Quantitative Assessment of Lipiodol-Related Artifact Reduction for Dual-Energy Computed Tomography After
Yun-Lung Ting, Hsiang-Yi Lin1, Tzu-Ting Huang2
1Departments of Radiology.
Objective:
This study used metal artifact reduction (MAR) software to examine the computed tomography (CT) number of dual-energy CT (DECT) of hepatocellular carcinoma after transcatheter arterial chemoembolization.
Methods:
Hollow columnar acrylic phantoms were filled with lipiodol and inserts of 2 sizes (large and small) were used to simulate liver tumors on a Revolution GSI CT scanner. The CT numbers of a single test object were collected twice: once with and once without the MAR algorithm. Lipiodol beam-hardening artifacts were quantified by measuring CT numbers in a region of interest around the tumor-simulating insert.
Results:
The virtual monochromatic CT numbers of large and small tumors were closely related to energy. For small tumors, CT numbers increased with energy. For large tumors, CT numbers increased with energy at 1 cm from the margin but decreased with an increase in energy at 5 cm. Regardless of the size, distance, or location of the tumor, the CT numbers fluctuated more at low energy levels.
Conclusions:
At 1 cm from the margin, the CT numbers with MAR were significantly different from those without MAR. Low-energy CT numbers with MAR were near reference values. Metal artifact reduction exhibited superior performance for small tumors. Tumor margin images are affected by artifacts caused by Lipiodol. However, with MAR, CT numbers can be effectively calibrated, thus enabling clinicians to more accurately evaluate hepatocellular carcinoma development and identify residual tumors and recurrent or metastatic lesions.
