Prognostic value of postoperative change in liver stiffness in patients with HBV-related hepatocellular carcinoma.

Introduction

Due to the prevalence of hepatitis B virus (HBV) and alcohol consumption, hepatocellular carcinoma (HCC) has become one of the most common cancers, with an increasing incidence worldwide, particularly in China.[1,2] Ablation and liver transplantation are effective in treating HCC, however, surgical resection remains the most important therapy, and is conducted in most patients with early stage HCC.[3-5] Despite the effectiveness of these treatments, recurrence and metastasis are still major problems for long-term survival in patients with HCC.[6-8]

Chronic HBV infection is well-known to be associated with the development of cirrhosis and HCC, resulting in hepatic disease-related deaths.[9-11] The incidence of HCC is reported to be 0.5% in patients infected with HBV, and ranges between 2.5% and 6% in patients with cirrhosis.[6,12-14] In Asia, over 60% of patients with HCC have an aetiology of chronic HBV infection.[15-17] Liver stiffness, measured by transient elastography, has proven to be an effective non-invasive procedure for evaluating liver fibrosis and assessing portal vein hypertension.[18-21] Liver stiffness can be measured using a FibroScan™ device (Echosens, Paris, France) that detects the propagation speed of an elastic sheer wave triggered by a transducer, which is then converted to a measurement of liver stiffness.[22] High levels of liver stiffness have been reported to be associated with the risk of HCC occurrence.[23,24]

To the best of the authors’ knowledge, there is no published study to date that evaluates the prognostic effectiveness of liver stiffness measurements in patients with HCC, and thus, the prognostic value of liver stiffness in patients with HCC remains a clinical issue that requires exploration. The aim of the present study was to assess the prognostic value of change in liver stiffness from preoperative values, within one year following hepatic resection in patients with HBV-related HCC.

Patients and methods

Study population

This retrospective study included consecutive patients who were diagnosed with HCC and who had received liver surgery at Heze Municipal Hospital (Shandong, China) between February 2011 and May 2015. Anti-virus treatment and postoperative clinical examinations were conducted at Yantai Infectious Disease Hospital and Yantai Yuhuangding Hospital (both in Yantai, Shandong, China). Liver stiffness measurements, using Fibroscan™ (Echosens), were conducted at Jining No.1 People’s Hospital (Jining, Shandong, China). Inclusion criteria were: (1) Histologically confirmed HCC with HBV infection; (2) Child–Turcotte–Pugh score ≤9; (3) Single lesion located in the liver; and (4) Valid clinical characteristics and laboratory outcomes. Exclusion criteria were: (1) Coinfection with hepatitis C virus and/or human immunodeficiency virus; (2) Alcoholic hepatic diseases; (3) Schistosomiasis; and (4) Invalid clinical characteristics and laboratory outcomes.

This study was conducted under compliance with the Declaration of Helsinki and was approved by the Human Ethics Committees of Yantai Infectious Disease Hospital, Yantai Yuhuangding hospital and Jining No.1 People’s Hospital (Approval No. YIDH-201703-K3). All patients provided written informed consent prior to enrolment into the study.

Transient elastography

All patients underwent liver stiffness measurements by transient elastography on at least two occasions: one prior to surgery (baseline) and the other within 1 year following liver resection surgery. Transient elastography was conducted by Fibroscan™ (Echosens) as previously described.[25] Results were expressed as kPa and the median value of 10 successful measurements in each patient was used as the liver stiffness value.

Diagnosis and anti-virus treatment

Fibrosis and HCC were diagnosed by histopathology of tissue specimens acquired during liver surgery. The maximum tumour diameter was calculated with images obtained during preoperative contrast-enhanced computed tomography (CT) using a GE LightSpeed VCT system (GE Healthcare, Tokyo, Japan). HBV infection was diagnosed with a positive serum viral marker and/or elevated serum HBV-DNA level (>1000 copies/ml during two consecutive detections). Briefly, serum HBV DNA was detected using a Daan test (Daan Gene Co., Ltd., Guangdong, China) according to the manufacturer’s instructions. DNA was extracted from 100 μl of serum and detected by real-time polymerase chain reaction (PCR) with HBV-DNA specific fluorescence probing, using a Roche LightCycler 480 system (Roche Diagnostics Ltd., Rotkreuz, Switzerland). HCC recurrence during follow-up was screened by contrast-enhanced CT and ultrasonography (GE Volusion™ E10 ultrasound system; GE Healthcare, Shanghai, China). Blood laboratory parameters were obtained for each patient, and Child–Turcotte–Pugh score classification was applied for consideration of prognosis, as previously described.[25,26] All patients received anti-HBV regimens (entecavir) for at least 6 months prior to surgery and after surgery, as previously described.[27,28]

Study outcomes

The primary outcome of this study was change in liver stiffness following hepatic resection, from preoperative (baseline) value, as a prognostic marker in patients with HBV-related HCC. Secondary outcomes included independent factors associated with prognosis following hepatic resection among patients with HBV-related HCC.

Statistical analyses

Data are presented as n (%) prevalence, median (range) or mean ± SD. Student’s t-test and Wilcoxon signed–rank test was used to analyse continuous variables with or without normal distribution, respectively. Categorical variables were analysed using χ2-test and Fisher’s exact test. Survival was assessed using Kaplan-Meier estimator and Log-rank test, as appropriate. Multivariate analysis was performed using Cox proportional hazards model. All statistical analyses were conducted using SPSS software, version 16.0 (SPSS Inc., Chicago, IL, USA) and a P value < 0.05 was regarded as statistically significant.

Results

Baseline patient characteristics

Out of 186 patients initially enrolled, a total of 158 patients were included in the final analyses (Figure 1; Table 1). Male patients were predominant (n = 123 [77.85%]), and median age was 50 years (range, 22–74 years). More than half of the patients were confirmed by pathology to have fibrosis (67.72%). Most patients were Child–Turcotte–Pugh class A (n = 147 [93.04%]), and all patients received anti-virus therapy prior to hepatic resection. The median serum HBV DNA level was 4.1 log copies/ml. A total of 102 patients (64.56%) had serum HBV DNA level <3 log copies/ml (the lower limit of detection) prior to surgery, and 56 patients (35.44%) were serum HBV DNA positive. The median maximum tumour diameter was 4.1 (range, 3.0–5.3) cm, and median liver stiffness was 6.7 (range, 4.3–12.5) kPa (Table 1).

Figure 1.

Flow chart showing selection of the study population. HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HDV, hepatitis D virus.

Table 1.

Baseline characteristics of patients with hepatitis B virus-related hepatocellular carcinoma.

Variable	Study population, n = 158
Sex, male	123 (77.85)
Age, years	50 (22–74)
Fibrosis	107 (67.72)
Cirrhosis	81 (51.27)
CTP class
A	147 (93.04)
B	11 (6.96)
C	0 (0)
Tumour maximum size, cm	3.6 (1–8)
HBV DNA, log copies/ml	4.1 (3.0–5.3)
HBeAg positive	96 (60.76)
AFP, ng/ml	347.82 (1.19–1210)
TBIL, mmol/l	6.87 (2.60–24.70)
ALT, IU/l	24 (11–46)
AST, IU/l	21 (17–58)
ALP, IU/l	69 (45–203)
ALB, g/l	4.2 (2.8–6.1)
Platelet count, × 109/l	193.67 (89–267)
Prothrombin time, s	12.7 (11.0–15.6)
INR	0.97 (0.78–1.54)
Liver stiffness, kPa	6.7 (4.3–12.5)

Data presented as n (%) prevalence or median (range).

AFP, α-fetoprotein; ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CTP, Child–Turcotte–Pugh score; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; INR, international normalized ratio; TBIL, total bilirubin.

Demographic and clinical characteristics in patients with increased or decreased postoperative liver stiffness

All patients received liver stiffness assessments on at least two occasions. Mean baseline liver stiffness was 8.53 ±2.41 Kpa and was 9.26 ± 2.29 Kpa after surgery. Compared with baseline values prior to surgery, 98 patients (62.03%) experienced increased liver stiffness (mean change, 3.8 ± 1.3 kPa) and 60 patients (37.97%) had decreased liver stiffness (mean change, 2.4 ± 1.1 kPa) within 12 months following surgery (dynamic changes are shown in Figure 2). In patients categorised according to postoperative increase or decrease in liver stiffness, there was a statistically significant between-group difference in platelet count only (P = 0.04; Table 2).

Figure 2.

Dynamic changes in liver stiffness values from baseline (before surgery) to within 1 year after surgery in 158 patients with hepatitis B virus-related hepatocellular carcinoma. Mean ± SD liver stiffness was 8.53 ± 2.41 Kpa at baseline and 9.26 ± 2.29 Kpa post-surgery; 98 patients (62.03%) experienced increased liver stiffness (mean change, 3.8 ± 1.3 kPa) and 60 patients (37.97%) had decreased liver stiffness (mean change, 2.4 ± 1.1 kPa) following surgery.

Table 2.

Comparison of baseline variables in 158 patients with increased or decreased liver stiffness following surgery.

Variable	Study subgroup		Statistical significance
	Increased LS	Decreased LS
Patients	98 (62.03)	60 (37.97)
Sex, male	71 (72.45)	52 (86.67)	NS
Age, years	48 (27–73)	50 (22–74)	NS
Fibrosis	68 (69.39)	39 (65.00)	NS
CTP class			NS
A	89 (90.82)	58 (96.67)
B	9 (9.18)	2 (3.33)
C	0	0
Tumour maximum size, cm	4.7 (2.4–8)	3.3 (1–8)	NS
HBV DNA, log copies/ml	3.7 (3.0–4.7)	3.9 (3.0–5.3)	NS
HBeAg positive	61 (62.24)	31 (58.33)	NS
AFP, ng/ml	379.42 (2.56–1210)	319.87 (1.19–1210)	NS
TBIL, mmol/l	4.78 (2.60–24.70)	7.41 (3.90–21.60)	NS
ALT, IU/l	21 (11–38)	27 (19–46)	NS
AST, IU/l	28 (17–46)	18 (17–58)	NS
ALP, IU/l	71 (47–203)	58 (45–102)	NS
ALB, g/l	4.7 (2.8–5.7)	3.9 (2.8–6.1)	NS
Platelet count, ×109/l	189.77 (103–267)	213.52 (89–241)	P = 0.04
Prothrombin time, s	13.2 (11.0–14.7)	12.6 (11.0–15.6)	NS
Differentiation			NS
Well–moderate	48 (49.0)	31 (51.7)
Poor–undifferentiated	50 (51.0)	29 (48.3)
Vascular invasion			NS
Yes	22 (22.4)	16 (26.7)
No	76 (77.6)	44 (73.3)
INR	1.01 (0.87–1.54)	0.88 (0.78–1.21)	NS

Data presented as n (%) prevalence or median (range).

AFP, α-fetoprotein; ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CTP, Child–Turcotte–Pugh score; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; INR, international normalized ratio; LS, liver stiffness; TBIL, total bilirubin.

NS, no statistically significant between-group difference (P > 0.05).

Prognostic performance of liver stiffness changes in patients with HCC

Patient outcomes were analysed using Kaplan–Meier estimator in patients categorised according to postoperative increase or decrease in liver stiffness, in order to investigate the potential difference in prognosis between the two groups. Patients with increased postoperative liver stiffness were revealed to have significantly worse overall survival outcomes than those with decreased postoperative liver stiffness (P = 0.042). Similar trends were observed for diseases-free survival and recurrence outcomes, which showed that patients with increase postoperative liver stiffness had significantly worse outcomes than those with decreased liver stiffness (P = 0.045 for disease-free survival and P = 0.041 for recurrence outcomes; Figure 3).

Figure 3.

Kaplan-Meier estimator curves in 158 patients with hepatitis B virus-related hepatocellular carcinoma, categorised according to increased or decreased postoperative liver stiffness (LS), showing: (a) significantly worse overall survival outcomes in patients with increased LS versus those with decreased LS (P = 0.042); (b) significantly worse diseases-free survival in patients with increased LS versus those with decreased LS (P = 0.045); and (c) significantly worse recurrence outcomes in patients with increased LS versus those with decreased LS (P = 0.041).

Univariate and multivariate analyses of prognostic variables in patients with HCC

The potential correlations between clinical parameters and overall survival, disease-free survival and recurrence were analysed using Cox proportional hazards model in patients with HCC. Univariate analyses showed that age, Child–Turcotte–Pugh score and postoperative change in liver stiffness were all prognostic variables for overall survival. Multivariate analysis revealed that only Child–Turcotte–Pugh score (hazard ratio [HR] 1.209, P = 0.039) and liver stiffness changes (HR 1.891, P = 0.042) were independent prognostic variables that were associated with overall survival (Table 3).

Table 3.

Univariate and multivariate analyses of prognostic variables for overall survival in patients with hepatitis B virus-related hepatocellular carcinoma.

Variable	Univariate analysis			Multivariate analysis
	HR	95% CI	Statistical significance	HR	95% CI	Statistical significance
Age	0.935	0.165, 0.974	P = 0.004
Sex	0.805	0.236, 2.110	NS
CTP score	1.235	1.156, 2.502	P = 0.011	1.209	1.010, 1.742	P = 0.039
HBV DNA	0.905	0.818, 1.290	NS
HBeAg status	1.606	0.270, 1.962	NS
AFP level	1.604	0.516, 2.021	NS
Platelet count	1.363	0.296, 2.366	NS
Total bilirubin	2.129	0.724, 2.363	NS
LS changes	2.112	1.839, 2.326	P = 0.003	1.891	1.441, 2.620	P = 0.042

AFP, α-fetoprotein; CTP, Child–Turcotte–Pugh score; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; LS, liver stiffness; HR, hazard ratio; CI, confidence interval.

NS, no statistically significant correlation (P > 0.05; Cox proportional hazards).

To evaluate whether postoperative change in liver stiffness was an independent risk factor for disease-free survival and recurrence outcomes, both univariate and multivariate analyses were conducted (Table 4 and Table 5). Multivariate analyses showed that α-fetoprotein (AFP) level (HR 1.210, P = 0.029) and change in liver stiffness (HR 1.521, P = 0.040) were independent prognostic variables for HCC disease-free survival (Table 4) while serum HBV DNA viral load (HR 1.011, P = 0.040), AFP level (HR 1.929, P = 0.035) and change in liver stiffness (HR 1.052, P = 0.032) were independent prognostic variables for HCC recurrence (Table 5).

Table 4.

Univariate and multivariate analyses of prognostic variables for disease-free survival in patients with hepatitis B virus-related hepatocellular carcinoma.

Variable	Univariate analysis			Multivariate analysis
	HR	95% CI	Statistical significance	HR	95% CI	Statistical significance
Age	1.001	0.991, 1.012	NS
Sex	0.972	0.670, 1.594	NS
CTP score	0.662	0.221, 0.976	P = 0.037
HBV DNA	0.495	0.330, 1.109	NS
HBeAg status	1.230	1.065, 1.692	P = 0.021
AFP level	1.291	1.061, 2.182	P = 0.015	1.210	1.115, 1.692	P = 0.029
Platelet count	1.211	1.067, 1.877	P = 0.027
Total bilirubin	1.159	0.698, 1.912	NS
LS changes	1.530	1.051, 2.248	P = 0.035	1.521	1.074, 2.081	P = 0.040

AFP, α-fetoprotein; CTP, Child–Turcotte–Pugh score; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; LS, liver stiffness; HR, hazard ratio; CI, confidence interval.

NS, no statistically significant correlation (P > 0.05; Cox proportional hazards).

Table 5.

Univariate and multivariate analyses of prognostic variables for recurrence in patients with hepatitis B virus-related hepatocellular carcinoma.

Variable	Univariate analysis			Multivariate analysis
	HR	95% CI	Statistical significance	HR	95% CI	Statistical significance
Age	0.172	0.129, 1.284	NS
Sex	1.201	0.882, 1.729	NS
CTP score	1.209	1.019, 1.812	P = 0.025
HBV DNA	1.029	1.008, 1.892	P = 0.037	1.011	1.002, 1.928	P = 0.040
HBeAg status	0.662	0.197, 1.236	NS
AFP level	1.861	1.081, 2.162	P = 0.029	1.929	1.028, 2.788	P = 0.035
Platelet count	1.191	0.772, 1.962	NS
Total bilirubin	1.099	0.294, 1.621	NS
LS changes	1.012	1.008, 1.425	P = 0.012	1.052	1.005, 1.752	P = 0.032

AFP, α-fetoprotein; CTP, Child–Turcotte–Pugh score; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; LS, liver stiffness; HR, hazard ratio; CI, confidence interval.

NS, no statistically significant correlation (P > 0.05; Cox proportional hazards).

Discussion

The present study demonstrated that increased postoperative liver stiffness may be a potential risk factor for poor HCC outcomes in patients with HBV-related HCC. The study showed that patients with decreased postoperative liver stiffness values had better survival outcomes than patients with elevated liver stiffness values. In addition, Child–Turcotte–Pugh score and liver stiffness changes were found to be independent variables associated with overall survival. AFP level and liver stiffness changes were found to be independent variables associated with HCC disease-free survival while serum HBV DNA viral load, AFP level and liver stiffness changes were independent variables associated with HCC recurrence. As such, these factors may have prognostic value in patients with HBV-related HCC.

Although liver histology is useful regarding diagnosis of hepatic diseases, such as cirrhosis, the invasive nature of the procedure limits its wider use in clinical practice.[29,30] FibroScan™ (Echosens) has been routinely applied for liver diseases detection worldwide, and allows clinicians to quantitatively evaluate the status of the liver.[31] As a non-invasive procedure, FibroScan™ is also suitable for patients following liver resections. Previous advances have indicated that elevated liver stiffness is associated with HCC, but the mechanisms for liver regeneration after surgery are complicated in patients with underlying HBV infection.[32] Postoperative screening with transient elastography provides a reproducible method of quantification to assess the alteration of liver regeneration, as well as potential impact on the prognosis of patients with HCC.[33] The present study found a correlation between dynamic changes of liver stiffness and prognosis in patients with HBV-related HCC, which may be utilized as a dynamic approach to monitor HCC or evaluate the recovery of patients.

Several markers have been identified as effective in evaluating prognosis in patients with HCC.[34-36] For example, AFP has been widely reported as a prognostic marker in patients with HCC,[37-39] and is a serum HCC marker that is used to diagnosis and monitor HCC progression. However, AFP levels are not only elevated in HCC. Other diseases, such as those involving tumours of the female reproductive system, may also manifest with elevated AFP levels.[40,41] Although high postoperative AFP levels are associated with poorer HCC prognosis, there is a certain proportion of patients with HCC who are AFP negative.[42] For those AFP-negative patients with HCC, AFP is not suitable as a prognostic marker. According to the present study, patients with HCC with abnormal or normal AFP levels may instead be effectively stratified using liver stiffness measured by transient elastography.

The present study results may be limited by several factors. The relatively small sample size may limit the generalizability of the results, and the retrospective study design may have biased the results. In the present study, the antiviral treatment was determined by clinical experts at Yantai Infectious Diseases Hospital, liver stiffness was measured at Jining No.1 People’s Hospital, and HCC diagnosis and surgery were performed at Heze Municipal Hospital. In addition, patients included in the present study received various different treatments following surgery, for example, some received no further treatment following radical resection, while some were treated with sorafenib, and these factors were not included in the present analyses. All of these additional variables may have affected the present results, and whether or not postoperative treatment may further improve the prognosis of patients with HCC remains unclear. Data were collected and analysed at a single centre. Larger multicentre prospective studies are required to validate the present results and further investigate the value of liver stiffness as a prognostic marker in patients with HCC.

In conclusion, the present study demonstrated an association between dynamic change in liver stiffness and prognosis in patients with HCC. Liver stiffness measurement with transient elastography may allow the evaluation of outcomes in patients with HCC, and thus, may be used to guide routine follow-up for these patients. In particular, patients with HCC with increased liver stiffness following surgery should undergo more frequent follow-up examinations, even if the most recent exam following liver resection indicates no cause for concern.