Metal artifact reduction and tumor detection using photon-counting multi-energy computed tomography.

• Main Authors: Chang-Lae Lee, Junyoung Park, Sangnam Nam, Jiyoung Choi, Yuna Choi, Sangmin Lee, Kyoung-Yong Lee, Minkook Cho
• Format: Article
• Language: English
• Published: Public Library of Science (PLoS) 2021-01-01
• Series: PLoS ONE
• Online Access: https://doi.org/10.1371/journal.pone.0247355

Metal artifacts are considered a major challenge in computed tomography (CT) as these adversely affect the diagnosis and treatment of patients. Several approaches have been developed to address this problem. The present study explored the clinical potential of a novel photon-counting detector (PCD) CT system in reducing metal artifacts in head CT scans. In particular, we studied the recovery of an oral tumor region located under metal artifacts after correction. Three energy thresholds were used to group data into three bins (bin 1: low-energy, bin 2: middle-energy, and bin 3: high-energy) in the prototype PCD CT system. Three types of physical phantoms were scanned on the prototype PCD CT system. First, we assessed the accuracy of iodine quantification using iodine phantoms at varying concentrations. Second, we evaluated the performance of material decomposition (MD) and virtual monochromatic images (VMIs) using a multi-energy CT phantom. Third, we designed an ATOM phantom with metal insertions to verify the effect of the proposed metal artifact reduction. In particular, we placed an insertion-mimicking an iodine-enhanced oral tumor in the beam path of metallic objects. Normalized metal artifact reduction (NMAR) was performed for each energy bin image, followed by an image-based MD and VMI reconstruction. Image quality was analyzed quantitatively by contrast-to-noise ratio (CNR) measurements. The results of iodine quantification showed a good match between the true and measured iodine concentrations. Furthermore, as expected, the contrast between iodine and the surrounding material was higher in bin 1 image than in bin 3 image. On the other hand, the bin 3 image of the ATOM phantom showed fewer metal artifacts than the bin 1 image because of the higher photon energy. The result of quantitative assessment demonstrated that the 40-keV VMI (CNR: 20.6 ± 1.2) with NMAR and MD remarkably increased the contrast of the iodine-enhanced region compared with that of the conventional images (CNR: 10.4 ± 0.5) having 30 to 140 keV energy levels. The PCD-based multi-energy CT imaging has immense potential to maximize the contrast of the target tissue and reduce metal artifacts simultaneously. We believe that it would open the door to novel applications for the diagnosis and treatment of several diseases.