Professional Documents
Culture Documents
LIVER
Hepatocellular Carcinoma Progression of Indeterminate
Nodules in Hepatitis B Virus-Related Cirrhotic Liver
Hyo Jung Cho, MD, PhD1, 4, Bohyun Kim, MD, PhD2, 4, Jung-Dong Lee, MS3, Dae Ryong Kang, PhD3, Jai Keun Kim, MD, PhD2,
Jei Hee Lee, MD, PhD2, Sung Jae Shin, MD, PhD1, Kee Myung Lee, MD, PhD1, Byung Moo Yoo, MD, PhD1, Kwang Jae Lee, MD, PhD1,
Soon Sun Kim, MD, PhD1, Jae Youn Cheong, MD, PhD1 and Sung Won Cho, MD, PhD1
OBJECTIVES: This study was performed to evaluate long-term outcome of indeterminate nodules detected on
cirrhotic liver and to develop risk prediction model for hepatocellular carcinoma (HCC) progression
of indeterminate nodules on hepatitis B virus (HBV)-related cirrhotic liver.
RESULTS: A total of 494 indeterminate nodules were included. Independent risk factors of HCC progression
were old age, arterial enhancement, large nodule size, low serum albumin level, high serum
α -fetoprotein (AFP) level, and prior HCC history in all included subjects. In subjects with chronic
hepatitis B, old age (year; hazard ratio (HR)=1.06; P<0.001), arterial enhancement (HR=2.62;
P=0.005), large nodule size (>1 cm; HR=7.34; P<0.001), low serum albumin level (≤3.5 g/dl;
HR=3.57; P=0.001), high serum AFP level (≥100 ng/ml; HR=6.04; P=0.006), prior HCC history
(HR=4.24; P=0.001), and baseline hepatitis B e antigen positivity (HR=2.31; P=0.007) were
associated with HCC progression. We developed a simple risk prediction model using these risk
factors and identified patients at low, intermediate, and high risk for HCC; 5-year cumulative
incidences were 1%, 14.5%, and 63.1%, respectively. The developed risk score model showed good
performance with area under the curve at 0.886 at 3 years, and 0.920 at 5 years in leave-one-out
cross-validation.
CONCLUSIONS: We developed a useful and accurate risk score model for predicting HCC progression of indeterminate
nodules detected on HBV-related cirrhotic liver.
Am J Gastroenterol advance online publication, 25 October 2016; doi:10.1038/ajg.2016.480
1
Department of Gastroenterology, Ajou University School of Medicine, Suwon, South Korea; 2Department of Radiology, Ajou University School of Medicine,
Suwon, South Korea; 3Office of Biostatistics, Ajou University School of Medicine, Suwon, South Korea; 4The first two authors contributed equally to this work.
Correspondence: Sung Won Cho, MD, PhD, Department of Gastroenterology, Ajou University School of Medicine, Worldcup-ro 164, Yeongtong-Gu, Suwon
443-380, South Korea. E-mail: sung_woncho@hotmail.com
Received 14 April 2016; accepted 14 September 2016
uncertain malignant potential, should be differentially diagnosed 4,363 patients with liver cirrhosis who underwent dynamic liver
CT between January 2005 and December 2013
with regenerative nodule, dysplastic nodule, early HCC, and “non-
nodule–nodule-like lesions” such as nontumorous arterioportal 2,130 patients were excluded
(AP) shunt (9,10). Although the American Association for the
LIVER
helpful; however, the cost is a considerable problem, and MR can- 194 nodules were excluded
not offer correct answers all the time. Moreover, regenerative or • 46 nodules diagnosed as HCC within 3 months
• 72 nodules diagnosed as HCC between 6 to 12 months
dysplastic nodules could progress to HCC along a well-described • 49 nodules were not followed up more than 12 months
without any conclusive diagnosis.
stepwise hepatocarcinogenic process, even if initially it is not • 27 nodules were excluded due to development of
metachronous HCC
diagnosed as HCC on biopsy or MR (15). Therefore, reliable and
easy-to-use clinical model for predicting long-term outcome of the
494 indeterminate nodules from 474 patients were included
indeterminate nodules would be helpful in clinical practice. • Subject without HCC, N=410 (393 patients)
• Subject with HCC, N=84 (81 patients)
This study was aimed to evaluate long-term outcome of indeter-
minate nodules and risk factors associated with HCC progression Figure 1. Flowchart for subject inclusion. CT, computed tomography; HCC,
of the indeterminate nodules in all cirrhotic patients, and also tried hepatocellular carcinoma.
to develop HCC risk prediction model especially for the indeter-
minate nodules detected on hepatitis B virus (HBV)-related cir-
rhotic liver.
for nodules >1 cm, (ii) growing of diameter >1 cm and changing
characteristics showing definite HCC findings as described above
METHODS for nodules < 1 cm, and/or (iii) pathologic confirmation of HCC.
Subject selection and definition of terminology Indeterminate nodules without HCC progression during follow-
This retrospective study was conducted at Ajou University Hos- up period were categorized to subjects without HCC.
pital, Suwon, South Korea, between January 2005 and December Figure 1 demonstrates a flowchart for subject inclusion of the
2013. Indeterminate nodule was defined as nodular lesions up present study. Among 4,363 consecutive patients with liver cirrho-
to 2 cm with undetermined malignant potential detected on cir- sis who underwent dynamic liver CT during study duration, 2,130
rhotic liver by dynamic liver computed tomography (CT), which patients were excluded based on exclusion criteria. Patients were
was performed for both primary HCC surveillance and diagnos- excluded if there was another primary tumor history and if there
tic CT to characterize a liver lesion identified on ultrasonography, was a definite HCC finding on the CT imaging. Patients with prior
during HCC surveillance with the following radiologic features: HCC history were excluded basically; however, patients with prior
(i) arterial enhancing nodular lesions without any other radio- HCC history who underwent curative therapy and were followed
logic features suggesting HCC, such as delayed phase washout or up for >2 years without recurrence were included if they fulfilled
capsular enhancement (selective arterial enhancing nodules), (ii) inclusion/exclusion criteria.
nonenhancing nodules on arterial phase with venous or delayed Among CT images of 2,233 patients, 688 indeterminate nodules
phase washout (selective portal or delayed washout nodules), and from 641 patients were detected. Subjects were excluded (i) if they
(iii) AP shunt lesion requiring follow-up imaging for differential were diagnosed as HCC within 3 months by immediate evalua-
diagnosis with small HCC or dysplastic nodule because the initial tion such as MR and/or biopsy, (ii) if they were diagnosed as HCC
imaging findings are not entirely typical of AP shunt. A typical AP between 6 and 12 months from their first detection, (iii) if they
shunt was defined as a wedge-shaped, peripheral locating lesion were not followed up for >12 months without any conclusive diag-
with transient parenchymal enhancement during the hepatic nosis, and (iv) if metachronous HCC was developed in a different
arterial phase (16). Nodules were not included if they had defi- location from those of the indeterminate nodules during follow-up
nite radiologic features of HCC (17), or definite benign features, period. Among 688 subjects, 145 nodules had undergone immedi-
such as typical hemangioma, definite cystic lesion, or typical AP ate evaluation such as MR (N=134) or MR and biopsy (N=11), and
shunt. Diagnosis of cirrhosis was made if one or more of following eventually 46 nodules were diagnosed as HCC with sequential MR
four criteria were met in the absence of acute liver failure: platelet imaging (N=37) and/or biopsy (N=9) within 3 months. Finally,
count <100,000/μ l; albumin level <3.5 g/dl; international normal- 494 indeterminate nodules from 474 patients were included in the
ized ratio >1.3; and surface nodularity on imaging study (18). study.
HCC diagnosis of the indeterminate nodules during follow- Included nodules were followed up using 4-phase CT scans con-
up was defined according to AASLD guidelines as follows (3); (i) taining pre-contrast, arterial, portal venous, and delayed phases
changing radiologic characteristics showing definite HCC find- using one of the four CT scanners in our institution (SOMATOM
ings such as arterial enhancement and delayed phase washout Sensation 16, Siemens AG, Forchheim, Germany; Brilliance 16,
Table 1. Baseline characteristics of the included subjects and comparison between subjects with and without hepatocellular carcinoma
Variables All included subjects Subjects associated Comparison between subjects with and without HCC
(N=494) with CHB (N=373)
LIVER
Subjects without HCC Subjects with HCC P
(N=410 ) (N=84)
Philips, Best, The Netherlands; Brilliance 64, Philips; SOMATOM tive portal or delayed washout nodules, and atypical AP shunts
Definition Flash, Siemens) every 3 or 6 months. Interpretation requiring follow-up imaging.
of imaging analysis was performed by three expert abdominal Data about the sex, age, whether HCC developed or not, cause of
radiologists with >5 years of experience. To test for interreader underlying cirrhosis, Child–Pugh class, prior HCC history, antivi-
agreement in categorizing indeterminate nodules, two reviewers ral therapy history, and follow-up periods were collected from the
(BLINDED) independently reviewed the indeterminate nodules medical records. Radiologic findings on nodule size and pattern
and classified them as selective arterial enhancing nodules, selec- of arterial enhancement were also reviewed. Baseline laboratory
1.0
Cumulative incidence rate Austria, http://www.r-project. org). Kaplan–Meier survival analy-
2 year 12.3% sis was performed to estimate HCC progression over time. Time
Cumulative incidence rate
0.6
tors associated with HCC progression among nodules, univariate
and multivariate analyses were performed using Cox regression
0.4 analysis. Independent risk factors obtained from multivariate
Cox analysis were entered into the HCC prediction risk score
0.2 model. The risk score was calculated by the weighted sum of the
selected variables, and the weights were defined as the quotient
0.0 (rounded to nearest integer) of estimated corresponding β -coef-
12 24 36 48 60 72
ficient divided by the smallest β derived from multivariate Cox
Time to HCC progression (months) regression analysis in a stepwise method. The performance of the
No. of
subject selected model was evaluated by leave-one-out cross-validation.
at risk 494 400 260 176 126 79 Receiver operating characteristic curves and area under the curve
Figure 2. Cumulative incidence of hepatocellular carcinoma (HCC) pro- (AUC) (95% confidence interval (CI)) at 36 and 60 months were
gression of indeterminate nodules detected on cirrhotic liver. computed using leave-one-out cross-validation. The calculated
risk score by the developed model was categorized into three
groups: low-, intermediate-, and high-risk groups. The cutoff val-
parameters, including complete blood count, total bilirubin, albu- ues for dividing each group were determined by considering the
min, alanine transaminase, international normalized ratio, hepati- best discriminatory ability and monotonicity.
tis B envelope antigen (HBeAg)/hepatitis B e antibody (anti-HBe), Unweighted birater κ-coefficient and its 95% CI were used to
HBV DNA titer, and α -fetoprotein (AFP), were collected. assess the degree of agreement between two reviewers (19). Strength
of κ agreement was defined as follows; < 0.00, poor; 0.00–0.20, slight;
Baseline characteristics and cumulative incidence of HCC 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and
progression 0.81–1.00, almost perfect (20). Statistical analysis for κ was performed
Table 1 shows baseline characteristics of study subjects and com- with MedCalc version 16.4.3 (MedCalc, Mariakerke, Belgium).
parison of baseline characteristics of study subjects according to the
final diagnosis, non-HCC vs. HCC. Among 494 nodules from 474
patients selected based on inclusion/exclusion criteria, 84 nodules RESULTS
(17.0%) progressed to HCC, whereas 410 nodules (83%) remained Interreader agreement for categorizing indeterminate nodules
indeterminate during follow-up. The median follow-up period of Overall agreement for interpreting indeterminate nodules between
enrolled subjects was 36 months (range, 12–124 months). two readers was substantial, with unweighed κ -coefficient value
In the subgroup analysis of nodules with HBV-related liver cir- of 0.77 (95% CI=0.72–0.82).
rhosis, 62 of 373 (16.6%) nodules developed to HCC. The HCC
progression group demonstrated significantly higher proportion Risk factors associated with HCC progression of indeterminate
of HBeAg positivity and significantly higher HBV DNA titer com- nodules in all included nodules
pared with the no HCC progression group. Patients with antiviral The Kaplan–Meier analysis and Cox regression analysis were per-
treatment were defined as patients who were already undergoing formed to identify risk factors associated with indeterminate nod-
antiviral therapy before detection of the nodular lesion or started ule to HCC progression. Figure 3 shows Kaplan–Meier curves
antiviral therapy within 3 months from the inclusion of the study. comparing groups divided by age, liver cirrhosis etiology, arterial
The patients without antiviral treatment group were further enhancement pattern, and nodule size. Multivariate Cox regres-
divided according to baseline HBV DNA viral load. High HBV sion analysis revealed that baseline parameters associated with
DNA viral load was defined as >20,000 IU/ml in HBeAg-positive HCC progression in indeterminate nodules were old age (year;
patients and >2,000 IU/ml in HBeAg-negative patients. The χ 2 test hazard ratio (HR)=1.04; 95% CI=1.01–1.07; P=0.002), arterial
revealed that more patients without antiviral treatment despite enhancement (yes; HR=2.72; 95% CI=1.58–4.68; P<0.001), large
high HBV titer were distributed to the HCC progression group nodule size (>1 cm; HR=7.18; 95% CI=4.08–12.67; P<0.001), low
significantly (33.9%) than to no HCC progression group (11.3%). serum albumin level (≤3.5 g/dl; HR=1.72; 95% CI=0.98–3.02;
The cumulative incidence of HCC progression is shown in P=0.060), high serum AFP level (≥100 ng/ml; HR=4.48; 95%
Figure 2. The cumulative incidence of HCC progression was CI=1.79–11.19; P=0.001), and presence of prior HCC history
12.3%, 17.6%, and 20.2% at 2, 3, and 5 years, respectively. (yes; HR=3.25; 95% CI=1.44–7.30; P=0.004). (Table 2).
Subgroup analysis was performed to identify risk factors asso-
Statistical analysis ciated with early HCC diagnosis within 12 months by using
Statistical analysis was performed using R software packages logistic regression analysis. A total of 72 indeterminate nodules
(R version 3.1.2, R Project for Statistical Computing, Vienna, diagnosed to HCC between 6 and 12 months from their first
a 1.0 b 1.0
Age ≤ 55 HCV HBV
of HCC progression
of HCC progression
0.8 0.8
LIVER
0.6 0.6 P=0.018
P<0.001
0.4 0.4
0.2 0.2
0.0 0.0
12 24 36 48 60 72 12 24 36 48 60 72
c 1.0 d
Arterial enhancement(–) 1.0
Size ≤ 1 cm
Cumulative incidence rate
Arterial enhancement(+)
of HCC progression
0.8
0.6 P<0.001
P<0.001 0.6
0.4
0.4
0.2 0.2
0.0 0.0
12 24 36 48 60 72 12 24 36 48 60 72
Time to HCC progression (months) Time to HCC progression (months)
No. of subject No. of subject
at risk at risk
Arterial enhancement(–) 291 240 161 108 77 49 Size≤1 cm 364 315 214 147 109 72
Arterial enhancement(+) 91 52 28 20 16 11 Size>1 cm 130 85 46 29 17 7
R/O AP shunt 112 108 71 48 33 19
Figure 3. Comparison of hepatocellular carcinoma progression by Kaplan–Meier curves in included subjects according to groups divided by (a) age (group
1, <55 and group 2, ≥55 years), (b) etiology of liver cirrhosis (group 1, HBV; group 2, HCV; group 3, alcoholic hepatitis; group 4, other causes), (c) arterial
enhancement pattern (group 1, arterial enhancement; group 2, no arterial enhancement; group 3, arterioportal (AP) shunt), and (d) nodule size (group 1,
≤1 cm and group 2, >1 cm). HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus.
detection were compared with subjects without HCC (n=410). DNA level. HBeAg positivity, high HBV DNA titer, and the no
Finally, old age, arterial enhancement, large nodule size, low antiviral treatment group despite high HBV DNA titer were sig-
platelet count, and high serum AFP level were identified as inde- nificantly associated with increasing risk of HCC progression.
pendent risk factors for predicting early HCC diagnosis within Cox regression analyses (Table 4) were performed to identify
12 months (Table 3), and those almost corresponded to the risk independent risk factors for the development of HCC. After
factors of HCC progression of indeterminate nodules demon- adjusting compounding factors, old age (year; HR=1.06; 95%
strated at Table 2. CI=1.04–1.09; P<0.001), arterial enhancement (yes; HR=2.62;
95% CI=1.33–5.18; P=0.005), large nodule size (>1 cm; HR=7.34;
Clinical and radiologic variables associated with HCC progression 95% CI=3.70–13.57; P<0.001), low serum albumin level (≤3.5 g/
of indeterminate nodules detected on HBV-related cirrhosis dl; HR=3.57; 95% CI=1.72–7.42; P=0.001), high serum AFP level
Subgroup analysis was performed in 373 subjects from 356 (≥100 ng/ml; HR=6.04; 95% CI=1.69–21.56; P=0.006), prior HCC
patients with HBV-related cirrhosis to identify risk factors asso- history (yes; HR=4.24; 95% CI=1.76–10.19; P=0.001), and HBeAg
ciated with indeterminate nodule to HCC progression in HBV- positivity (HR=2.31; 95% CI=1.26–4.24; P=0.007) were identi-
related cirrhosis. Figure 4 shows comparison of Kaplan–Meier fied as independent risk factors for predicting HCC progression
plot between groups divided by HBeAg positivity, baseline HBV of indeterminate nodule detected on HBV-related liver cirrhosis.
DNA titer, and groups according to antiviral therapy and/or HBV These risk factors were used to develop risk prediction model.
Table 2. Univariate and multivariate analysis to evaluate the risk factors associated with indeterminate nodule to hepatocellular carcinoma
progression (N=494)
Table 3. Univariate and multivariate analysis to identify risk factors associated with early HCC diagnosis of indeterminate nodules by
comparing subjects diagnosed to hepatocellular carcinoma between 6 to 12 months from first detection (N=72) and subjects without HCC
diagnosis during follow up (N=410)
Risk prediction model for indeterminate nodule to HCC Risk score = 1× age (years) + 19 × enhancement pattern (arterial
progression in B-viral cirrhosis non-enhancement or AP shunt = 0; arterial enhancement = 1)
The risk score was calculated by weighted sum using β -coefficient + 42 × size (≤ 1 cm = 0; > 1 cm = 1) + 16 × serum albumin
derived from multivariate Cox regression analysis. The formula of (> 3.5 g/dL = 0; ≤ 3.5 g/dL = 1) + 31× serum AFP (< 100 ng/
the risk score was as follows. The β -coefficient of the correspond- mL = 0; ≥ 100 ng/mL = 1) + 32× prior HCC history (no = 0;
ing risk factors and their risk score is listed in Table 5. yes = 1) + 18 × HBeAg (negative = 0; positive = 1)
a b
1.0 1.0
Low HBV DNA titer HBeAg (–)
High HBV DNA titer HBeAg (+)
of HCC progression
of HCC progression
LIVER
0.6 0.6
0.2 0.2
0.0 0.0
12 24 36 48 60 72 12 24 36 48 60 72
No. of subject Time to HCC progression (months) No. of subject Time to HCC progression (months)
at risk at risk
Low HBV DNA titer 172 144 92 58 41 26 HBeAg (–) 247 209 139 90 70 43
High HBV DNA titer 182 145 102 75 53 32 HBeAg (+) 117 87 59 47 29 20
c
1.0
Antiviral therapy
No antiviral therapy, low HBV DNA titer
Cumulative incidence rate
0.6
P<0.001
0.4
0.2
0.0
12 24 36 48 60 72
Figure 4. Comparison of hepatocellular carcinoma progression by Kaplan–Meier curves in subjects with HBV-related cirrhosis according to groups divided
by (a) baseline HBV DNA titer (group 1, high HBV DNA titer and group 2, low HBV DNA titer), (b) HBeAg positivity (group 1, negative and group 2,
positive), and (c) whether to perform antiviral therapy or not according to HBV DNA titer level (group 1, antiviral treatment group; group 2, no antiviral
treatment in low HBV DNA titer; group 3, no antiviral treatment in high HBV DNA titer). HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; HCC,
hepatocellular carcinoma; High HBV DNA viral load, >20,000 IU/ml in HBeAg-positive patients and >2,000 IU/ml in HBeAg-negative patients.
We plotted time-dependent receiver operating characteris- the low-risk group, 9.2%, 14.5%, and 14.5% at 24, 36, 60 months,
tic curves and calculated AUCs in the development cohort and respectively, in the intermediate-risk group, whereas progression
leave-one-out cross-validation (Figure 5). The developed model rate was 41.6%, 53.3%, and 63.1% at 24, 36, and 60 months, respec-
showed relatively good performance with AUC of 0.879 (95% tively, in the high-risk group.
CI=0.827–0.931) and 0.922 (0.878–0.966) at 3 and 5 years, respec-
tively, in the derivation cohort, and AUC of 0.886 (0.836–0.935)
and 0.920 (0.875–0.965) at 3 and 5 years, respectively, in leave- DISCUSSION
one-out cross-validation. This study demonstrated the outcome of indeterminate nodules
Our cohort was categorized into three groups by cutoff point at discovered on cirrhotic liver and risk factors associated with HCC
60 and 105 of calculated risk score: low-risk group (risk score <60), progression of the nodules based on long-term follow-up data. In
intermediate-risk group (60<risk score<105), and high-risk group addition, we developed risk score model for predicting nodule
(risk score >105). These three groups were compared by Kaplan– to HCC progression risk in patients with HBV-related cirrhosis.
Meier analysis using log rank test and predicted risk was well Although several studies developed risk score model to identify
calibrated with observed HCC risk (Figure 6). HCC progression high-risk patients of HCC development in Asian HBV carriers
rate was 0%, 1%, and 1% at 24, 36, and 60 months, respectively, in (21–24), only few studies investigated clinical risk estimating
Table 4. Univariate and multivariate analysis to evaluate the risk factors associated with indeterminate nodule to hepatocellular carcinoma
progression in patients with hepatitis B virus-related cirrhosis (N=373)
a b
ROC curve at months = 36 ROC curve at months = 36
1.0 1.0
P=0.9847 P=0.9955
LIVER
0.8 0.8
Sensitivity
Sensitivity
0.6 0.6
0.4 0.4
0.2 0.2
Total AUC = 0.879 (se = 0.266) Total AUC = 0.922 (se = 0.266)
0.0
Leave-one-out AUC = 0.886 (se = 0.252) 0.0 Leave-one-out AUC = 0.920 (se = 0.23)
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
1 - Specificity 1 - Specificity
AUC 87.91 (82.70–93.12) 88.56 (83.62–93.50) AUC 92.21 (87.78–96.64) 92.00 (87.49–96.51)
Sensitivity 80.06 (68.36–91.76) 82.34 (71.19–93.49) Sensitivity 82.92 (72.57–93.27) 84.65 (74.73–94.57)
Specificity 78.63 (71.59–85.67) 79.58 (72.94–86.22) Specificity 84.85 (76.19–93.51) 86.08 (78.44–93.72)
Figure 5. Time-dependent receiving operating characteristic (ROC) curves and calculated area under the curves (AUC) in development cohort and leave-
one-out cross-validation. (a) Time-dependent receiving operating curve and calculated area under the curve at 36 months. (b) Time-dependent receiving
operating curve and calculated area under the curve at 60 months.
1.0 High risk pared with 24 months. Therefore, the nodular lesions categorized
Cumulative incidence rate
Intermediate risk
in low-risk groups could be followed up by regular surveillance
of HCC progression
Although the risk score model is very useful and reproducible, it 4. European Association For The Study Of The Liver. European Organisation For
has a limitation of generalizability. Because this model was derived Research And Treatment Of Cancer. EASL–EORTC clinical practice guide-
lines: management of hepatocellular carcinoma. J Hepatol 2012;56:908–43.
from the subjects detected on only HBV-related cirrhotic liver, it 5. Omata M, Lesmana LA, Tateishi R et al. Asian Pacific Association for the
would not be generally applicable for subjects detected on noncir-
LIVER
31. Velázquez RF, Rodriguez M, Navascues CA et al. Prospective analysis of 34. Chang TT, Liaw YF, Wu SS et al. Long‐term entecavir therapy results in the
risk factors for hepatocellular carcinoma in patients with liver cirrhosis. reversal of fibrosis/cirrhosis and continued histological improvement in
Hepatology 2003;37:520–7. patients with chronic hepatitis B. Hepatology 2010;52:886–93.
32. Yuen M-F, Yuan H-J, Wong DK et al. Prognostic determinants for chronic 35. Colombo M, Iavarone M. Role of antiviral treatment for HCC prevention.
hepatitis B in Asians: therapeutic implications. Gut 2005;54:1610–4. Best Pract Res Clin Gastroenterol 2014;28:771–81.
LIVER
33. Marcellin P, Gane E, Buti M et al. Regression of cirrhosis during treatment 36. Fernández-Rodríguez CM, Gutiérrez-García ML. Prevention of hepatocel-
with tenofovir disoproxil fumarate for chronic hepatitis B: a 5-year open- lular carcinoma in patients with chronic hepatitis B. World J Gastrointest
label follow-up study. Lancet 2013;381:468–75. Pharmacol Ther 2014;5:175–82.
AQ3