMAR and NMAR are the only reference algorithms, which are used in almost every publication for both conventional CT [19][20][21][22][23][24][25][26] as well as cone-beam CT. [27][28][29] Sinogram inpainting describes the replacement of the metal projections, mostly by interpolating linearly in between the metal boundaries of the uncorrected sinogram to minimize the impact of the metal projections onto the reconstructed image. Because all metal projections are considered invalid information, the resulting LI image contains strong interpolation artifacts, which blur surrounding tissue.…”

“…LIMAR is equivalent to NMAR, just without the tissue‐normalization step 18 . Primarily in recent deep learning‐based MAR methods, LIMAR and NMAR are the only reference algorithms, which are used in almost every publication for both conventional CT 19–26 as well as cone‐beam CT 27–29 …”

Nonlinearly scaled prior image‐controlled frequency split for high‐frequency metal artifact reduction in computed tomography

“…Most of them can be categorized in the following two major groups: Image-based approaches [18][19][20] aim to correct artifacts solely in image space. In contrast, sinogram-based approaches [21][22][23] try to correct the information within the metal trace of the sinogram instead of applying linear interpolation. Hybrid approaches combine image-and sinogram-based methods.…”

A New Iterative Metal Artifact Reduction Algorithm for Both Energy-Integrating and Photon-Counting CT Systems

“…Those could be carried over to various steps within iMAR or iMARv2, such as the metal segmentation, which is still based on a thresholding-based imaging mask 41 . Related recent studies in this area have already shown promising performance when incorporating deep learning into MAR by learning an end-to-end pipeline in either image, 42 sinogram space, 43 or mixed space 44 …”