ORIGINAL ARTICLE
Blood Gene Signature for Early Hepatocellular Carcinoma Detection in Patients With Chronic Hepatitis B Haniza Omar, MD,* Chun Ren Lim, PhD,w Samuel Chao, MSc,z Michelle Mei Lin Lee, MSc,w Chin Wei Bong, PhD,w Edie Jian Jiek Ooi, BSc,w Choon Geok Yu, MD,w Soek Siam Tan, MD,* Muhammad Radzi Abu Hassan, MD,y Jayaram Menon, MD,8 Raman Muthukaruppan, MD,8 Mandeep Singh, MD,8 Nik Azim Nik Abdullah, MD,z Boon Phoe Ooi, MD,# Robert Phooi Huat Ding, MD,** Eng Joo Low, MD,ww Francis Tan, MD,ww David Novak, MSc,z David F. Harris, MD,zz Hengxuan Yang, MD,z Ismail Merican, MD,* and Choong-Chin Liew, PhDzyy
Purpose: Up to 25% of chronic hepatitis B (CHB) patients eventually develop hepatocellular carcinoma (HCC), a disease with poor prognosis unless detected early. This study identifies a bloodbased RNA biomarker panel for early HCC detection in CHB. Materials and Methods: A genome-wide RNA expression study was performed using RNA extracted from blood samples from Malaysian patients (matched HCC, CHB, controls). Genes differentiating HCC from controls were selected for further testing using quantitative real-time polymerase chain reaction. Finally, a 6-gene biomarker panel was identified and characterized using a training set (cohort I = 126), and tested against 2 test sets (cohort II = 222; cohort III = 174). The total number of samples used for each group is: HCC + CHB = 143, CHB = 211, control = 168. Results: Our gene panel displays a consistent trend distinguishing HCC from controls in our test sets, with an area under receiveroperating characteristic curve of 0.9 in cohort III. Our independent test set (cohort III) showed that the gene panel had a sensitivity of 70% with a specificity of 92%. The biomarker profile for HCC was consistently detected in a small subgroup of CHB patients, thus potentially predicting early, preclinical cases of cancer that should be screened more intensively. Conclusion: The biomarkers identified in this study can be used as the basis of a blood-based test for the detection of early HCC in CHB.
Received for publication August 12, 2013; accepted February 4, 2014. From the *Selayang Hospital, Lebuhraya Selayang-Kepong, Batu Caves, Selangor Darul Ehsan; wGeneNews (Malaysia) Sdn. Bhd., Mount Miriam Cancer Hospital, Tanjung Bungah; **Island Hospital, Penang; ySultanah Bahiyah Hospital, Jalan Langgar, Alor Setar, Kedah; 8Queen Elizabeth Hospital, Kota Kinabalu, Sabah; zUmum Kuching Sarawak Hospital, Jalan Tun Ahmad Zaidi Adruce, Kuching, Sarawak; #Penang Hospital, Jalan Residensi; wwLam Wah Ee Hospital, Pulau Pinang; zzGeneNews Diagnostics, Office/Lab 9, Biotechnology Incubation Centre, Technology Park Malaysia, Kuala Lumpur, Malaysia; zGeneNews Ltd., Richmond Hill, ON, Canada; and yyBrigham and Women’s Hospital, Harvard Medical School, Boston, MA. Supported by GeneNews Ltd. and the Malaysian Ministry of Health (NMRR-09-317-3991 Ministry of Health, Government of Malaysia). C.R.L., H.Y., S.C., C.W.B., C.G.Y., E.J.J.O., M.M.L.L., D.F.H., D.N., and C.-C.L. are employees of GeneNews Ltd., that also sponsored this research. C.-C.L. is the Chief Scientist of GeneNews Ltd, and holds shares in the company. The remaining authors declare that they have nothing to disclose. Reprints: Choong-Chin Liew, PhD, GeneNews Ltd., 2 East Beaver Creek Road, Building 2, Richmond Hill, ON, Canada L4B 2N3 (e-mail: [email protected]). Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved.
150 | www.jcge.com
Key Words: hepatocellular carcinoma, gene expression, diagnostics, blood, biomarkers
(J Clin Gastroenterol 2015;49:150–157)
H
epatocellular carcinoma (HCC) is the fifth most common cancer in men worldwide and the third leading cause of cancer-related death. According to GLOBOCAN estimates, approximately 750,000 new cases were diagnosed in 2008 and 700,000 people died of liver cancer.1 HCC disproportionately affects populations in East and Southeast Asia, which account for >70% of all newly diagnosed cases. The high prevalence of the disease in Asia can be explained by the strong linkage between liver cancer and chronic hepatitis B (CHB) virus infection. Of the world’s 400 million current chronic carriers of hepatitis B virus, three quarters live in China and Southeast Asia,2 and some 15% to 25% of those infected will eventually develop HCC.3,4 Other risk factors for HCC include hepatitis C virus infection, alcohol-related cirrhosis, and nonalcoholic steatohepatitis. The incidence of the disease is increasing in developed countries. For example, HCC incidence has doubled in the United States over the past 20 years mainly because of immigrations from Asia and currently represents the most rapidly increasing cause of cancer-related death in that country. HCC is highly lethal; it develops gradually and has a clinically indolent course in its early stages. It often presents first at a late stage with liver failure or other serious complications. At this stage, a cure is no longer feasible; almost all patients who develop late-stage liver cancer will be dead within 1 year.5 Early-stage HCC is potentially curable. Therefore, the focus in hepatocellular oncology has been the early detection of small lesions more amenable to resection, transplantation, or radiofrequency ablation.6 To this end, several groups have recommended surveillance programs to monitor at-risk populations.7,8 The American Association for the Study of Liver Diseases7 and the Asian Pacific Association for the Study of the Liver8 recommend surveillance of certain defined populations at high risk of developing liver cancer, such as Asian men and women with CHB infection over age 40 and age 50, respectively. In this study we describe a blood-based gene-expression profile for HCC which is particularly directed to surveillance strategies for HCC in CHB patients. Our hypothesis is that a subgroup of CHB patients has already J Clin Gastroenterol
Volume 49, Number 2, February 2015
J Clin Gastroenterol
Volume 49, Number 2, February 2015
Liver Cancer Biomarkers for Hepatitis B Patients
TABLE 1. Clinical Characteristics of the Patient Cohorts for Microarray Hybridization (Cohort I)
P* Characteristics No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female Medical conditions # (%) None Hypertension Diabetes mellitus Hyperlipidemia and dyslipidemia Coronary disease Asthma Others
Control
CHB
HCC + CHB
Control vs. HCC + CHB
50 43.9 ± 11.2 26.0 ± 4.9
49 51.6 ± 12.7 25.9 ± 3.7
47 56.3 ± 10.0 24.7 ± 4.2
< 0.01 0.32
< 0.01 0.53
0.09 0.16
(19.2) (53.2) (12.8) (0.0) (14.9)
< 0.01 < 0.01 1.00 0.24 < 0.01
< 0.01 < 0.01 1.00 0.62 1.00
1.00 0.41 1.00 1.00 0.03
< 0.01 < 0.01
0.14 0.14
0.14 0.14
(51.1) (27.7) (14.9) (6.4)
0.42 0.33 1.00 0.32
0.23 0.17 1.00 0.06
0.84 0.82 1.00 0.36
2 (4.3) 0 (0.0) 9 (19.1)
1.00 0.50 0.40
1.00 1.00 1.00
1.00 0.50 0.26
28 12 7 3 0
(56.0) (24.0) (14.0) (6.0) (0.0)
10 31 6 1 1
(20.4) (63.3) (12.2) (2.0) (2.0)
9 25 6 0 7
28 (56.0) 22 (44.0)
35 (71.4) 14 (28.6)
40 (85.1) 7 (14.9)
30 9 7 7
23 15 7 1
24 13 7 3
(60.0) (18.0) (14.0) (14.0)
2 (4.0) 2 (4.0) 6 (12.0)
(46.9) (30.6) (14.3) (2.0)
2 (4.1) 2 (4.1) 5 (10.2)
Control vs. CHB
CHB vs. HCC + CHB
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
TABLE 2. Clinical Characteristics of the Patients in Cohort II
P* Characteristics No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female Medical conditions # (%) None Hypertension Diabetes mellitus Hyperlipidemia Dyslipidemia Coronary disease Asthma Others
Control
CHB
HCC + CHB
75 52.5 ± 13.5 26.2 ± 7.9
75 51.3 ± 12.0 25.8 ± 6.9
72 55.9 ± 12.2 24.4 ± 10.8
53 10 5 6 1
(70.7) (13.3) (6.7) (8.0) (1.3)
19 50 4 2 0
(25.3) (66.7) (5.3) (2.7) (0.0)
11 43 10 0 8
CHB vs. HCC + CHB
0.39 0.71
0.04 0.03
(15.3) (59.7) (13.9) (0.0) (11.1)
< 0.01 < 0.01 0.18 0.03 0.02
< 0.01 < 0.01 1.00 0.28 1.00
0.15 0.40 0.10 0.50 < 0.01
< 0.01 < 0.01
0.13 0.13
0.37 0.37
0.14 0.70 0.23 1.00 1.00 0.02 1.00 0.49
0.74 0.70 0.43 1.00 1.00 0.04 1.00 0.21
0.32 1.00 0.83 0.61 0.50 1.00 1.00 0.56
61 (81.3) 14 (18.7)
63 (87.5) 9 (12.5)
37 16 19 2 1 13 1 3
40 19 14 1 2 4 1 8
45 18 12 2 0 3 0 5
(53.3) (25.3) (18.7) (1.3) (2.7) (5.3) (1.3) (10.7)
Control vs. CHB
0.22 0.02
52 (69.3) 23 (30.7) (49.3) (21.3) (25.3) (2.7) (1.3) (17.3) (1.3) (4.0)
Control vs. HCC + CHB
(62.5) (25.5) (16.7) (2.8) (0.0) (4.2) (0.0) (6.9)
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
151
J Clin Gastroenterol
Omar et al
Volume 49, Number 2, February 2015
TABLE 3. Clinical Characteristics of the Patients in Cohort III
P* Characteristics No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female Medical conditions # (%) None Hypertension Diabetes mellitus Hyperlipidemia Dyslipidemia Coronary disease Asthma Others
Control
CHB
HCC + CHB
Control vs. HCC + CHB
50 44.1 ± 9.8 27.2 ± 5.4
91 50.4 ± 5.4 26.0 ± 6.9
33 56.3 ± 10.8 24.6 ± 5.0
< 0.01 0.02
< 0.01 0.04
< 0.01 0.41
36 4 3 5 2
18 59 14 0 0
(30.3) (60.6) (0.0) (0.0) (9.1)
< 0.01 < 0.01 0.27 0.15 0.38
< 0.01 < 0.01 0.11 < 0.01 0.12
0.23 0.68 0.02 1.00 0.02
< 0.01 < 0.01
0.04 0.04
< 0.01 < 0.01
0.82 0.42 0.78 0.15 1.00 0.70 0.52 1.00
0.07 0.19 0.19 0.05 0.55 0.28 0.29 0.21
0.20 0.02 0.13 1.00 0.56 0.66 1.00 0.29
(72.0) (8.0) (6.0) (10.0) (4.0)
(19.8) (64.8) (15.4) (0.0) (0.0)
10 20 0 0 3
19 (38.0) 31 (62.0)
52 (57.1) 39 (42.9)
32 (97.0) 1 (3.0)
27 9 9 4 0 5 2 2
64 9 9 1 3 4 1 10
19 9 7 0 0 2 0 1
(54.0) (18.0) (18.0) (8.0) (0.0) (10.0) (4.0) (4.0)
(70.3) (9.9) (9.9) (1.1) (3.3) (4.4) (1.1) (11.0)
(57.6) (27.3) (21.2) (0.0) (0.0) (6.1) (0.0) (3.0)
Control vs. CHB
CHB vs. HCC + CHB
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
developed early, preclinical HCC and that these patients should have the same gene-expression profile as patients with confirmed, overt HCC. These early HCC patients will progress to overt HCC much faster than the general CHB group and should therefore be followed up more closely. The aim of this study is to verify our hypothesis and to determine if a larger study is justified. A higher rate of “false”-positive predictions in CHB patients relative to controls would constitute a partial confirmation of our hypothesis. The long-term aim of our research is to develop a noninvasive diagnostic tool for the detection of early, preclinical HCC that cannot be diagnosed by conventional methods in patients with CHB.
Between July 2010 and June 2012, we collected 1388 blood samples. Confirmed cancer (HCC + CHB) samples that met the selection criteria and that passed RNA qualitycontrol checks were selected for cohort I. Control and CHB samples were selected to match these samples for age, body mass index, ethnicity, and other clinical parameters (Table 1). We aimed for a cohort comprising about 50 samples in each category, as this number was shown to achieve adequate power in our other studies involving gene expression signatures in whole blood. The remaining HCC + CHB samples were matched to CHB and controls and divided into independent cohorts II and III (Tables 2 to 4). In total, we selected 152 HCC + CHB, 215 CHB, and 175 control blood samples (Table 5).
MATERIALS AND METHODS
Blood Collection, RNA Isolation, and RNA Quality Control
Blood samples were collected from patients who presented to 9 hospitals in Malaysia and who were diagnosed with HCC with CHB (HCC + CHB). HCC + CHB patients were mainly diagnosed using noninvasive imaging methods, such as computed tomography scan and/or ultrasound scan, and must have confirmed CHB virus infection. Controls comprised patients without hepatitis B/C virus infection or HCC. CHB was defined as a >6-month history of having detectable HBsAg, or other antigens or antibodies indicating hepatitis B virus infection, and at least 1 test proving liver cell damage. Patients coinfected with human immunodeficiency virus or hepatitis C virus were excluded from this study. All study participants were at least 21 years old and able and willing to give informed consent according to protocols approved by the Medical Research and Ethics Committee, Ministry of Health Malaysia.
152 | www.jcge.com
Peripheral whole blood (2 2 5 mL) was collected using PaxGene tubes (PreAnalytix, Hombrechtikon, Switzerland), and whole-blood RNA was isolated as described previously.9 Isolated RNA was checked using 2100 Bioanalyzer RNA 6000 Nano Chips (Agilent Technologies, CA). Samples were excluded from microarray analysis that did not meet the following quality criteria: RINZ7 0; 28S:18S rRNAZ1 0. RNA quantity was determined by absorbance at 260 nm in a DU 800 Spectrophotometer (Beckman Coulter, CA) or a NanoDrop 2000c UV-vis Spectrophotometer (Thermo Scientific, DE).
Microarray Hybridization RNA from 146 samples (cohort I: HCC + CHB = 47; CHB = 49, controls = 50) was analyzed by microarray hybridization. The clinical characteristics of these patient cohorts are presented in Table 1.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
J Clin Gastroenterol
Volume 49, Number 2, February 2015
Liver Cancer Biomarkers for Hepatitis B Patients
TABLE 4. Comparison of Clinical Characteristics Between Cohort I, Cohort II, and Cohort III
P* Characteristics Control No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female CHB No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female HCC + CHB No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female
Cohort I
Cohort II
Cohort III
Cohort I vs. Cohort II
50 43.9 ± 11.2 26.0 ± 4.9
75 52.5 ± 13.5 26.2 ± 7.9
50 44.1 ± 9.8 27.2 ± 5.4
< 0.01 0.64
0.80 0.24
< 0.01 0.08
(72.0) (8.0) (6.0) (10.0) (4.0)
0.13 0.15 0.22 1.00 0.52
0.14 0.05 0.32 0.71 0.49
1.00 0.40 1.00 0.75 0.56
28 12 7 3 0
(56.0) (24.0) (14.0) (6.0) (0.0)
53 10 5 6 1
(70.7) (13.3) (6.7) (8.0) (1.3)
36 4 3 5 2
Cohort I vs. Cohort III
Cohort II vs. Cohort III
28 (56.0) 22 (44.0)
52 (69.3) 23 (30.7)
19 (38.0) 31 (62.0)
0.13 0.13
0.11 0.11
< 0.01 < 0.01
49 51.5 ± 12.6 25.8 ± 3.7
75 51.3 ± 12.0 25.8 ± 6.9
91 50.4 ± 5.4 26.0 ± 6.9
0.91 0.20
0.10 0.19
0.18 0.99
18 59 14 0 0
(19.8) (64.8) (15.4) (0.0) (0.0)
0.52 0.85 0.20 1.00 0.40
1.00 1.00 0.80 0.35 0.35
0.46 0.87 0.05 0.20 1.00
10 31 6 1 1
(20.4) (63.3) (12.2) (2.0) (2.0)
19 50 4 2 0
(25.3) (66.7) (5.3) (2.7) (0.0)
35 (71.4) 14 (28.6)
61 (81.3) 14 (18.7)
52 (57.1) 39 (42.9)
0.27 0.27
0.10 0.10
< 0.01 < 0.01
47 56.3 ± 10.0 24.9 ± 4.1
72 55.9 ± 12.2 24.4 ± 10.8
33 56.3 ± 10.8 24.6 ± 5.0
0.91 0.10
0.96 0.66
0.90 0.34
(30.3) (60.6) (0.0) (0.0) (9.1)
0.62 0.57 1.00 1.00 0.58
0.29 0.65 0.04 1.00 0.51
0.11 1.00 0.03 1.00 1.00
32 (97.0) 1 (3.0)
0.79 0.79
0.13 0.13
0.17 0.17
9 25 6 0 7
(19.2) (53.2) (12.8) (0.0) (14.9)
40 (85.1) 7 (14.9)
11 43 10 0 8
(15.3) (59.7) (13.9) (0.0) (11.1)
63 (87.5) 9 (12.5)
10 20 0 0 3
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
TABLE 5. Sample Size and Results
Groups Control CHB HCC + CHB (early/intermediate/advanced/ unknown) Total ROC AUC: control vs. HCC + CHB
Cohort I Microarray
Cohort I qRT-PCR
Cohort II qRT-PCR
Cohort III qRT-PCR
50 49 47
43 45 38 (8/7/13/10)
75 75 72 (18/9/29/16)
50 91 33 (5/10/12/6)
146 0.8
126 0.9
222 0.8
174 0.9
CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma; qRT-PCR, quantitative real-time polymerase chain reaction.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
153
J Clin Gastroenterol
Omar et al
In accordance with NuGen Ovation RNA Amplification System V2 and NuGen Encore Biotin Module protocols (NuGen, CA), 50 ng of each sample were used for cDNA synthesis and hybridization onto Affymetrix U133 Plus 2 0 GeneChip oligonucleotide arrays (Affymetrix, CA). Hybridizations were assessed by the Affymetrix GeneChip quality threshold suggested by NuGen. QC criteria were: (1) raw Q < 5 0; (2) scaling factor
Blood Gene Signature for Early Hepatocellular Carcinoma Detection in Patients With Chronic Hepatitis B Haniza Omar, MD,* Chun Ren Lim, PhD,w Samuel Chao, MSc,z Michelle Mei Lin Lee, MSc,w Chin Wei Bong, PhD,w Edie Jian Jiek Ooi, BSc,w Choon Geok Yu, MD,w Soek Siam Tan, MD,* Muhammad Radzi Abu Hassan, MD,y Jayaram Menon, MD,8 Raman Muthukaruppan, MD,8 Mandeep Singh, MD,8 Nik Azim Nik Abdullah, MD,z Boon Phoe Ooi, MD,# Robert Phooi Huat Ding, MD,** Eng Joo Low, MD,ww Francis Tan, MD,ww David Novak, MSc,z David F. Harris, MD,zz Hengxuan Yang, MD,z Ismail Merican, MD,* and Choong-Chin Liew, PhDzyy
Purpose: Up to 25% of chronic hepatitis B (CHB) patients eventually develop hepatocellular carcinoma (HCC), a disease with poor prognosis unless detected early. This study identifies a bloodbased RNA biomarker panel for early HCC detection in CHB. Materials and Methods: A genome-wide RNA expression study was performed using RNA extracted from blood samples from Malaysian patients (matched HCC, CHB, controls). Genes differentiating HCC from controls were selected for further testing using quantitative real-time polymerase chain reaction. Finally, a 6-gene biomarker panel was identified and characterized using a training set (cohort I = 126), and tested against 2 test sets (cohort II = 222; cohort III = 174). The total number of samples used for each group is: HCC + CHB = 143, CHB = 211, control = 168. Results: Our gene panel displays a consistent trend distinguishing HCC from controls in our test sets, with an area under receiveroperating characteristic curve of 0.9 in cohort III. Our independent test set (cohort III) showed that the gene panel had a sensitivity of 70% with a specificity of 92%. The biomarker profile for HCC was consistently detected in a small subgroup of CHB patients, thus potentially predicting early, preclinical cases of cancer that should be screened more intensively. Conclusion: The biomarkers identified in this study can be used as the basis of a blood-based test for the detection of early HCC in CHB.
Received for publication August 12, 2013; accepted February 4, 2014. From the *Selayang Hospital, Lebuhraya Selayang-Kepong, Batu Caves, Selangor Darul Ehsan; wGeneNews (Malaysia) Sdn. Bhd., Mount Miriam Cancer Hospital, Tanjung Bungah; **Island Hospital, Penang; ySultanah Bahiyah Hospital, Jalan Langgar, Alor Setar, Kedah; 8Queen Elizabeth Hospital, Kota Kinabalu, Sabah; zUmum Kuching Sarawak Hospital, Jalan Tun Ahmad Zaidi Adruce, Kuching, Sarawak; #Penang Hospital, Jalan Residensi; wwLam Wah Ee Hospital, Pulau Pinang; zzGeneNews Diagnostics, Office/Lab 9, Biotechnology Incubation Centre, Technology Park Malaysia, Kuala Lumpur, Malaysia; zGeneNews Ltd., Richmond Hill, ON, Canada; and yyBrigham and Women’s Hospital, Harvard Medical School, Boston, MA. Supported by GeneNews Ltd. and the Malaysian Ministry of Health (NMRR-09-317-3991 Ministry of Health, Government of Malaysia). C.R.L., H.Y., S.C., C.W.B., C.G.Y., E.J.J.O., M.M.L.L., D.F.H., D.N., and C.-C.L. are employees of GeneNews Ltd., that also sponsored this research. C.-C.L. is the Chief Scientist of GeneNews Ltd, and holds shares in the company. The remaining authors declare that they have nothing to disclose. Reprints: Choong-Chin Liew, PhD, GeneNews Ltd., 2 East Beaver Creek Road, Building 2, Richmond Hill, ON, Canada L4B 2N3 (e-mail: [email protected]). Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved.
150 | www.jcge.com
Key Words: hepatocellular carcinoma, gene expression, diagnostics, blood, biomarkers
(J Clin Gastroenterol 2015;49:150–157)
H
epatocellular carcinoma (HCC) is the fifth most common cancer in men worldwide and the third leading cause of cancer-related death. According to GLOBOCAN estimates, approximately 750,000 new cases were diagnosed in 2008 and 700,000 people died of liver cancer.1 HCC disproportionately affects populations in East and Southeast Asia, which account for >70% of all newly diagnosed cases. The high prevalence of the disease in Asia can be explained by the strong linkage between liver cancer and chronic hepatitis B (CHB) virus infection. Of the world’s 400 million current chronic carriers of hepatitis B virus, three quarters live in China and Southeast Asia,2 and some 15% to 25% of those infected will eventually develop HCC.3,4 Other risk factors for HCC include hepatitis C virus infection, alcohol-related cirrhosis, and nonalcoholic steatohepatitis. The incidence of the disease is increasing in developed countries. For example, HCC incidence has doubled in the United States over the past 20 years mainly because of immigrations from Asia and currently represents the most rapidly increasing cause of cancer-related death in that country. HCC is highly lethal; it develops gradually and has a clinically indolent course in its early stages. It often presents first at a late stage with liver failure or other serious complications. At this stage, a cure is no longer feasible; almost all patients who develop late-stage liver cancer will be dead within 1 year.5 Early-stage HCC is potentially curable. Therefore, the focus in hepatocellular oncology has been the early detection of small lesions more amenable to resection, transplantation, or radiofrequency ablation.6 To this end, several groups have recommended surveillance programs to monitor at-risk populations.7,8 The American Association for the Study of Liver Diseases7 and the Asian Pacific Association for the Study of the Liver8 recommend surveillance of certain defined populations at high risk of developing liver cancer, such as Asian men and women with CHB infection over age 40 and age 50, respectively. In this study we describe a blood-based gene-expression profile for HCC which is particularly directed to surveillance strategies for HCC in CHB patients. Our hypothesis is that a subgroup of CHB patients has already J Clin Gastroenterol
Volume 49, Number 2, February 2015
J Clin Gastroenterol
Volume 49, Number 2, February 2015
Liver Cancer Biomarkers for Hepatitis B Patients
TABLE 1. Clinical Characteristics of the Patient Cohorts for Microarray Hybridization (Cohort I)
P* Characteristics No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female Medical conditions # (%) None Hypertension Diabetes mellitus Hyperlipidemia and dyslipidemia Coronary disease Asthma Others
Control
CHB
HCC + CHB
Control vs. HCC + CHB
50 43.9 ± 11.2 26.0 ± 4.9
49 51.6 ± 12.7 25.9 ± 3.7
47 56.3 ± 10.0 24.7 ± 4.2
< 0.01 0.32
< 0.01 0.53
0.09 0.16
(19.2) (53.2) (12.8) (0.0) (14.9)
< 0.01 < 0.01 1.00 0.24 < 0.01
< 0.01 < 0.01 1.00 0.62 1.00
1.00 0.41 1.00 1.00 0.03
< 0.01 < 0.01
0.14 0.14
0.14 0.14
(51.1) (27.7) (14.9) (6.4)
0.42 0.33 1.00 0.32
0.23 0.17 1.00 0.06
0.84 0.82 1.00 0.36
2 (4.3) 0 (0.0) 9 (19.1)
1.00 0.50 0.40
1.00 1.00 1.00
1.00 0.50 0.26
28 12 7 3 0
(56.0) (24.0) (14.0) (6.0) (0.0)
10 31 6 1 1
(20.4) (63.3) (12.2) (2.0) (2.0)
9 25 6 0 7
28 (56.0) 22 (44.0)
35 (71.4) 14 (28.6)
40 (85.1) 7 (14.9)
30 9 7 7
23 15 7 1
24 13 7 3
(60.0) (18.0) (14.0) (14.0)
2 (4.0) 2 (4.0) 6 (12.0)
(46.9) (30.6) (14.3) (2.0)
2 (4.1) 2 (4.1) 5 (10.2)
Control vs. CHB
CHB vs. HCC + CHB
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
TABLE 2. Clinical Characteristics of the Patients in Cohort II
P* Characteristics No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female Medical conditions # (%) None Hypertension Diabetes mellitus Hyperlipidemia Dyslipidemia Coronary disease Asthma Others
Control
CHB
HCC + CHB
75 52.5 ± 13.5 26.2 ± 7.9
75 51.3 ± 12.0 25.8 ± 6.9
72 55.9 ± 12.2 24.4 ± 10.8
53 10 5 6 1
(70.7) (13.3) (6.7) (8.0) (1.3)
19 50 4 2 0
(25.3) (66.7) (5.3) (2.7) (0.0)
11 43 10 0 8
CHB vs. HCC + CHB
0.39 0.71
0.04 0.03
(15.3) (59.7) (13.9) (0.0) (11.1)
< 0.01 < 0.01 0.18 0.03 0.02
< 0.01 < 0.01 1.00 0.28 1.00
0.15 0.40 0.10 0.50 < 0.01
< 0.01 < 0.01
0.13 0.13
0.37 0.37
0.14 0.70 0.23 1.00 1.00 0.02 1.00 0.49
0.74 0.70 0.43 1.00 1.00 0.04 1.00 0.21
0.32 1.00 0.83 0.61 0.50 1.00 1.00 0.56
61 (81.3) 14 (18.7)
63 (87.5) 9 (12.5)
37 16 19 2 1 13 1 3
40 19 14 1 2 4 1 8
45 18 12 2 0 3 0 5
(53.3) (25.3) (18.7) (1.3) (2.7) (5.3) (1.3) (10.7)
Control vs. CHB
0.22 0.02
52 (69.3) 23 (30.7) (49.3) (21.3) (25.3) (2.7) (1.3) (17.3) (1.3) (4.0)
Control vs. HCC + CHB
(62.5) (25.5) (16.7) (2.8) (0.0) (4.2) (0.0) (6.9)
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
151
J Clin Gastroenterol
Omar et al
Volume 49, Number 2, February 2015
TABLE 3. Clinical Characteristics of the Patients in Cohort III
P* Characteristics No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female Medical conditions # (%) None Hypertension Diabetes mellitus Hyperlipidemia Dyslipidemia Coronary disease Asthma Others
Control
CHB
HCC + CHB
Control vs. HCC + CHB
50 44.1 ± 9.8 27.2 ± 5.4
91 50.4 ± 5.4 26.0 ± 6.9
33 56.3 ± 10.8 24.6 ± 5.0
< 0.01 0.02
< 0.01 0.04
< 0.01 0.41
36 4 3 5 2
18 59 14 0 0
(30.3) (60.6) (0.0) (0.0) (9.1)
< 0.01 < 0.01 0.27 0.15 0.38
< 0.01 < 0.01 0.11 < 0.01 0.12
0.23 0.68 0.02 1.00 0.02
< 0.01 < 0.01
0.04 0.04
< 0.01 < 0.01
0.82 0.42 0.78 0.15 1.00 0.70 0.52 1.00
0.07 0.19 0.19 0.05 0.55 0.28 0.29 0.21
0.20 0.02 0.13 1.00 0.56 0.66 1.00 0.29
(72.0) (8.0) (6.0) (10.0) (4.0)
(19.8) (64.8) (15.4) (0.0) (0.0)
10 20 0 0 3
19 (38.0) 31 (62.0)
52 (57.1) 39 (42.9)
32 (97.0) 1 (3.0)
27 9 9 4 0 5 2 2
64 9 9 1 3 4 1 10
19 9 7 0 0 2 0 1
(54.0) (18.0) (18.0) (8.0) (0.0) (10.0) (4.0) (4.0)
(70.3) (9.9) (9.9) (1.1) (3.3) (4.4) (1.1) (11.0)
(57.6) (27.3) (21.2) (0.0) (0.0) (6.1) (0.0) (3.0)
Control vs. CHB
CHB vs. HCC + CHB
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
developed early, preclinical HCC and that these patients should have the same gene-expression profile as patients with confirmed, overt HCC. These early HCC patients will progress to overt HCC much faster than the general CHB group and should therefore be followed up more closely. The aim of this study is to verify our hypothesis and to determine if a larger study is justified. A higher rate of “false”-positive predictions in CHB patients relative to controls would constitute a partial confirmation of our hypothesis. The long-term aim of our research is to develop a noninvasive diagnostic tool for the detection of early, preclinical HCC that cannot be diagnosed by conventional methods in patients with CHB.
Between July 2010 and June 2012, we collected 1388 blood samples. Confirmed cancer (HCC + CHB) samples that met the selection criteria and that passed RNA qualitycontrol checks were selected for cohort I. Control and CHB samples were selected to match these samples for age, body mass index, ethnicity, and other clinical parameters (Table 1). We aimed for a cohort comprising about 50 samples in each category, as this number was shown to achieve adequate power in our other studies involving gene expression signatures in whole blood. The remaining HCC + CHB samples were matched to CHB and controls and divided into independent cohorts II and III (Tables 2 to 4). In total, we selected 152 HCC + CHB, 215 CHB, and 175 control blood samples (Table 5).
MATERIALS AND METHODS
Blood Collection, RNA Isolation, and RNA Quality Control
Blood samples were collected from patients who presented to 9 hospitals in Malaysia and who were diagnosed with HCC with CHB (HCC + CHB). HCC + CHB patients were mainly diagnosed using noninvasive imaging methods, such as computed tomography scan and/or ultrasound scan, and must have confirmed CHB virus infection. Controls comprised patients without hepatitis B/C virus infection or HCC. CHB was defined as a >6-month history of having detectable HBsAg, or other antigens or antibodies indicating hepatitis B virus infection, and at least 1 test proving liver cell damage. Patients coinfected with human immunodeficiency virus or hepatitis C virus were excluded from this study. All study participants were at least 21 years old and able and willing to give informed consent according to protocols approved by the Medical Research and Ethics Committee, Ministry of Health Malaysia.
152 | www.jcge.com
Peripheral whole blood (2 2 5 mL) was collected using PaxGene tubes (PreAnalytix, Hombrechtikon, Switzerland), and whole-blood RNA was isolated as described previously.9 Isolated RNA was checked using 2100 Bioanalyzer RNA 6000 Nano Chips (Agilent Technologies, CA). Samples were excluded from microarray analysis that did not meet the following quality criteria: RINZ7 0; 28S:18S rRNAZ1 0. RNA quantity was determined by absorbance at 260 nm in a DU 800 Spectrophotometer (Beckman Coulter, CA) or a NanoDrop 2000c UV-vis Spectrophotometer (Thermo Scientific, DE).
Microarray Hybridization RNA from 146 samples (cohort I: HCC + CHB = 47; CHB = 49, controls = 50) was analyzed by microarray hybridization. The clinical characteristics of these patient cohorts are presented in Table 1.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
J Clin Gastroenterol
Volume 49, Number 2, February 2015
Liver Cancer Biomarkers for Hepatitis B Patients
TABLE 4. Comparison of Clinical Characteristics Between Cohort I, Cohort II, and Cohort III
P* Characteristics Control No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female CHB No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female HCC + CHB No. Age (mean ± SD) Body mass index (mean ± SD) Ethnicity # (%) Malay Chinese Indonesian Indian Others Sex Male Female
Cohort I
Cohort II
Cohort III
Cohort I vs. Cohort II
50 43.9 ± 11.2 26.0 ± 4.9
75 52.5 ± 13.5 26.2 ± 7.9
50 44.1 ± 9.8 27.2 ± 5.4
< 0.01 0.64
0.80 0.24
< 0.01 0.08
(72.0) (8.0) (6.0) (10.0) (4.0)
0.13 0.15 0.22 1.00 0.52
0.14 0.05 0.32 0.71 0.49
1.00 0.40 1.00 0.75 0.56
28 12 7 3 0
(56.0) (24.0) (14.0) (6.0) (0.0)
53 10 5 6 1
(70.7) (13.3) (6.7) (8.0) (1.3)
36 4 3 5 2
Cohort I vs. Cohort III
Cohort II vs. Cohort III
28 (56.0) 22 (44.0)
52 (69.3) 23 (30.7)
19 (38.0) 31 (62.0)
0.13 0.13
0.11 0.11
< 0.01 < 0.01
49 51.5 ± 12.6 25.8 ± 3.7
75 51.3 ± 12.0 25.8 ± 6.9
91 50.4 ± 5.4 26.0 ± 6.9
0.91 0.20
0.10 0.19
0.18 0.99
18 59 14 0 0
(19.8) (64.8) (15.4) (0.0) (0.0)
0.52 0.85 0.20 1.00 0.40
1.00 1.00 0.80 0.35 0.35
0.46 0.87 0.05 0.20 1.00
10 31 6 1 1
(20.4) (63.3) (12.2) (2.0) (2.0)
19 50 4 2 0
(25.3) (66.7) (5.3) (2.7) (0.0)
35 (71.4) 14 (28.6)
61 (81.3) 14 (18.7)
52 (57.1) 39 (42.9)
0.27 0.27
0.10 0.10
< 0.01 < 0.01
47 56.3 ± 10.0 24.9 ± 4.1
72 55.9 ± 12.2 24.4 ± 10.8
33 56.3 ± 10.8 24.6 ± 5.0
0.91 0.10
0.96 0.66
0.90 0.34
(30.3) (60.6) (0.0) (0.0) (9.1)
0.62 0.57 1.00 1.00 0.58
0.29 0.65 0.04 1.00 0.51
0.11 1.00 0.03 1.00 1.00
32 (97.0) 1 (3.0)
0.79 0.79
0.13 0.13
0.17 0.17
9 25 6 0 7
(19.2) (53.2) (12.8) (0.0) (14.9)
40 (85.1) 7 (14.9)
11 43 10 0 8
(15.3) (59.7) (13.9) (0.0) (11.1)
63 (87.5) 9 (12.5)
10 20 0 0 3
*P values for age and body mass index were calculated by Mann-Whitney test, P values for ethnicity, sex, and medical conditions were calculated by Fisher exact test. CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma.
TABLE 5. Sample Size and Results
Groups Control CHB HCC + CHB (early/intermediate/advanced/ unknown) Total ROC AUC: control vs. HCC + CHB
Cohort I Microarray
Cohort I qRT-PCR
Cohort II qRT-PCR
Cohort III qRT-PCR
50 49 47
43 45 38 (8/7/13/10)
75 75 72 (18/9/29/16)
50 91 33 (5/10/12/6)
146 0.8
126 0.9
222 0.8
174 0.9
CHB indicates chronic hepatitis B; HCC, hepatocellular carcinoma; qRT-PCR, quantitative real-time polymerase chain reaction.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
153
J Clin Gastroenterol
Omar et al
In accordance with NuGen Ovation RNA Amplification System V2 and NuGen Encore Biotin Module protocols (NuGen, CA), 50 ng of each sample were used for cDNA synthesis and hybridization onto Affymetrix U133 Plus 2 0 GeneChip oligonucleotide arrays (Affymetrix, CA). Hybridizations were assessed by the Affymetrix GeneChip quality threshold suggested by NuGen. QC criteria were: (1) raw Q < 5 0; (2) scaling factor
![[Incidence of hepatocellular carcinoma in chronic infection with hepatitis B].](/assets/img/hold.png)
