ORIGINAL CONTRIBUTIONS 1

see related editorial on page x

Development of Risk Prediction Model for

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.

METHODS: Indeterminate nodules up to 2 cm with uncertain malignant potential detected on computed

tomography of cirrhotic liver during HCC surveillance were analyzed retrospectively. HCC risk
prediction model of indeterminate nodules in HBV-related cirrhotic liver was deduced based on
result of Cox regression analysis.

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

INTRODUCTION During HCC imaging surveillance, cirrhosis-associated nodules

Hepatocellular carcinoma (HCC) is the fifth most common
malignancy and the third most common cause of cancer mortal-
ity worldwide (1,2). Patients with liver cirrhosis are considered as
high-risk group of developing HCC, and most guidelines recom-
mend HCC surveillance in patients with liver cirrhosis to detect
early HCC amenable to curative treatment (3–5).
caused by localized proliferation of hepatocytes and their support-
ing stroma due to continuous liver injury, are often detected (6–8).
In clinical practice, imaging interpretation and management deci-
sion for these nodules are difficult because accurate diagnosis and
estimation of their malignant potential is difficult, particularly in
nodules < 2 cm. Indeterminate nodules, which are up to 2 cm with

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

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 2 Cho et al.
uncertain malignant potential, should be differentially diagnosed
with regenerative nodule, dysplastic nodule, early HCC, and “non-
nodule–nodule-like lesions” such as nontumorous arterioportal
(AP) shunt (9,10). Although the American Association for the
LIVER
Study of Liver Diseases (AASLD) guideline recommends liver
biopsy for the lesions (11), it is also hard to perform when con-
sidering pitfall of image-guided biopsy such as relatively low diag-
nostic yield (12), false-negative result (13), and procedure-related
complications (14). Magnetic resonance (MR) imaging might be
helpful; however, the cost is a considerable problem, and MR can-
not offer correct answers all the time. Moreover, regenerative or
dysplastic nodules could progress to HCC along a well-described
stepwise hepatocarcinogenic process, even if initially it is not
diagnosed as HCC on biopsy or MR (15). Therefore, reliable and
easy-to-use clinical model for predicting long-term outcome of the
indeterminate nodules would be helpful in clinical practice.
This study was aimed to evaluate long-term outcome of indeter-
minate nodules and risk factors associated with HCC progression
of the indeterminate nodules in all cirrhotic patients, and also tried
to develop HCC risk prediction model especially for the indeter-
minate nodules detected on hepatitis B virus (HBV)-related cir-
rhotic liver.
METHODS
Subject selection and definition of terminology
This retrospective study was conducted at Ajou University Hos-
pital, Suwon, South Korea, between January 2005 and December
2013. Indeterminate nodule was defined as nodular lesions up
to 2 cm with undetermined malignant potential detected on cir-
rhotic liver by dynamic liver computed tomography (CT), which
was performed for both primary HCC surveillance and diagnos-
tic CT to characterize a liver lesion identified on ultrasonography,
during HCC surveillance with the following radiologic features:
(i) arterial enhancing nodular lesions without any other radio-
logic features suggesting HCC, such as delayed phase washout or
capsular enhancement (selective arterial enhancing nodules), (ii)
nonenhancing nodules on arterial phase with venous or delayed
phase washout (selective portal or delayed washout nodules), and
(iii) AP shunt lesion requiring follow-up imaging for differential
diagnosis with small HCC or dysplastic nodule because the initial
imaging findings are not entirely typical of AP shunt. A typical AP
shunt was defined as a wedge-shaped, peripheral locating lesion
with transient parenchymal enhancement during the hepatic
arterial phase (16). Nodules were not included if they had defi-
nite radiologic features of HCC (17), or definite benign features,
such as typical hemangioma, definite cystic lesion, or typical AP
shunt. Diagnosis of cirrhosis was made if one or more of following
four criteria were met in the absence of acute liver failure: platelet
count <100,000/μ l; albumin level <3.5 g/dl; international normal-
ized ratio >1.3; and surface nodularity on imaging study (18).
HCC diagnosis of the indeterminate nodules during follow-
up was defined according to AASLD guidelines as follows (3); (i)
changing radiologic characteristics showing definite HCC find-
ings such as arterial enhancement and delayed phase washout
The American Journal of GASTROENTEROLOGY
4,363 patients with liver cirrhosis who underwent dynamic liver
CT between January 2005 and December 2013
2,130 patients were excluded
• 273 patients with another primary tumor in their history
• 794 patients with definite HCC on their CT imaging
• 578 patients with previous HCC history within 2 years
• 285 patients with previous HCC history more than 2
years ago who underwent non-curative treatment
Among CT images of 2,233 patients, 688 indeterminate nodules
from 641 patients were detected
194 nodules were excluded
• 46 nodules diagnosed as HCC within 3 months
• 72 nodules diagnosed as HCC between 6 to 12 months
• 49 nodules were not followed up more than 12 months
without any conclusive diagnosis.
• 27 nodules were excluded due to development of
metachronous HCC
494 indeterminate nodules from 474 patients were included
• Subject without HCC, N=410 (393 patients)
• Subject with HCC, N=84 (81 patients)
Figure 1. Flowchart for subject inclusion. CT, computed tomography; HCC,
hepatocellular carcinoma.
for nodules >1 cm, (ii) growing of diameter >1 cm and changing
characteristics showing definite HCC findings as described above
for nodules < 1 cm, and/or (iii) pathologic confirmation of HCC.
Indeterminate nodules without HCC progression during follow-
up period were categorized to subjects without HCC.
Figure 1 demonstrates a flowchart for subject inclusion of the
present study. Among 4,363 consecutive patients with liver cirrho-
sis who underwent dynamic liver CT during study duration, 2,130
patients were excluded based on exclusion criteria. Patients were
excluded if there was another primary tumor history and if there
was a definite HCC finding on the CT imaging. Patients with prior
HCC history were excluded basically; however, patients with prior
HCC history who underwent curative therapy and were followed
up for >2 years without recurrence were included if they fulfilled
inclusion/exclusion criteria.
Among CT images of 2,233 patients, 688 indeterminate nodules
from 641 patients were detected. Subjects were excluded (i) if they
were diagnosed as HCC within 3 months by immediate evalua-
tion such as MR and/or biopsy, (ii) if they were diagnosed as HCC
between 6 and 12 months from their first detection, (iii) if they
were not followed up for >12 months without any conclusive diag-
nosis, and (iv) if metachronous HCC was developed in a different
location from those of the indeterminate nodules during follow-up
period. Among 688 subjects, 145 nodules had undergone immedi-
ate evaluation such as MR (N=134) or MR and biopsy (N=11), and
eventually 46 nodules were diagnosed as HCC with sequential MR
imaging (N=37) and/or biopsy (N=9) within 3 months. Finally,
494 indeterminate nodules from 474 patients were included in the
study.
Included nodules were followed up using 4-phase CT scans con-
taining pre-contrast, arterial, portal venous, and delayed phases
using one of the four CT scanners in our institution (SOMATOM
Sensation 16, Siemens AG, Forchheim, Germany; Brilliance 16,
www.nature.com/ajg
 Outcome of Indeterminate Nodules Detected on Cirrhotic Liver 3

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)

Size (mm), mean±s.d. 10.75±2.84 11.07±3.03 10.32±2.52 12.83±3.36 <0.001

Nodule size <1 cm, n (%) 364 (73.7) 274 (735) 342 (83.4) 22 (26.2) <0.001
Pattern of arterial enhancement
No arterial enhancement, n (%) 291 (58.9) 226 (60.6) 244 (59.5) 47 (56.0) <0.001
Arterial enhancement, n (%) 112 (22.7) 67 (18.0) 60 (14.6) 31 (36.9)
R/O arterioportal shunt, n (%) 91 (18.4) 80 (21.4) 106 (25.9) 6 (7.1)
Age (years), mean±s.d. (range) 52.78±11.18 (19–84) 51.51±11.19 (28–80) 51.58±11.02 (19–84) 58.67±10.14 (36–82) <0.001
Male, n (%) 361 (73.1) 299 (72.9) 66 (73.8) 0.868
Cause of background liver cirrhosis
Hepatitis B, n (%) 373 (75.5) 311 (75.9) 62 (73.8) 0.042
Hepatitis C, n (%) 19 (3.8) 11 (2.7) 8 (9.5)
Alcoholic, n (%) 77 (15.6) 67 (16.3) 10 (11.9)
Other cause, n (%) 25 (5.1) 21 (5.1) 4 (4.7)
History of prior HCC, n (%) 27 (5.5) 11 (2.9) 16 (3.9) 11 (13.1 ) 0.002
Child–Pugh class A/B/C, n (%) 443 (87.7)/58 (11.7)/3 339 (90.9)/33 (8.8)/1 366 (89.3)/41 (10.0)/3 67 (79.8)/17 (20.2)/0 0.023
(0.6) (0.3) (0.7)
Platelet (×109/l), mean±s.d. 120.66±60.19 116.29±54.05 131.92±59.62 105.25±51.07 0.001
Albumin (g/dl), mean±s.d. 3.86±0.62 3.91±0.60 4.13±0.54 3.82±0.58 <0.001
Bilirubin (μ mol/l), mean±s.d. 1.26±1.15 1.26±1.14 1.26±1.13 1.27±0.79 0.432
AFP (ng/ml), mean±s.d. 37.59±374.68 97.41±652.60 18.60±181.77 115.26±738.61 0.245
AFP >100 ng/ml, n (%) 13 (2.7) 11 (2.9) 4 (1.0) 9(11.1) <0.001
ALT (U/l), mean±s.d. 55.75±118.11 82.13±196.43 52.93±117.96 69.84±112.86 0.277
INR, mean±s.d. 1.19±0.25 1.18±0.24 1.18±0.26 1.21±0.19 0.599
Subjects associated with CHB, N (%) N=311 N=62
HBeAg positivity, n (%) 117 (32.1) 85 (28.1) 32 (51.6) 0.001
HBV DNA (log 10 IU/ml) 3.16±2.74 2.91±2.70 4.38±2.62 <0.001
(mean±s.d.)
Patients with antiviral treatment, 240 (64.3) 207 (66.6) 33 (53.2)
n (%)
Patients without antiviral treatment,
n (%)
High baseline HBV DNA, n (%) 56 (15.0) 35 (11.3) 21 (33.9) <0.001
Low baseline HBV DNA, n (%) 77 (20.6) 69 (22.1) 8 (12.9)
AFT, α -fetoprotein; ALT, alanine transaminase; CHB, chronic hepatitis B; HCC, hepatocellular carcinoma; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; High HBV
DNA viral load, >20,000 IU/ml in HBeAg-positive patients and >2,000 IU/ml in HBeAg-negative patients; INR, international normalized ratio.

Philips, Best, The Netherlands; Brilliance 64, Philips; SOMATOM
Definition Flash, Siemens) every 3 or 6 months. Interpretation
of imaging analysis was performed by three expert abdominal
radiologists with >5 years of experience. To test for interreader
agreement in categorizing indeterminate nodules, two reviewers
(BLINDED) independently reviewed the indeterminate nodules
and classified them as selective arterial enhancing nodules, selec-
tive portal or delayed washout nodules, and atypical AP shunts
requiring follow-up imaging.
Data about the sex, age, whether HCC developed or not, cause of
underlying cirrhosis, Child–Pugh class, prior HCC history, antivi-
ral therapy history, and follow-up periods were collected from the
medical records. Radiologic findings on nodule size and pattern
of arterial enhancement were also reviewed. Baseline laboratory

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 4 Cho et al
1.0
Cumulative incidence rate
2 year 12.3%
Cumulative incidence rate
0.8 3 year 17.6%
5 year 20.2%
of HCC progression
LIVER
0.6
0.4
0.2
0.0
12 24 36 48 60
Time to HCC progression (months)
No. of
subject
at risk 494 400 260 176 126
Figure 2. Cumulative incidence of hepatocellular carcinoma (HCC) pro-
gression of indeterminate nodules detected on cirrhotic liver.
parameters, including complete blood count, total bilirubin, albu-
min, alanine transaminase, international normalized ratio, hepati-
tis B envelope antigen (HBeAg)/hepatitis B e antibody (anti-HBe),
HBV DNA titer, and α -fetoprotein (AFP), were collected.
Baseline characteristics and cumulative incidence of HCC
progression
Table 1 shows baseline characteristics of study subjects and com-
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
(17.0%) progressed to HCC, whereas 410 nodules (83%) remained
indeterminate during follow-up. The median follow-up period of
enrolled subjects was 36 months (range, 12–124 months).
In the subgroup analysis of nodules with HBV-related liver cir-
rhosis, 62 of 373 (16.6%) nodules developed to HCC. The HCC
progression group demonstrated significantly higher proportion
of HBeAg positivity and significantly higher HBV DNA titer com-
pared with the no HCC progression group. Patients with antiviral
treatment were defined as patients who were already undergoing
antiviral therapy before detection of the nodular lesion or started
antiviral therapy within 3 months from the inclusion of the study.
The patients without antiviral treatment group were further
divided according to baseline HBV DNA viral load. High HBV
DNA viral load was defined as >20,000 IU/ml in HBeAg-positive
patients and >2,000 IU/ml in HBeAg-negative patients. The χ 2 test
revealed that more patients without antiviral treatment despite
high HBV titer were distributed to the HCC progression group
significantly (33.9%) than to no HCC progression group (11.3%).
The cumulative incidence of HCC progression is shown in
Figure 2. The cumulative incidence of HCC progression was
12.3%, 17.6%, and 20.2% at 2, 3, and 5 years, respectively.
Statistical analysis
Statistical analysis was performed using R software packages
(R version 3.1.2, R Project for Statistical Computing, Vienna,
The American Journal of GASTROENTEROLOGY
Austria, http://www.r-project. org). Kaplan–Meier survival analy-
sis was performed to estimate HCC progression over time. Time
to HCC progression was defined as the time from the first detec-
tion of the nodular lesion to HCC diagnosis. To identify risk fac-
tors associated with HCC progression among nodules, univariate
and multivariate analyses were performed using Cox regression
analysis. Independent risk factors obtained from multivariate
Cox analysis were entered into the HCC prediction risk score
model. The risk score was calculated by the weighted sum of the
selected variables, and the weights were defined as the quotient
(rounded to nearest integer) of estimated corresponding β -coef-
72
ficient divided by the smallest β derived from multivariate Cox
regression analysis in a stepwise method. The performance of the
selected model was evaluated by leave-one-out cross-validation.
79 Receiver operating characteristic curves and area under the curve
(AUC) (95% confidence interval (CI)) at 36 and 60 months were
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-
ues for dividing each group were determined by considering the
best discriminatory ability and monotonicity.
Unweighted birater κ-coefficient and its 95% CI were used to
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;
0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and
0.81–1.00, almost perfect (20). Statistical analysis for κ was performed
with MedCalc version 16.4.3 (MedCalc, Mariakerke, Belgium).
RESULTS
Interreader agreement for categorizing indeterminate nodules
Overall agreement for interpreting indeterminate nodules between
two readers was substantial, with unweighed κ -coefficient value
of 0.77 (95% CI=0.72–0.82).
Risk factors associated with HCC progression of indeterminate
nodules in all included nodules
The Kaplan–Meier analysis and Cox regression analysis were per-
formed to identify risk factors associated with indeterminate nod-
ule to HCC progression. Figure 3 shows Kaplan–Meier curves
comparing groups divided by age, liver cirrhosis etiology, arterial
enhancement pattern, and nodule size. Multivariate Cox regres-
sion analysis revealed that baseline parameters associated with
HCC progression in indeterminate nodules were old age (year;
hazard ratio (HR)=1.04; 95% CI=1.01–1.07; P=0.002), arterial
enhancement (yes; HR=2.72; 95% CI=1.58–4.68; P<0.001), large
nodule size (>1 cm; HR=7.18; 95% CI=4.08–12.67; P<0.001), low
serum albumin level (≤3.5 g/dl; HR=1.72; 95% CI=0.98–3.02;
P=0.060), high serum AFP level (≥100 ng/ml; HR=4.48; 95%
CI=1.79–11.19; P=0.001), and presence of prior HCC history
(yes; HR=3.25; 95% CI=1.44–7.30; P=0.004). (Table 2).
Subgroup analysis was performed to identify risk factors asso-
ciated with early HCC diagnosis within 12 months by using
logistic regression analysis. A total of 72 indeterminate nodules
diagnosed to HCC between 6 and 12 months from their first
www.nature.com/ajg
 Outcome of Indeterminate Nodules Detected on Cirrhotic Liver 5

a 1.0 b 1.0
Age ≤ 55 HCV HBV

Cumulative incidence rate

Cumulative incidence rate
Age > 55 Alcoholic Other cause

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

Time to HCC progression (months) Time to HCC progression (months)

No. of subject No. of subject
at risk at risk
Age≤55 304 263 177 130 94 61 HBV 373 304 206 142 103 66
Age>55 190 137 83 46 32 18 HCV 19 13 7 3 2 1
Alcoholic 77 61 37 23 15 10
Other cause 25 22 10 8 6 2

c 1.0 d
Arterial enhancement(–) 1.0
Size ≤ 1 cm
Cumulative incidence rate

Arterial enhancement(+)

Cumulative incidence rate

0.8 R/O AP shunt Size > 1 cm
of HCC progression

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).
Finally, old age, arterial enhancement, large nodule size, low
platelet count, and high serum AFP level were identified as inde-
pendent risk factors for predicting early HCC diagnosis within
12 months (Table 3), and those almost corresponded to the risk
factors of HCC progression of indeterminate nodules demon-
strated at Table 2.
Clinical and radiologic variables associated with HCC progression
of indeterminate nodules detected on HBV-related cirrhosis
Subgroup analysis was performed in 373 subjects from 356
patients with HBV-related cirrhosis to identify risk factors asso-
ciated with indeterminate nodule to HCC progression in HBV-
related cirrhosis. Figure 4 shows comparison of Kaplan–Meier
plot between groups divided by HBeAg positivity, baseline HBV
DNA titer, and groups according to antiviral therapy and/or HBV
DNA level. HBeAg positivity, high HBV DNA titer, and the no
antiviral treatment group despite high HBV DNA titer were sig-
nificantly associated with increasing risk of HCC progression.
Cox regression analyses (Table 4) were performed to identify
independent risk factors for the development of HCC. After
adjusting compounding factors, old age (year; HR=1.06; 95%
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;
95% CI=3.70–13.57; P<0.001), low serum albumin level (≤3.5 g/
dl; HR=3.57; 95% CI=1.72–7.42; P=0.001), high serum AFP level
(≥100 ng/ml; HR=6.04; 95% CI=1.69–21.56; P=0.006), prior HCC
history (yes; HR=4.24; 95% CI=1.76–10.19; P=0.001), and HBeAg
positivity (HR=2.31; 95% CI=1.26–4.24; P=0.007) were identi-
fied as independent risk factors for predicting HCC progression
of indeterminate nodule detected on HBV-related liver cirrhosis.
These risk factors were used to develop risk prediction model.

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 6 Cho et al.

Table 2. Univariate and multivariate analysis to evaluate the risk factors associated with indeterminate nodule to hepatocellular carcinoma
progression (N=494)

Variables Univariate Multivariate

LIVER

HR (95% CI) P HR (95% CI) P

Gender, male 1.05 (0.65–1.71) 0.847

Age, year 1.06 (1.04–1.08) <0.001 1.04 (1.01–1.07) 0.002
Background liver disease, CHC 2.99 (144–6.20) 0.003
Arterial enhancement, yes 3.42 (2.19–5.34) <0.001 2.72 (1.58–4.68) <0.001
Nodule size, >1 cm 9.70 (5.96–15.80) <0.001 7.18 (4.08–12.67) <0.001
Child–Pugh class, B and C 2.22 (1.30–3.78) 0.003
Platelet, <100×109/l 2.01 (1.23–3.30) 0.006
Albumin, ≤3.5 g/dl 3.31 (2.00–5.47) <0.001 1.72 (0.98–3.02) 0.060
Bilirubin, >2 mg/dl 1.13 (0.49–2.62) 0.773
ALT, >80 IU/l 2.25 (1.21–4.21) 0.011
AFP, ≥100 ng/ml 7.21 (3.60–14.46) <0.001 4.48 (1.79–11.19) 0.001
Prior HCC history, yes 3.04 (1.61–5.74) 0.001 3.25 (1.44–7.30) 0.004
AFP, α -fetoprotein; ALT, alanine transaminase; CHC, chronic hepatitis C; CI, confidence interval; HCC, hepatocellular carcinoma; HR, hazard ratio.
Harrell’s C index=0.843 (s.e.=0.038).

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)

Variables Univariate Multivariate

OR (95% CI) P OR (95% CI) P

Gender, male 0.90 (0.51–1.60) 0.714

Age, year 1.04 (1.02–1.07) <0.001 1.05 (1.02–1.09) 0.004
Background liver disease, CHC 6.54 (2.72–15.74) <0.001
Arterial enhancement, yes 4.67 (2.72–8.00) <0.001 6.62 (3.11–14.07) <0.001
Nodule size, >1 cm 11.43 (6.50–20.11) <0.001 8.47 (4.12–17.41) <0.001
Child–Pugh class, B and C 3.20 (1.75–5.85) <0.001
Platelet, <100×109/l 2.91 (1.62–6.23) <0.001 1.86 (0.91–3.82) 0.089
Albumin, ≤3.5 g/dl 3.261 (1.77–6.00) <0.001
Bilirubin, >2 mg/dl 2.49 (1.20–5.19) 0.014
ALT, >80 IU/l 1.74 (0.75–4.02) 0.198
AFP, ≥100 ng/ml 9.05 (2.49–32.92) 0.001 8.87 (1.26–62.50) 0.029
Prior HCC history, yes 1.84 (0.65–5.18) 0.250
AFP, α -fetoprotein; ALT, alanine transaminase; CHC, chronic hepatitis C; CI, confidence interval; HCC, hepatocellular carcinoma; OR, odds ratio.

Risk prediction model for indeterminate nodule to HCC
progression in B-viral cirrhosis
The risk score was calculated by weighted sum using β -coefficient
derived from multivariate Cox regression analysis. The formula of
the risk score was as follows. The β -coefficient of the correspond-
ing risk factors and their risk score is listed in Table 5.
Risk score = 1× age (years) + 19 × enhancement pattern (arterial
non-enhancement or AP shunt = 0; arterial enhancement = 1)
+ 42 × size (≤ 1 cm = 0; > 1 cm = 1) + 16 × serum albumin
(> 3.5 g/dL = 0; ≤ 3.5 g/dL = 1) + 31× serum AFP (< 100 ng/
mL = 0; ≥ 100 ng/mL = 1) + 32× prior HCC history (no = 0;
yes = 1) + 18 × HBeAg (negative = 0; positive = 1)

The American Journal of GASTROENTEROLOGY www.nature.com/ajg

 Outcome of Indeterminate Nodules Detected on Cirrhotic Liver 7

a b
1.0 1.0
Low HBV DNA titer HBeAg (–)
High HBV DNA titer HBeAg (+)

Cumulative incidence rate

Cumulative incidence rate
0.8 0.8

of HCC progression
of HCC progression

LIVER
0.6 0.6

0.4 P<0.001 0.4 P<0.001

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

0.8 No antiviral therapy, high HBV DNA titer

of HCC progression

0.6
P<0.001
0.4

0.2

0.0

12 24 36 48 60 72

Time to HCC progression (months)

No. of subject
at risk
Antiviral therapy 240 202 135 85 62 33
Low HBV DNA titer 77 61 45 34 26 22
High HBV DNA titer 56 41 26 23 15 11

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

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 8 Cho et al.

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)

Variables Univariate Multivariate

LIVER

HR (95% CI) P HR (95% CI) P

Gender, male 0.90 (0.521–1.54) 0.700

Age, years 1.06 (1.04–1.09) <0.001 1.05 (1.02–1.08) 0.0033
Arterial enhancement, yes 3.19 (1.89–5.37) <0.001 2.45 (1.24–4.84) 0.001
Nodule size,>1 cm 11.92 (6.57–21.65) <0.001 7.44 (3.75–14.78) <0.001
Child–Pugh class, B and C 2.50 (1.63–3.85) <0.001
Platelet, <100×109/l 1.88 (1.085–3.27) 0.026
Albumin, ≤3.5 g/dl 4.81 (2.73–8.48) <0.001 2.14 (1.10–4.14) 0.024
Bilirubin, >2 mg/dl 1.23 (0.44–3.04) 0.690
ALT, >80 IU/l 2.11 (1.06–4.21) 0.034
AFP, ≥100 ng/ml 6.46 (2.93–14.25) <0.001 4.43 (1.38–14.17) 0.012
Prior HCC history, yes 3.13 (1.55–6.35) 0.002 4.57 (1.93–10.81) 0.006
HBeAg positivity, positive 2.41 (1.47–3.97) 0.001 2.30 (1.27–4.16) 0.006
HBV DNA, high HBV DNA load 2.69 (1.52–4.76) 0.001
Antiviral treatment, no treatment in high HBV DNA titer 3.18 (1.88–5.39) <0.001
AFP, α -fetoprotein; ALT, alanine transaminase; CI, confidence interval; 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; HR, hazard ratio.
Harrell’s C index=0.878 (s.e.=0.042).

In this study, the cumulative risk of HCC progression in indeter-

Table 5. β -Coefficient of the included risk factors identified from minate nodules was 11.4, 17.6, and 20.1 for 24, 36, and 60 months,
multivariate Cox regression analysis and their corresponding risk respectively. Similarly, Kobayashi et al. (25) reported that 18.8%
scores
of nodules developed to HCC during the follow-up period among
Variables β -Coef- HR (95% CI) P Risk 154 nodules detected on chronic liver disease. All 154 nodules
ficient score were subdivided by pathologic diagnosis, high-grade dysplastic,
Age, years 0.0475 1.05 (1.02–1.08) 0.0033 1 low-grade dysplastic, and regenerative nodules. Cumulative HCC
Arterial 0.8951 2.45 (1.24–4.84) 0.001 19
development rate at year 5 was ∼ 80%, 36.6%, and 12.4% for high-
enhancement, yes grade dysplastic, low-grade dysplastic, and regenerative nodules,
Nodule size, 2.0071 7.44 (3.75–14.78) <0.001 42 respectively. Similar to this literature, pathologic confirmation
>1 cm would be the reliable and precise method for predicting the out-
Albumin, ≤3.5 g/dl 0.7594 2.14 (1.10–4.14) 0.024 16 come of nodular lesions (25,26); however, biopsy for all nodu-
lar lesions is impractical considering false-negative results and
AFP, ≥100 ng/ml 1.4882 4.43 (1.38–14.17) 0.012 31
procedure-related complications. Regarding radiologic studies,
Prior HCC history, 1.5190 4.57 (1.93–10.81) 0.006 32
sequential imaging strategy such as CT followed by MR imaging
yes
was proposed for differential diagnosis on inconclusive cirrhosis-
HBeAg positivity, 0.8320 2.30 (1.27–4.16) 0.006 18
positive
associated nodules (3). However, the diagnostic accuracy of this
strategy is relatively unsatisfactory than expected (27–29), and
AFP, α -fetoprotein; HBeAg, hepatitis B envelope antigen; HCC, hepatocellular
carcinoma. therefore we developed a simple-to-use risk-forecasting model
composed of routinely available clinical variables to facilitate using
it in clinical practice.
model for indeterminate nodule to HCC progression in cirrhotic The indeterminate nodules could be categorized into three
liver. In our knowledge, this is the first study to develop risk score groups based on the calculated score using the developed risk
model for predicting nodule to HCC progression in patients with score model developed from the present study. In the low-risk
HBV-related cirrhosis. This risk prediction score model was vali- group, HCC progression rate was measured as 0% and 1% at 24
dated using leave-one-out cross-validation, and the result of the and 36 months, respectively, and it stayed the same at 60 months.
validation test revealed high accuracy and reproducibility of the In the intermediate-risk group, HCC progression rate was 9.2%,
developed model in our study. 14.5%, and 14.5% at 24, 36 and 60 months, respectively, and in

The American Journal of GASTROENTEROLOGY www.nature.com/ajg

 Outcome of Indeterminate Nodules Detected on Cirrhotic Liver 9

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

Total Leave-one-out Total Leave-one-out

Optima cutoff 87 87 Optima cutoff 87 87

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)

Positive predictive Positive predictive

46.51 (34.48–58.54) 48.09 (36.15–60.03) 60.69 (44.58–76.80) 62.46 (47.21–77.71)
value value
Negative predictive Negative predictive
94.44 (90.87–98.01) 95.15 (91.84–98.46) 94.63 (91.16–98.10) 95.35 (92.16–98.54)
value value

Positive likelihood Positive likelihood

3.75 (2.79–5.35) 4.03 (2.99–5.38) 5.47 (3.85–5.67) 6.08 (4.21–5.75)
ratio ratio

Negative likelihood Negative likelihood

0.25 (0.16–0.40) 0.22 (0.14–0.36) 0.20 (0.12–0.33) 0.18 (0.11–0.30)
ratio ratio

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

0.8 Low risk

schedule. However, the HCC progression risk for the high-risk
0.6 group was ∼ 40% at 24 months, and it was increased up to 63.1% at
60 months; therefore, aggressive diagnostic approach using biopsy
0.4 or MR imaging should be considered, and they should be followed
up very cautiously.
0.2
Several variables that could not be entered into the final model,
such as the Child–Pugh class, platelet count, HBV DNA titer, and
0.0
antiviral treatment, are important risk factors of HCC progression
12 24 36 48 60 72
based on previous literatures (30–32). Particularly, all B-viral-
Time to HCC progression (months) associated variables, including HBeAg status, HBV DNA level, and
No. of subject
at risk antiviral treatment, are well-known risk factors associated with
Low risk 138 129 96 73 61 38 HCC development in CHB patients. All these variables were sig-
Intermediate risk 103 85 50 36 22 17
nificant in the univariate analysis of our study; however, only the
High risk 67 39 26 16 11 6
HBeAg status could be included in the risk-calculating model after
Figure 6. Comparison of hepatocellular carcinoma progression (HCC)
multivariate analysis. Long-term suppression of HBV by potent
incidence between three groups divided by calculated risk score as low-,
intermediate-, and high-risk groups.
antiviral therapy could lead to regression of fibrosis and cirrhosis
based on previous literatures (33,34), and recently HCC-preven-
tive effect by long-term suppression of HBV using antiviral agent
the high-risk group, it was 41.6%, 53.3%, and 63.1% at 24, 36 and has also been accumulated (35,36). Therefore, aggressive antiviral
60 months, respectively. In the low-risk group, cumulative HCC treatment should be considered in patients with HBV-related
progression rate was not significantly increased at 60 months com- cirrhosis.

© 2016 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

 10 Cho et al.
Although the risk score model is very useful and reproducible, it
has a limitation of generalizability. Because this model was derived
from the subjects detected on only HBV-related cirrhotic liver, it
would not be generally applicable for subjects detected on noncir-
LIVER
rhotic liver or non-HBV-related-cirrhotic liver. Further study would
be required for overcoming the limitation of the present study.
In conclusion, we developed a potent and simple risk model
for predicting indeterminate nodule to HCC progression in
HBV-related liver cirrhosis by combining widely available clin-
ical variables to be used easily in clinical practice. This model
could be helpful in clinical decision in cirrhosis-associated inde-
terminate nodules and, finally, it could provide comprehensive
and individualized medicine for each patient with HBV-related
cirrhosis.
CONFLICT OF INTEREST
Guarantor of the article: Sung Won Cho, MD, PhD.
Specific author contributions: Analysis of data and drafting of the
manuscript: Hyo Jung Cho; drafting of the manuscript and critical
supervision on radiologic aspect: Bohyun Kim; statistical analysis:
Jung-Dong Lee and Dae Ryong Kang; study supervision on radiologic
aspect: Jei Hee Lee and Jai Keun Kim; study planning: Sung Jae Shin
and Kee Myung Lee; study supervision: Byung Moo Yoo and Kwang
Jae Lee; critical revision of the manuscript for important intellectual
content: Soon Sun Kim and Jae Youn Cheong; study concept: Sung
Won Cho. All authors have approved the final revision.
Financial support: None.
Potential competing interests: None.
Study Highlights
WHAT IS CURRENT KNOWLEDGE
✓ Indeterminate nodules with undetermined malignant po-
tential are often detected on cirrhotic liver during hepato-
cellular carcinoma (HCC) surveillance.
✓ Indeterminate nodules should be differentially diagnosed
with benign lesions, premalignant lesions, and early HCC.
✓ Estimation of malignant potential for indeterminate nod-
ules is difficult particularly in nodules < 2 cm.
✓ Noninvasive and easy-to-use risk estimating model for in-
determinate nodules would be helpful in clinical practice.
WHAT IS NEW HERE
✓ We developed easy-to-use risk estimating model for
predicting indeterminate nodules to HCC progression.
✓ We categorized indeterminate nodules by the risk model:
low, intermediate, and high risk for HCC.
REFERENCES
1. Jemal A, Ward E, Hao Y et al. Trends in the leading causes of death in the
United States, 1970-2002. JAMA 2005;294:1255–9.
2. Bosch FX, Ribes J, Diaz M et al. Primary liver cancer: worldwide incidence
and trends. Gastroenterology 2004;127:S5–S16.
3. Bruix J, Sherman M. Management of hepatocellular carcinoma: an update.
Hepatology 2011;53:1020–2.
The American Journal of GASTROENTEROLOGY
4. European Association For The Study Of The Liver. European Organisation For
Research And Treatment Of Cancer. EASL–EORTC clinical practice guide-
lines: management of hepatocellular carcinoma. J Hepatol 2012;56:908–43.
5. Omata M, Lesmana LA, Tateishi R et al. Asian Pacific Association for the
Study of the Liver consensus recommendations on hepatocellular carci-
noma. Hepatol Int 2010;4:439–74.
6. Brown JJ, Naylor MJ, Yagan N. Imaging of hepatic cirrhosis. Radiology
1997;202:1–16.
7. Guyot C, Lepreux S, Combe C et al. Hepatic fibrosis and cirrhosis: the (myo)
fibroblastic cell subpopulations involved. Int J Biochem Cell Biol 2006;38:
135–51.
8. International Working Party. Terminology of nodular hepatocellular le-
sions. Hepatology 1995;22:983–93.
9. Hanna RF, Aguirre DA, Kased N et al. Cirrhosis-associated hepatocellular
nodules: correlation of histopathologic and MR imaging features 1. Radio-
graphics 2008;28:747–69.
10. Kim TK, Lee E, Jang H-J. Imaging findings of mimickers of hepatocellular
carcinoma. Clin Mol Hepatol 2015;21:326–43.
11. Yu SJ. A concise review of updated guidelines regarding the management of
hepatocellular carcinoma around the world: 2010–2016. Clin Mol Hepatol
2016;22:7.
12. Stewart CJ, Coldewey J, Stewart IS. Comparison of fine needle aspiration
cytology and needle core biopsy in the diagnosis of radiologically detected
abdominal lesions. J Clin Pathol 2002;55:93–7.
13. Durand F, Regimbeau JM, Belghiti J et al. Assessment of the benefits and
risks of percutaneous biopsy before surgical resection of hepatocellular
carcinoma. J Hepatol 2001;35:254–8.
14. Grant A, Neuberger J. Guidelines on the use of liver biopsy in clinical prac-
tice. Gut 1999;45:IV1–IV11.
15. Sakamoto M, Hirohashi S, Shimosato Y. Early stages of multistep hepato-
carcinogenesis: adenomatous hyperplasia and early hepatocellular carci-
noma. Hum Pathol 1991;22:172–8.
16. Choi BI, Lee KH, Han JK et al. Hepatic arterioportal shunts: dynamic CT
and MR features. Korean J Radiol 2002;3:1–15.
17. Choi J-Y, Lee J-M, Sirlin CB. CT and MR imaging diagnosis and staging of
hepatocellular carcinoma: part II. Extracellular agents, hepatobiliary agents,
and ancillary imaging features. Radiology 2014;273:30–50.
18. Lee HS, Kim JK, Cheong JY et al. Prediction of compensated liver cirrhosis
by ultrasonography and routine blood tests in patients with chronic viral
hepatitis. Korean J Hepatol 2010;16:369–75.
19. Crewson PE. Reader agreement studies. AJR Am J Roentgenol
2005;184:1391–7.
20. Landis JR, Koch GG. The measurement of observer agreement for categori-
cal data. Biometrics 1977;33:159–74.
21. Chen CJ, Yang HI. Natural history of chronic hepatitis B REVEALed. J
Gastroenterol Hepatol 2011;26:628–38.
22. Tseng TC, Liu CJ, Yang HC et al. High levels of hepatitis B surface antigen
increase risk of hepatocellular carcinoma in patients with low HBV load.
Gastroenterology 2012;142:1140–9 e3, quiz e13-4.
23. Yang HI, Yuen MF, Chan HL et al. Risk estimation for hepatocellular carci-
noma in chronic hepatitis B (REACH-B): development and validation of a
predictive score. Lancet Oncol 2011;12:568–74.
24. Kao JH. Risk stratification of HBV infection in Asia-Pacific region. Clin
Mol Hepatol 2014;20:223–7.
25. Kobayashi M, Ikeda K, Hosaka T et al. Dysplastic nodules frequently de-
velop into hepatocellular carcinoma in patients with chronic viral hepatitis
and cirrhosis. Cancer 2006;106:636–47.
26. Sato T, Kondo F, Ebara M et al. Natural history of large regenerative
nodules and dysplastic nodules in liver cirrhosis: 28-year follow-up study.
Hepatol Int 2015;9:330–6.
27. Khalili K, Kim TK, Jang H-J et al. Optimization of imaging diagnosis of 1–2
cm hepatocellular carcinoma: an analysis of diagnostic performance and
resource utilization. J Hepatol 2011;54:723–8.
28. Sangiovanni A, Manini MA, Iavarone M et al. The diagnostic and economic
impact of contrast imaging techniques in the diagnosis of small hepatocel-
lular carcinoma in cirrhosis. Gut 2010;59:638–44.
29. Sersté T, Barrau V, Ozenne V et al. Accuracy and disagreement of com-
puted tomography and magnetic resonance imaging for the diagnosis of
small hepatocellular carcinoma and dysplastic nodules: role of biopsy.
Hepatology 2012;55:800–6.
30. Bolondi L, Sofia S, Siringo S et al. Surveillance programme of cirrhotic
patients for early diagnosis and treatment of hepatocellular carcinoma: a
cost effectiveness analysis. Gut 2001;48:251–9.
www.nature.com/ajg

31. Velázquez RF, Rodriguez M, Navascues CA et al. Prospective analysis of
risk factors for hepatocellular carcinoma in patients with liver cirrhosis.
Hepatology 2003;37:520–7.
32. Yuen M-F, Yuan H-J, Wong DK et al. Prognostic determinants for chronic
hepatitis B in Asians: therapeutic implications. Gut 2005;54:1610–4.
33. Marcellin P, Gane E, Buti M et al. Regression of cirrhosis during treatment
with tenofovir disoproxil fumarate for chronic hepatitis B: a 5-year open-
label follow-up study. Lancet 2013;381:468–75.
© 2016 by the American College of Gastroenterology
Outcome of Indeterminate Nodules Detected on Cirrhotic Liver 11
34. Chang TT, Liaw YF, Wu SS et al. Long‐term entecavir therapy results in the
reversal of fibrosis/cirrhosis and continued histological improvement in
patients with chronic hepatitis B. Hepatology 2010;52:886–93.
35. Colombo M, Iavarone M. Role of antiviral treatment for HCC prevention.
Best Pract Res Clin Gastroenterol 2014;28:771–81.
LIVER
36. Fernández-Rodríguez CM, Gutiérrez-García ML. Prevention of hepatocel-
lular carcinoma in patients with chronic hepatitis B. World J Gastrointest
Pharmacol Ther 2014;5:175–82.
AQ3
The American Journal of GASTROENTEROLOGY