Prediction of posthepatectomy liver failure using transient elastography in patients with hepatitis B related hepatocellular carcinoma

Background It is essential to accurately predict Postoperative liver failure (PHLF) which is a life-threatening complication. Liver hardness measurement (LSM) is widely used in non-invasive assessment of liver fibrosis. The aims of this study were to explore the application of preoperative liver stiffness measurements (LSM) by transient elastography in predicting postoperative liver failure (PHLF) in patients with hepatitis B related hepatocellular carcinoma. Methods The study included 247 consecutive patients with hepatitis B related hepatocellular carcinoma who underwent hepatectomy between May 2015 and September 2015. Detailed preoperative examinations including LSM were performed before hepatectomy. The endpoint was the development of PHLF. Results All of the patients had chronic hepatitis B defined as the presence of hepatitis B surface antigen (HBsAg) for more than 6 months and 76 (30.8%) had cirrhosis. PHLF occurred in 37 (14.98%) patients. Preoperative LSM (odds ratio, OR, 1.21; 95% confidence interval, 95% CI: 1.13–1.29; P < 0.001) and international normalized ratio (INR) (OR, 1.07; 95% CI: 1.01–1.12; P < 0.05) were revealed to be independent risk factors for PHLF, and a new model was defined as LSM-INR index (LSM-INR index = 0.191*LSM + 6.317*INR-11.154). The optimal cutoff values of LSM and LSM-INR index for predicting PHLF were 14 kPa (AUC 0.86, 95% CI: 0.811–0.901, P < 0.001) and −1.92 (AUC 0.87, 95% CI: 0.822–0.909, P < 0.001), respectively. Conclusions LSM can be helpful for surgeons to make therapeutic decisions in patients with hepatitis B related hepatocellular carcinoma.


Background
Hepatocellular carcinoma (HCC) is one of the common malignant tumors worldwide. Surgical resection is the most effective treatment for patients with localized HCC [1,2]. However, postoperative liver failure (PHLF) is a life-threatening complication and intrinsic risk of mortality [3]. It is not only correlated with the volumes of liver resection, but also the insufficient function of hepatic reserve (FHR) [4].
More than 80% of HCCs arise in patients with hepatic fibrosis or cirrhosis [5], which has a very important impact on liver function [4]. Hence, we attempted to determine whether FHR can be indirectly predicted by assessing the degree of liver fibrosis.
LSM is a technology involving obtaining the liver instantaneous elastic modulus to estimate the degree of liver fibrosis by transient elastography (TE), an easy and noninvasive method with high accuracy [6][7][8].
However, baseline characteristics of patients differed greatly among these studies, and few of the studies focused on PHLF. Unlike those in the western countries, about 80% HCC patients in China had hepatitis B virus infection [13].

Aim of the study
In this study, we aimed to assess the usefulness of liver stiffness measured by transient elastography Fibro Touch® (Wuxi HISKY Medical Technologies Co., Ltd. Beijing, China) for predicting PHLF in patients with hepatitis B-related hepatocellular carcinoma.

Patients
This study was approved by the Institutional Ethics Committee of the Eastern Hepatobiliary Surgery Hospital (EHBH).
HCC patients who underwent liver resection at EHBH were prospectively recruited between May 2015 and September 2015. The exclusion criteria were as follows: (i) patients with hepatolithiasis or patients who will receive hepatectomy because of intrahepatic cholangiocarcinoma (ICC) or hepatic maligancies other than HCC; (ii) patients with cirrhosis due to schistosomiasis, alcoholic liver disease or non-alcoholic fatty liver disease (NAFLD); (iii) patients undergoing preoperative transhepatic arterial chem otherapy and embolization (TACE).

Transient elastography
All patients fasted for at least 6 h before receiving LSM examination by transient elastography FibroTouch®. The examination was performed by two trained and certified operators who were blinded to the patients' clinical data, according to the operation manual and the Liver Stiffness Study Group "Elastica" of the Italian Association for the Study of the Liver [14]. LSM was expressed in kiloPascals (kPa) and was considered reliable only if 10 successful measurements were obtained, with an IQR/median of LSM of < 30% and a success rate of > 60% [15].

Liver surgery
During surgery, right costal margin incision was chosen and the fluid infusion was minimal to keep central venous pressure lower than 5 mmHg to reduce bleeding from hepatic veins [16,17]. Intraoperative ultrasound (US) was performed systematically to detect the presence of any additional nodules not detected preoperatively. Major hepatectomy was defined as removal of 3 or more Couinaud segments [18,19]. Diuretics and Ampicillin were used for routine postoperative care.
PHLF was defined as the presence of at least one of the following variables: 1) occurrence of refractory ascites causing a delay in the removal of surgical drainages and/or postoperative drainage exceeding 500 ml/day, a continuous elevation of total serum bilirubin concentration (≥60umol/ l) beyond postoperative day 7; 2) alteration of coagulation factors requiring fresh frozen plasma infusion with an International Normalized Ratio (INR) of more than 1.50 [20]. The endpoint of this study is the presence of PHLF.

Statistical analysis
Continuous variables were expressed as the mean and standard deviation. Differences between the subgroups were compared by t-test or Mann-Whitney U test.

Characteristics of the study population
The demographic and clinicopathologic characteristics of these patients were shown in Table 1.
All the patients (213 men and 34 women), with a mean age of 53.27 ± 10.33 years, had postive HBsAg lasting for > 6 months. All of them were Child-Pugh class A, and a small proportion had complications (17%). As shown by the pathological results, 76 (31%) patients had cirrhosis and 18 (7%) had esophageal varices consistent with gastroscopy results. 37 (14.98%) patients developed PHLF postoperatively, and they had a significantly higher preoperative mean LSM (21.4 ± 6.3 kPa) than those without PHLF (12.7 ± 5.7 kPa, P < 0.001).

Independent predictors for PHLF in HCC patients
Univariate and multivariate analysis were used for analyzing the potential influencing factors associated with PHLF, and the results were reported in Table 2. Univariate analysis revealed that the factors including platelet count (p = 0.037), prealbumin (p = 0.032), gammaglutamyl transpeptidase (p = 0.027), prothrombin time (p = 0.005) and INR (p = 0.005), LSM (p < 0.001), the use of intraoperative blood transfusions (p = 0.005), the presence of cirrhosis (p < 0.001) and esophageal varices (p = 0.003) were significant predictors for PHLF.

Diagnostic performance of LSM for predicting PHLF
The diagnostic performance and corresponding ROC curves of LSM are shown in Fig. 1 (Fig. 2).

Discussion
Surgical resection is the first-line therapeutic option for early HCC [21]. However, insufficient FHR may result in postoperative complications and even PHLF (4), which is a major cause of postoperative morbidity and mortality after elective hepatic resection [22,23].
Liver resection for HCC patients with chronic liver diseases still carries higher risk of PHLF than normal liver resection [22,24].To the best of our knowledge, this is the first published study on the effectiveness of LSM measurement in predicting PHLF in patients with hepatitis B related hepatocellular carcinoma since previous studies only suggested a potential role of LSM in predicting postresection hepatic insufficiency. Furthermore, in those previous studies [10,25], the background hepatic conditions causing HCC showed a great variability and included HCV, alcoholic and nonalcoholic steatohepatitis in addition to HBV. The baseline characteristics of patients differed by studies and it may result in a selection bias.
In this study, we used a new generation transient elastography, FibroTouch® for liver stiffness measurement (LSM) [9,26], which has enhanced 2D-image-guided positioning function. It's particularly advantageous for precise preoperative examination, because the examiners are able to set the the region of interest(ROI) in the non-tumor area. It may also explain why the optimal cut-off LSM value (14.0 Kpa) in this study is lower than that in previous studies [10,12]. Although the conclusions of their studies were similar to that of this study, some differences have to be pointed out. First, although the study population was very similar to ours, the background of the population was not homogenous. Secondly, their definition of PHLF was only based on postoperative serum bilirubin levels, which configures a very high risk of irreversible PHLF (5 mg/dL for more than 5 days post operation) [27]. Perhaps for this reason, their LSM cutoff was higher and may miss milder grades of PHLF [25].
Using the calculated cutoff value of 14.0 Kpa, LSM had high specificity and negative predictive value for predicting PHLF. The value in liver failure group was significantly higher than that in non-liver failure group (21.4 Kpa vs. 12.7Kpa). It implied that FHR may become worse with the increase of LSM value. While the effectiveness of LSM for diagnosing cirrhosis is not high (AUC = 0.78, 95% confidence interval: 0.727-0.834), it may be confounded by the enrollment of patients with Child-Pugh Class A. Precisely because of this reason, one interesting result of this study is that the degree of resection is not a significant predictor of PHLF.  In our study, HBV DNA showed no statistical differences, which is not in line with a previous report [28], and a possible explanation is the inclusion of patients with HBV DNA higher than 10 4 IU/ml. For such patients, we will first use antiviral drugs to control the amount of HBV, and then choose surgery. Univariate analysis showed that platelet count, LSM and others were significant prognostic factors for PHLF. This is consistent with the previous finding of risk factors for postoperative complications [29].
We must acknowledge this study has several potential limitations. Compared to similar articles, LSM was significantly better than ICG-15 and MELD score in the prediction of postoperative complications [10,12], we do not routinely use ICG-15 or MELD to assess liver function. The number of the outcome of PHLF was relatively small, and further acquisition of cases and the external validation should be accomplished in the future. Second, our analysis did not include other variables that may affect the outcomes of surgery, such as the resected liver volume which is closely correlated with the functional liver reserve and postoperative results [11]. A welldesigned, well-controlled, randomized study of a large population is required.

Conclusions
In summary, our study showed that the preoperative LSM is a valid tool for surgeons in making therapeutic decisions in patients with hepatitis B related hepatocellular carcinoma. We also established a new index, LSM-INR index, which quantitatively evaluated the risk of the INR and LSM that should be useful for surgeons in making therapeutic decisions in patients with hepatitis B related hepatocellular carcinoma before hepatectomy.