INTRODUCTION

Liver function assessment is of great significance for the diagnosis of preoperative liver reserve. The indocyanine green test has been used to evaluate liver function in clinical practice. However, it cannot accurately predict liver reserve in the presence of partial liver decompensation¹. Therefore, it is necessary to establish a morphology-based liver function assessment method. Gadoxetic acid-enhanced magnetic resonance imaging (MRI) has been proposed as a method to evaluate liver function^2–4. Gadoxetic acid is a hepatocyte-specific contrast agent used for liver MRI. The T₁ value of the liver in the hepatobiliary phase shortens with higher concentration of gadoxetic acid. Because the uptake of gadoxetic acid by hepatocytes is related to its function, previous studies have reported that the rate of reduction in T₁ values of the liver (T₁ reduction rate) after the administration of gadoxetic acid is a valuable biomarker of liver function^3,4. However, in severe liver dysfunction, it is challenging to accurately assess liver function using the T₁ reduction rate. Motosugi et al. reported that the signal intensity of the liver in the hepatobiliary phase in patients with severe liver function abnormalities increased with larger doses of gadoxetic acid, and that the concentration of the contrast agent in the hepatic focal area was dependent on the administered dose of gadoxetic acid⁵. Therefore, the uptake of gadoxetic acid by hepatocytes may depend not only on liver function, but also on its dose. We hypothesized that bias correction by administration dose of gadoxetic acid can improve the accuracy of stratification of liver function. The purpose of our study was to clarify whether the dose-normalized gadoxetic acid concentration could be an index of local liver function.

METHODS

Forty-two patients who underwent gadoxetic acid-enhanced MRI using a 3.0 T MRI scanner (Ingenia, Philips Healthcare, Best, The Netherlands) were enrolled in this study (mean age, 67.3 years; age range, 26–85 years). They were allocated to groups according to their Child-Pugh classification scores. There were 13 patients with normal liver function (NLF), 18 categorized with Child-Pugh A (CPA), 8 with Child-Pugh B (CPB), and 3 with Child-Pugh C (CPC). A 3D phase-sensitive inversion recovery (PSIR) sequence was performed before and 15 min after the injection of gadoxetic acid. The T₁ values of the liver were calculated on a pixel-by-pixel basis using the signal intensities observed in the PSIR and reference images obtained from the imaging, as previously reported. The following formula was used to calculate the T₁ reduction rate, gadoxetic acid concentration in the hepatic focal area at hepatobiliary phase (C_Hep), and normalized C_Hep:
T₁ reduction rate (in %) = (T_1pre-T_1post) × 100/T_1pre
C_Hep = (1/r₁) × (1/T_1post-1/T_1pre)
where r₁ represents the relaxivity of gadoxetic acid in hepatocytes at 3.0 T
Normalized C_Hep = C_Hep/Dose
where ”Dose” represents the dose of gadoxetic acid
The Steel-Dwass test was used to compare the differences in the T₁ reduction rate, C_Hep, and normalized C_Hep between the NLF, CPA, and CPB+CPC groups. The Spearman rank correlation analysis was performed to assess the correlation coefficients between each index and the biochemical data (the total bilirubin, albumin, and Child-Pugh score) and between C_Hep and dose. Receiver operating characteristic (ROC) analysis was used to evaluate the discrimination of each index between the NLF+CPA and CPB+CPC groups.

RESULTS

Figure 1 shows the representative quantitative images of T₁ reduction rate, C_Hep, and normalized C_Hep for NLF, CPA, CPB+CPC, respectively. The T₁ reduction rate and C_Hep showed statistically significant differences between the NLF and CPA groups and between the NLF and CPB+CPC groups (P < 0.05), but not between the CPA and CPB+CPC groups (P > 0.05). Normalized C_Hep showed statistically significant differences between all groups (P < 0.05)(Fig. 2). Normalized C_Hep was moderately correlated with the biochemical data (Table 1). The area under the ROC curves of T₁ reduction rate, C_Hep, and normalized C_Hep were 0.81, 0.81, and 0.87, respectively (Table 2, Fig. 3). The C_Hep showed no correlation with the dose of gadoxetic acid in the NLF and CPA groups and moderate correlation in the CPB+CPC group.

DISCUSSION

We showed that normalized C_Hep could be accurately assessed for liver function with a severe stage. In the CPB+CPC group, there was a moderate correlation between the dose and gadoxetic acid concentration. The concentration of gadoxetic acid in the liver may be overestimated because of the additional residual gadoxetic acid in the extracellular space due to severe liver dysfunction. Therefore, bias correction based on the administration dose of gadoxetic acid may be effective, especially in stratifying patients with severe liver dysfunction.

CONCLUSION

The use of normalized C_Hep allows for the accurate assessment of severe liver dysfunction.