Measurement of Ultrasound Attenuation for Quantifying Amount ff Fat in Chronic Liver Disease

Abstract

Several studies about the evaluation of steatosis based on ultrasound attenuation measured by using ultrasound scanners have been reported (Fujii Y. et al, 2002, Kanayama Y. et al, 2013). However, some more investigations seem to be needed toward the practical use. CAP (controlled attenuation parameters), installed in Fibroscan (Echosens) is an example of practical use, but it is not a genre of imaging modality, also requires a dedicated probe. In this study, we tried to measure ultrasound attenuation from B-mode-based RF signals taken by a conventional ultrasound scanner. RF signals for 125 consecutive chronic liver disease (CLD) patients who underwent liver biopsy on the same day were acquired at transmission and reception frequency of 4.0 MHz. Ultrasound scanner used was LOGIQ E9 (GE Healthcare) with a 3.5 MHz convex probe (C1-6-D). Steatosis and liver fibrosis were staged and graded during the pathological analysis of bioptic specimens. Steatosis was categorized as S0: < 5%; S1: 5-33%; S2: 34-66%; or S3: ≥ 67%. Attenuation coefficient (AC) [dB/cm/MHz] of a region of interest (ROI) was measured from the RF signals of the corresponding B-mode image. As the calculation method, first we compensated the envelope of the signal with the reference signals taken from a reference agar-graphite phantom with known attenuation (SONO403, Gammex Inc., 0.5±0.05 dB/cm/MHz). This procedure can eliminate complicity of the beam shape due to the focal ultrasound. Secondly, we measured AC by calculating a slope of the signal envelope compensated. A position of the ROI was selected manually by an operator to avoid the diaphragm or large vessels. The slope was calculated by least squire method. The mean and standard deviation of attenuation coefficient values in steatosis stages S0 (n = 52), S1 (n = 45), S2 (n = 20), and S3 (n = 9) were 0.47±0.08, 0.56±0.08, 0.67±0.06, and 0.73±0.08 dB/cm/MHz, respectively. The mean of AC values demonstrated a stepwise increase with increasing severity of steatosis (P < 0.0001). A significant correlation between AC and the percentage steatosis was found (r = 0.81, P < 0.0001). On the other hand, the mean and standard deviation of AC values in liver fibrosis grades F0 (n = 16), F1 (n = 14), F2 (n = 13), F3 (n = 3), and F4 (n = 21) were 0.49±0.07, 0.47±0.10, 0.46±0.11, 0.50±0.02, and 0.52±0.06 dB/cm/MHz. A significant correlation between AC and the percentage liver fibrosis wasn’t found (r = -0.03, P > 0.5). Attenuation coefficient measured from B-mode-based RF signals demonstrated a stepwise increase with increasing severity of steatosis and had a strong correlation with the percentage steatosis. On the other hand, AC much less correlated with the percentage liver fibrosis. B-mode images are thought to be a good assistance for the measurement. We conclude that our method is useful for a quantifying amount of fat in CLD.