Background Halo artifacts from a urinary catheter due to inaccurate scatter correction can cause significantly reduced tumor visibility in 18F-FDG PET/CT images. We investigated the incidence and mechanism of halo-artifact generation, and we attempted to establish appropriate scatter correction techniques to reduce artifacts.
Methods We first retrospectively reviewed patients with urinary catheters who were undergoing 18F-FDG PET/CT scans. Halo artifacts on the PET images were visually assessed with a standard scatter correction based on a tail-fitted single-scatter simulation (TF-SSS) using 4-mm voxel μ-maps (TFS 4-mm). We then performed phantom studies to reproduce the appearance of halo artifacts and evaluate scatter correction techniques regarding halo-artifact suppression. We measured the standardized uptake values (SUVs) to quantitatively assess the PET images. The scatter correction also used TF-SSS with 2-mm voxel μ-maps (TFS 2-mm) and a Monte Carlo-based single-scatter simulation (MC-SSS). Finally, we investigated whether TFS 2-mm and MC-SSS can be applied to patient data.
Results There were 61 patients with urinary catheters; in five, halo artifacts were observed in the TFS 4-mm images. The phantom study clearly demonstrated the appearance of halo artifacts. The SUVs for the TFS 4-mm and TFS-2mm images were underestimated at the halo-artifact regions, whereas the SUV for the MC-SSS images became the true value (SUV = 1). In four of the five patients, halo artifacts were not present in the TFS 2-mm images. In all five patients, halo artifacts were absent in the MC-SSS images. The patient and phantom studies demonstrated that halo artifacts were caused by overestimates of the scatters due to mismatch by urine movement in the interval between the CT and the PET scan, or by the partial volume effect and downsizing in the converting process from the CT images to the μ-maps.
Conclusions Halo artifacts were due to the mismatch between the μ-maps and PET images, which induces the scatter correction error. With the TF-SSS, halo artifacts were improved by using a small voxel size for the μ-maps, but the artifacts remained in one case. With the MC-SSS, it was possible to accurately estimate the scatters without generating halo artifacts.