Metal artifact reduction for improving quantitative SPECT/CT imaging

• Published in: Annals of nuclear medicine 2021-03, Vol.35 (3), p.291-298
• Main Authors: Konishi, Takahiro, Shibutani, Takayuki, Okuda, Koichi, Yoneyama, Hiroto, Moribe, Riku, Onoguchi, Masahisa, Nakajima, Kenichi, Kinuya, Seigo
• Format: Article
• Language: English
• Subjects: Attenuation; Computed tomography; Humans; Image Processing, Computer-Assisted - methods; Image reconstruction; Imaging; Iterative methods; Localization; Medical imaging; Medicine; Medicine & Public Health; Metals; Nuclear Medicine; Original Article; Phantoms, Imaging; Photon emission; Prostheses; Prosthetics; Radioactivity; Radiology; Reduction (metal working); Single photon emission computed tomography; Single Photon Emission Computed Tomography Computed Tomography - methods; Surgical implants; Tomography
• ISSN: 0914-7187; 1864-6433; 1864-6433
• DOI: 10.1007/s12149-020-01560-w

Objective This study aimed to evaluate the effect of the metal artifact reduction (MAR) method on quantitative single-photon emission computed tomography (SPECT)/computed tomography (CT) to reveal the usefulness of MAR in patients with metal implants. Methods We performed a phantom experiment simulating patients with artificial hip prostheses using SPECT/CT equipped with the iterative MAR (iMAR). The phantom was filled with Tc-99m solution (29.5 kBq/mL). For the CT scan conditions, tube current time products were applied to obtain volume CT dose indexes (CTDIvols) of 1.4, 2.8, and 5.6 mGy. Six types of quantitative SPECT images were reconstructed using data from different doses of CT processed with and without iMAR for CT attenuation correction. Thirty circular regions of interest (ROIs) were placed in each of the dark-band artifact areas, the white-streak artifact areas, and the non-artifact areas. We calculated radioactivity concentrations from quantitative SPECT images with and without iMAR to evaluate the quantitative accuracy. The differences of the effect of iMAR with different CT doses were also evaluated. Results The results obtained using CT data with a CTDIvol of 2.8 mGy are described below. For quantitative SPECT data without iMAR, we observed the underestimation of radioactivity concentration in the dark-band artifact areas and overestimation in the white-streak artifact areas. We observed quantification errors ranging from − 41.1% to + 20.0% without iMAR, depending on the ROI localization. When iMAR was used, these errors were reduced to a range of − 22.8% to + 14.2%. The mean absolute error from the true value in the artifact regions was also significantly reduced from 4.00 to 1.74 kBq/mL. In the non-artifact areas, the radioactivity concentrations obtained from the quantitative SPECT data with and without iMAR were equivalent to the true value and did not differ significantly between the two conditions. Similar results were observed for procedures with CTDIvols of 1.4 and 5.6 mGy. Conclusions This study indicated that iMAR could improve the quantitative accuracy of SPECT/CT independent of the CT dose. iMAR can serve as a practicable technique for quantitative SPECT/CT in patients with metal implants.