Purpose: To evaluate posthepatectomy liver failure (PHLF) using gadoxetic acid-enhanced magnetic resonance imaging (MRI) with a measure of relative liver enhancement (RLE) on hepatobiliary phase images, thereby facilitating safe liver resection. Methods: Twenty patients in Child-Pugh class A underwent tumor excision surgery and indocyanine green (ICG) clearance of future remnant liver (FRL) (ICG-Krem) values were >0.05. PHLF was evaluated using the grading system of the International Study Group of Liver Surgery (ISGLS). The RLE value was defined as the signal gain percentage between the precontrast and hepatocellular images. In the whole liver and FRL, theRLE value measured the tumor-free liver parenchyma in RLE images. We examined the correlation between indocyanine green clearance (ICG-K) and MRI-based liver function in the whole liver. Preoperative PHLF evaluation was predicted using remnant hepatocellular uptake index (rHUI), remnant RLE (rRLE), coefficient variation of Rrle [Cv(rRLE)], and ICG-Krem corrected by heterogeneous liver function(HLF-ICG-Krem). Results: HLF-ICG-Krem and rRLE values correlated with INRs after postoperative day five (r = -0.55 and 0.46, p = 0.01 and 0.04, respectively). Furthermore, HLF-ICG-Krem values ≤0.05 detected two patients with higher INRs after postoperative day five. On the other hand, neither rHUI nor Cv(rRLE) was correlated with INRs after postoperative day five (r = 0.28, and -0.03, respectively; p >0.05 for both). HLF-ICG-Krem was significantly lower with PHLF than without PHLF (p = 0.005). Conclusion: HLF-ICG-Krem is useful for evaluating PHLF more correctly.
Preoperative evaluation of future remnant liver (FRL) function is crucial in determining whether a patient can safely undergo liver resection. One reliable index for estimating posthepatectomy liver failure (PHLF) is indocyanine green (ICG) clearance of the FRL (ICG-Krem), which is calculated as the ratio of FRL volume to total liver volume [
Gadoxetic acid disodium is a developed magnetic resonance (MR) contrast agent for hepatocellular imaging. Several reports have discussed liver function assessments using MR images acquired using a variety of liver- specific contrast agents [
This prospective study was approved by the institutional review committee of our institution and all patients provided written informed consent prior to participation. From April 2013 to March 2015, 20 patients (15 men, 5 women; age range, 43 - 82 years; median age, 67 years) with HCC (8 patients) and metastatic liver tumors (12 patients) underwent gadoxetic acid disodium-enhanced MRI and surgery for tumor excision; among these eight underwent a partial hepatectomy. Of these, one patient had grade 1 fibrosis, one had grade 3, and one had grade 4. All patients without portal hypertension had Child-Pugh class A scores [
The medical records of all 20 patients were reviewed to determine serum albumin levels, total bilirubin levels, international normalized ratios (INRs), and platelet counts from samples acquired within one week before or after MRI.ICG clearance tests were examined within two weeks before or after MRI. A 0.5 mg/kg ICG dose was administered intravenously and venous blood was sampled in which 5 ml before and at 5, 10, and 15 min after ICG administration [
PHLF was evaluated after postoperative day five using increased INRs that were defined according to the ISGLS. In this grading system, PHLF is characterized by increased INRs and concomitant hyperbilirubinemia on or after postoperative day five. The severity of PHLF is graded according to its impact on clinical management. Grade A PHLF requires no change in the patient’s clinical management. The clinical management of patients with grade B PHLF deviates from the regular course but does not require invasive therapy. Grade C PHLF defines the need for invasive treatment. We made preoperative PHLF predictions according to the HLF-ICG- Krem values.
All patients underwent MRI using a Signa HDxt 3.0Tclinical scanner with a superconducting magnet (GE Healthcare, Milwaukee, WI, USA) and an 8-channel phased-array coil. Dynamic images using fat-suppressed T1-weighted gradient-echo images with a 3-dimensional acquisition sequence (liver acquisition with volume acceleration) were acquired before and at 35 s, 90 s, 180 s, and 20 min after the administration of gadoxetic acid disodium (0.1 mL/kg body weight). The gadoxetic acid disodium was administered intravenously as a bolus at a rate of 1 mL/s through an intravenous cubital line that was flushed with 40 mL of saline via a power injector (Dual Shot GX, Nemotokyorindo, Tokyo, Japan). Images were acquired in the transverse plane at a section thickness of 3.8 mm (zero-fill interpolation: zip, 2). The parameters were as follows: repetition time/echo time, 3.1/1.4 s; flip angle, 12˚; number of signals acquired, 1; field of view, 38 cm; matrix, 256 × 224 (512 zip); and acquisition time, 20 s.
MRI-based liver function indices were the FRL-hepatocellular uptake index (rHUI), RLE, FRL-RLE (rRLE), coefficient variation of RLE [Cv(RLE)], coefficient variation of rRLE [Cv(rRLE)], and ICG-Krem corrected by HLF (HLF-ICG-Krem). In 20 patients who had PHLF, we investigated the correlations after postoperative day five between increased INRs and rHUI, rRLE, Cv(rRLE), and HLF-ICG-Krem.
rHUI values were defined on hepatobiliary phase images as follows:
where rVLis the volume of the FRL. rL20 and S20 are the signal intensities of FRL, and spleen in the hepatobiliary phase, respectively.
Two sets of axial images were acquired from two types of liver images: precontrast enhancement images and hepatobiliary phase images. Hepatobiliary phase images were acquired 20 min after the intravenous administration of gadoxetic acid disodium. The RLE on hepatobiliary phase images were considered to primarily reflect hepatocellular enhancement effects. RLE values were defined as the signal gain percentage between the precontrast images and hepatobiliary phase images as follows:
where SIH and SIP are the signal intensities in the hepatobiliary phase images and precontrast enhanced images, respectively. RLE images were constructed on the basis of RLE values at each pixel and calculated using equation (2). Therefore, the signal intensity of each pixel was representative of the corresponding RLE value in the RLE images, which were displayed according to the National Institutes of Health color look-up table.
HLF was evaluated using the Cv(RLE) and was defined as follows:
ICG-Krem in consideration of HLF (HLF-ICG-Krem) was defined as follows:
Correlations were evaluated between rHUI, rRLE, Cv(rRLE), HLF-ICG-Krem, and INRs after postoperative day five.
RLE values and liver parenchymal volumes were measured by two radiologists (Y.S. and S.M., with 18 and three years of experience reading MR images, respectively) and a PhD researcher (S.M., with 31 years of experience reading MR images). Region of interest (ROI) and volume of interest (VOI) acquisitions were performed using OsiriX (ver. 2.7.5) medical imaging software (http://www.osirix-viewer.com). In the whole liver and FRL, hepatobiliary phase images were used to define ROIs and VOIs for the tumor-free liver parenchyma by freehand contouring. The major vessels in the liver, such as the portal and hepatic veins, were excluded during ROI acquisition (
Statistical analyses were performed using the software package SPSS for Macintosh (Version 16.0, SPSS Inc., Chicago, IL, USA). The Spearman correlation coefficient was used to evaluate correlations between rHUI, rRLE, Cv(rRLE), HLF-ICG-Krem, and INR values after postoperative day five. The Mann-Whitney test and Tukey’s multiple comparison test were used to evaluate differences in rHUI and HLF-ICG-Krem values according to the classified PHLF grade. In the preoperative evaluation of PHLF, HLF-ICG-Krem values and PHLF grades were calculated and compared by receiver operating characteristic (ROC) curves. Two-tailed p values <0.05 were considered to be statistically significant.
INR after postoperative day 5 | |||
---|---|---|---|
Indices | r | 95% confidence interval | P value |
HLF-ICG-Krem | −0.55 | −0.81 to −0.13 | 0.01 |
rREL | 0.46 | −0.05 to 0.75 | 0.04 |
Cv(rREL) | −0.03 | −0.48 to 0.43 | 0.89 |
rHUI | 0.28 | −0.20 to 0.65 | 0.24 |
HLF-ICG-Krem: heterogeneous liver function ICG-Krem, rRLE: Future remnant liver RLE, Cv(rREL): Coefficient variation rRLE, rHUI: remnant hepatocellular uptake index, REL: Relative enhancement of the liver.
Cace No. | Sex | Age | Disease | Hepatectomy | ICG K | INR after postoperative five day | Bilirubin after postoperative five day | PHLF grade | Fibrosis grade | rVL/VL | Cv(RLE)/Cv(rRLE) | ICG-Krem | HLF-ICG-Krem | RLE (%) | rRLE(%) | Cv(rRLE) | rHUI | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | F | 43 | HCC | left hepatectomy | 0.177 | 1.28 | 1.6 | A | F0 | 0.531 | 0.730 | 0.094 | 0.069 | 143.2 | 155.2 | 27.45 | 483.9 | |
2 | M | 72 | HCC | Partial hepatectomy | 0.120 | 1.28 | 2.2 | A | F0 | 0.737 | 1.000 | 0.088 | 0.088 | 150.8 | 151.7 | 21.95 | 698.1 | |
3 | M | 62 | HCC | right hepatectomy | 0.173 | 1.47 | 1.1 | B | F4 | 0.350 | 0.708 | 0.061 | 0.043 | 111.8 | 106.7 | 38.52 | 850.6 | |
4 | M | 65 | HCC | central bisegmentectomy | 0.118 | 1.98 | 3.6 | B | F0 | 0.439 | 0.926 | 0.052 | 0.048 | 128.8 | 138.9 | 51.33 | 373.8 | |
5 | F | 69 | HCC | medial left segmentectomy | 0.172 | 1.20 | 0.4 | No | F3 | 0.605 | 0.978 | 0.104 | 0.102 | 164.3 | 164.8 | 21.97 | 403.7 | |
6 | F | 59 | Metastasis | lateral segmentectomy | 0.196 | 1.00 | 0.8 | No | F0 | 0.698 | 0.762 | 0.137 | 0.104 | 147.3 | 149.5 | 33.04 | 569.2 | |
7 | M | 60 | Metastasis | right hepatectomy | 0.181 | 1.11 | 0.7 | No | F0 | 0.406 | 1.552 | 0.073 | 0.114 | 132.2 | 128.4 | 16.28 | 215.5 | |
8 | M | 64 | Metastasis | partial hepatectomy | 0.223 | 1.21 | 1.1 | No | F0 | 0.946 | 0.955 | 0.211 | 0.201 | 168.6 | 167 | 21.80 | 639.8 | |
9 | M | 63 | Metastasis | lateral segmentectomy | 0.193 | 1.18 | 0.8 | No | F0 | 0.803 | 0.904 | 0.155 | 0.140 | 198.9 | 203.2 | 18.01 | 1201.4 | |
10 | M | 82 | HCC | partial hepatectomy | 0.182 | 1.20 | 0.7 | No | F1 | 0.774 | 0.840 | 0.141 | 0.118 | 140.2 | 140 | 27.07 | 896.8 | |
11 | M | 60 | HCC | medial left segmentectomy | 0.178 | 1.20 | 1.2 | No | F0 | 0.777 | 0.973 | 0.138 | 0.135 | 172.3 | 177.1 | 23.26 | 1208.1 | |
12 | M | 68 | HCC | right hepatectomy | 0.191 | 1.27 | 0.7 | No | F0 | 0.361 | 0.941 | 0.069 | 0.065 | 153.3 | 153.4 | 20.86 | 298.2 | |
13 | M | 71 | Metastasis | partial hepatectomy | 0.149 | 1.10 | 0.5 | No | F0 | 0.880 | 0.816 | 0.131 | 0.107 | 98.8 | 100.1 | 31.37 | 130.4 | |
14 | F | 46 | Metastasis | partial hepatectomy | 0.220 | 1.16 | 0.3 | No | F0 | 0.749 | 0.586 | 0.165 | 0.097 | 94.4 | 97 | 46.08 | 190.7 | |
15 | M | 69 | Metastasis | partial hepatectomy | 0.131 | 0.99 | 1.5 | No | F0 | 0.895 | 0.937 | 0.117 | 0.110 | 91.6 | 91.8 | 39.98 | 514.1 | |
16 | F | 70 | Metastasis | posterior sectionectomy | 0.255 | 0.96 | 1.0 | No | F0 | 0.727 | 0.770 | 0.185 | 0.143 | 118.1 | 127.6 | 48.35 | 523.7 | |
17 | M | 75 | Metastasis | partial hepatectomy | 0.147 | 1.10 | 1.0 | No | F0 | 0.863 | 1.201 | 0.127 | 0.152 | 102.3 | 105.7 | 8.14 | 779.1 | |
18 | F | 68 | Metastasis | posterior sectionectomy | 0.210 | 1.05 | 0.4 | No | F0 | 0.772 | 0.972 | 0.162 | 0.158 | 126 | 124.4 | 20.50 | 288.6 | |
19 | M | 60 | Metastasis | left hepatectomy | 0.148 | 1.16 | 1.0 | No | F0 | 0.749 | 1.176 | 0.111 | 0.130 | 80.8 | 84.2 | 29.57 | 208.2 | |
20 | M | 73 | Metastasis | left lateral segmentectomy | 0.135 | 1.08 | 0.6 | No | F0 | 0.832 | 1.100 | 0.112 | 0.124 | 113.2 | 113.9 | 21.60 | 63.0 |
PHLF: posthepatectomy liver failure, VL: Volume of whole liver, rVL: Volume of whole future remnant liver, RLE: relative liver enhancement, rRLE: Future remnant liver RLE, Cv(REL): Coefficient variation RLE, Cv(rREL): Coefficient variation rRLE, HLF-ICG-Krem: heterogeneous liver function ICG-Krem, rHUI: remnant hepatocellular uptake index.
n | HLF-ICG-Krem | rRLE | ||
---|---|---|---|---|
No-PHLF | 16 | 0.125 ± 0.031 | 133.0 ± 34.1 | |
PHLF | 4 | 0.062 ± 0.021 | * | 138.1 ± 22.1 |
PHLF grade A | 2 | 0.079 ± 0.014 | 153.5 ± 2.5 | |
PHLF grade B | 2 | 0.045 ± 0.004 | ** | 122.8 ± 22.8 |
*: significant difference in No-PHLF and PHLF by Mann Whitney test; **: significant difference in No-PHLF and PHLF gred A, B by Tukey’s Multiple comparison test; HLF-ICG-Krem: heterogeneous liver function, rRLE: relative liver enhancement of future remnant liver. PHLF: posthepatectomy liver failure.
No-PHLF vs. PHLF | No-PHLF vs. PHLF B | |
---|---|---|
Cut off value | 0.05 | 0.05 |
Sensitivity (%) | 50 (2/4) | 100 (2/2) |
Specificity (%) | 100 (16/16) | 100 (18/18) |
Efficiency | 0.8 (16/20) | 1 (20/20) |
Prevalence | 0.2 (4/20) | 0.1 (2/18) |
Positive predictive value (%) | 100 (2/2) | 100 (2/2) |
Negative predictive value (%) | 89 (16/18) | 100 (18/18) |
100%, and 89%, respectively. The sensitivity, specificity, efficiency, prevalence, positive predictive value, and negative predictive value using a threshold HLF-ICG-Krem value of 0.05 that determined the presence of PHLF grade B were 100%, 100%, 1%, 0.1%, 100%, and 100%, respectively.
Liver function is usually evaluated using laboratory data obtained by blood sampling.ICG is one of the important test that comprehensively evaluates hepatic function, including uptake, metabolism, and excretion [
Preoperative evaluations of FRL function are crucial for determining whether a patient can safely undergo liver resection. Numerous attempts have been made to define prognostic factors that would facilitate liver failure risk assessments [
This study included both patients with metastatic liver tumors and those with HCC, and the liver excision methods included both major and partial liver resection. Both rRLE and HLF-ICG-Krem values correlated with INRs after postoperative day five. However, neither Cv(rRLE) nor rHUI was correlated with INRs after postoperative day five. This result suggests that rRLE and HLF-ICG-Krem are parameters that could be used to evaluate PHLF independently of the case or operation method.
HLF-ICG-Krem was significantly lower in patients with PHLF than in those without PHLF (p < 0.05). When the HLF-ICG-Krem threshold value was set at 0.05, HLF-ICG-Krem detected two patients among four with PHLF, which was not detected by ICG-Krem. HLF-ICG-Krem was significantly lower in patients with PHLF grade B than in those without PHLF (p < 0.05), at the <0.05 threshold value. The ability to predict which patients will develop PHLF grade B prior to surgery is clinically meaningful because the postoperative management of such patients deviates from the regular course. This study has some limitations. First, the study population included a small sample size, and second, the study population included few cases of PHLF grade A or B, and no cases of grade C.
In summary, the HLF-ICG-Krem value incorporates HLF into the standard ICG-Krem value and thus provides a useful parameter that can evaluate PHLF more correctly. This MRI examination method also provides morphological and functional information that can be used to optimize the preoperative selection of patients who can safely undergo liver resection.
The authors thank Mamoru Sogami, MD, PhD for providing valuable advice and all the study participants for their cooperation.
This work was supported by JSPS KAKENHI (26461844).
Shigeru Matsushima,Yozo Sato,Hidekazu Yamaura,Mina Kato,Yui Onoda,Shinichi Murata,Yasuhiro Shimizu,Yasutomi Kinosada,Hideyuki Nishiofuku,Yoshitaka Inaba, (2016) Preoperative Evaluation of Posthepatectomy Liver Failure Using MRI-Based Liver Function Indices in Child-Pugh Class A Patient. Open Journal of Radiology,06,147-156. doi: 10.4236/ojrad.2016.62022