Dichotomous associations of liver pathology with hepatocellular carcinoma morphology in Middle Africa: the situation in Cameroon
Amougou et al. BMC Res Notes
Dichotomous associations of liver pathology with hepatocellular carcinoma morphology in Middle Africa: the situation in Cameroon
Marie Atsama Amougou 0 2 3
Paul Jean Adrien Atangana 0 3
Alice Ghislaine Ndoumba Afouba 0 3
Paul Fewou Moundipa 2
Pascal Pineau 1
Richard Njouom 0 3
0 Virology Unit, Centre Pasteur of Cameroon , BP 1274 Yaoundé , Cameroon
1 Unité « Organisation nucléaire et Oncogenèse », INSERM U993, Institut Pasteur , Paris, 75015 Paris , France
2 Laboratory of Pharmacology and Toxicology, University of Yaounde I , BP 815 Yaounde , Cameroon
3 Virology Unit, Centre Pasteur of Cameroon , BP 1274 Yaoundé , Cameroon
Objective: This study evaluates the occurrence of the various morphological subtypes of hepatocellular carcinoma (HCC) and their connections with some risk factors in Cameroonian patients. The database of the 360 liver biopsies received and associated medical records were reviewed for histological and demographic analysis. Archival formalinfixed and paraffin embedded liver biopsy specimens or slide were re-evaluated in malignancies patients. HCC classification was determined according to the World Health Organization criteria. Results: Malignancies were confirmed in 24.7% (89/360) of liver biopsies. Primary liver tumors consisted in 80 cases of HCC and one case of hepatoblastoma. The distribution of the morphological variants of HCC was trabecular pattern (n = 45/80, 56.25%), acinar/pseudoglandular (32.5%) or scirrhous (11.2%). Remarkably, liver steatosis was present in 60.0% (48/80) of patients with HCC, most of them infected with hepatitis C virus (75.8%). Well-differentiated trabecular tumors were significantly associated with important fibrotic and necro-inflammatory activities in livers (P = 0.008) whereas acinar pattern was more frequent on fatty livers (P = 0.02). Our finding indicates that in Middle Africa the morphology of HCC subtypes correlates with changes affecting non-tumor liver tissue. Trabecular subtype is installed by strong liver injury whereas acinar pattern is more often associated with lipid metabolism defects.
Hepatocellular carcinoma; Liver biopsy; Hepatoblastoma; Steatosis; Trabecular; Cameroon
Primary liver cancer (PLC) is associated with a high
mortality rate and it is the second cause of cancer-related
death worldwide [
]. Sub Saharan Africa (SSA) represents
a major region both in terms of incidence and
mortality from HCC, the principal form of PLC [
Cameroon, little is known about the cancer burden in general.
According to the GLOBOCAN 2012, however, PLC is the
second type of cancers in term of incidence and the first
cause of cancer-related mortality in the country .
Globally, the two main histological types of PLC are
hepatocellular carcinoma (HCC) and
cholangiocarcinoma (CCA) [
]. Of these two types, HCC is known as
the leading form of liver cancer worldwide [
addition, HCC is a heterogeneous disease with four majors
morphological subtypes such as trabecular,
pseudoglandular/acinar or solid patterns [
]. It is now clear that
hepatitis B virus (HBV) and hepatitis C virus are the main
etiological viral factor for HCC development worldwide
] and both viruses are highly endemic in middle
No recent study describing the distribution of PLC
histotypes and morphologies is available in the
country. Furthermore, despite the importance of HCC in the
region, recent pathological reports about this tumor are
scarce in SSA as a whole. Such descriptions are all the
more important that the global epidemiology of HCC is
currently changing with the growing importance taken
by dysmetabolic conditions such as liver steatosis and
non-alcoholic steatohepatitis among major risk factors
]. The importance of these conditions is poorly
known in SSA despite the ongoing nutritional transition
that concern segments of the African populations and the
growing prevalence of non-communicable diseases such
diabetes and obesity. The aim of the present study was,
thus, to fill this gap using liver biopsies diagnosed over a
10-year period in Yaounde.
Materials and methods
Data collection and inclusion criteria
This retrospective study was conducted in the Pathology
Unit of Centre Pasteur of Cameroon recognized as the
national reference Laboratory of the Country. This study
was the continuity of another project on primary liver
Cancers approved by the Cameroonian Ethics Commitee
and the Ministry of Health. All the biopsies received from
January 2004 to January 2013 were reviewed and revealed
a total of 2068 biopsies. Only the 360 patients identified
with liver biopsy were include in the present study.
Histological and demographic data regarding age at
diagnosis and sex information were obtained from medical
record and the database of the Unit. The study did not
include information regarding clinical features
concerning patients affected with the tumors because these data
were not available in many cases. All histological
evaluation was performed by the single experienced pathologist
and the histological residents of the Unit.
For histological diagnosis, slides were retrieved from the
archives of the department and red first by the
histological resident and finally by the single experienced
pathologist of the Unit. Microscopic evaluation of the multiples
selected sections was accomplish from sections of
paraffin embedded tissue mounted on glass and stained with
Hematin-Eosin and Masson Trichrome. The size and the
quality (at least 3 portal spaces) of the biopsy specimens
were noted. The liver fibrosis and necro-inflammatory
activity were assessed using the METAVIR scoring
]. Fibrosis was therefore scored on a scale from 0
to 4 (F0 = no fibrosis, F1 = portal fibrosis without septa,
F2 = portal fibrosis and few septa, F3 = numerous septa
without cirrhosis and F4 = cirrhosis) and activity on a
scale from 0 to 3 (A0 = none, A1 = mild, A2 =
moderate and A3 = severe). HCC cases were further group
into morphological sub-types as described according to
WHO classification of tumors [
]. One representative
paraffin-embedded section from each case was selected
to determine the grade of steatosis represented by the
percentage of hepatocytes containing fat droplets [
Data were represented as mean ± standard deviation
(SD). Descriptive analysis was performed to characterize
the demographic variables of the patients. An
independent t test or nonparametric test was used to determine
the difference between groups. Frequencies and
proportions were used for categorical variables. The
differences were determined by Chi square test or a Fisher
exact test. The difference was considered statistically
significant for P < 0.05. All tests were two-sided. Analyses
were performed using SPSS 16.0 and Prism 6.0 statistical
During a 10-year period (January 2004 to January 2013)
a total of 2068 of different types of biopsies were received
and a total of 360 (17.4%) liver biopsies were included
and examined. Among these, 25.8% (93/360) have been
practiced on liver masses suspected to be hepatic
malignancies. Primary liver cancer (PLC) was found in 87.0%
(81/93) of them. For the remaining 12 specimens,
metastatic tumors (5 adenocarcinomas and 1 lymphoma)
represented 6.4% (6/93) of cases, mesenchymal tumors were
found in two patients (angiosarcoma and neurofibroma,
2.3%) whereas kysts or inflammatory pseudo-tumors
composed the last 4 cases (4.3%). The different types of
malignancies found in this study are presented in Table 1.
Two types of PLC were found, HCC representing
98.8% (n = 80/81) and hepatoblastoma (one case, 1.2%).
The mean age of patients diagnosed with HCC was
38.1 ± 15.2 years (median = 33) with the peak of HCC
occurrence between 20 and 39 years age (n = 49/80,
61.3%) and males were predominant among HCC
patients 65% (n = 52/80). Demographic characteristics of
patients diagnosed with HCC are presented in Table 1.
As resume in Table 1, a total of 42 HCC patients
(52.50%) were found with HBV infection whereas 36.2%
(29/80) were found with HCV infection. The etiology
was unknown for 9 (11.5%) HCC patients and the only
2 years-old girl with hepatoblastoma.
The distribution of the morphological subtypes of HCC
are presented in Table 1 and Additional file 1. In the
present study, trabecular patterns were prevalently found
(n = 45/80; 56.25%) whereas scirrhous subtypes were
found in (n = 9/80; 11.25%) and acinar/pseudoglandular
in (n = 26/80; 32.5%).
A minor subset of tumor cases 20.8% (15/72) occurred
on cirrhotic livers while another 25% (n = 18/72)
displayed severe fibrosis stage at F3. Necro-inflammatory
activity was scored as moderate-to-severe (A2–A3)
in 63.0% of HCC cases (Table 1). In addition, 60.2% of
patients with HCC presented liver steatosis with four of
them displaying more than 20% of hepatocyte
involvement. Finally, two cytological features differed
significantly between HCC and hepatitis. Clear cell variants
and ballooning degeneration were less prevalent in
tumors than in non-tumor samples (Table 1).
We subsequently tried to find clinico-biological
correlations that might explain the different forms taken by
tumors found in Cameroonian patients. As presented in
Fig. 1, infectious risk factors were important
determinant for the age of patients with HCC with HBV-infected
being the youngest (33.1 ± 10.5), the nonBnonC the
oldest (51.7 ± 13.2) and HCV-infected subjects occupying an
intermediate position (41.5 ± 1.1).
As expected, HCV infection was associated with two
well-known hallmarks of its histological impact on the
liver. Severe Fibrosis scores (F3-F4) were significantly
more prevalent in case of infection with this agent than
without it (65.5 vs 30.0%, OR = 4.4, 95% CI 1.6–12.3,
P = 0.0011). Similarly, liver steatosis was present in 75.8%
of the livers from HCV-infected patients whereas it was
observed in only 48.9% of others (OR = 3.2, 95% CI 1.1–
9.0, P = 0.031). The reverse was at least true concerning
fibrosis stages for HBV, which presence was less frequent
in F3-F4 than in F0-F2 livers (36.3 vs 71.9%, OR = 0.22,
95% CI 0.08–0.60, P = 0.0041).
Analyses of histological features were informative as
well. Interestingly, characters of the underlying
pathology affecting non-tumor livers were apparently
influencing tumor morphology. Of course, necro-inflammatory
activity and fibrosis were positively correlated. Indeed,
tissues scored as A2–A3 were more often the siege of
a F3–F4 fibrosis than A0–A1 specimens (74.4 vs 0.0%,
OR = 156.2, P < 0.0001). More interestingly, we observed
that trabecular HCC specimens were more often
developing from non-tumor livers with intense A2–A3
necroinflammatory (78.9 vs 43.3%, P = 0.0048) or with severe
F3–F4 fibrosis (60.5 vs 24.1%, P = 0.0059; Fig. 2). By
contrast, when the non-tumor liver was the siege of steatosis,
tumor morphology was more frequently
acinar/pseudoglandular instead (83.3 vs 51.0%, OR = 4.8, 95% CI:
1.4–16.1, P = 0.0101, Fig. 2). These two associations, high
activity-fibrosis/trabecular pattern and steatosis/acinar
pattern tended to be mutually exclusive as the presence
of high activity and pervasive fibrosis were negatively
correlated with acinar/pseudoglandular pattern (P < 0.008
for both features). In parallel, steatosis was strongly
anticorrelating with the macrotrabecular morphology (4.4 vs
28.5%, P = 0.018).
This study identified the different morpho-types of HCC
diagnosed in Cameroon through a retrospective
analysis in the Pathology Unit of the Centre Pasteur of
Cameroon from January 2004 to January 2013. It is the first
0-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90
Class of Ages (years)
Fig. 1 The distribution of age between patients with HCC and
patients biopsied for hepatitis HCC: hepatocellular carcinoma
review on the distribution of the different types of PLC in
HCC represents an overwhelming majority of the PLC
recorded (98.8%), as only one hepatoblastoma case was
reported (1.2%) in the corresponding period. Others
types of PLC such as cholangiocarcinoma (CCA) were
not diagnosed in the present study thus confirming that
this tumor type did not emerged recently in SSA. This
observation suggests that the pattern of tumor
development remains stable since decades in the region. This is
particularly interesting regarding the role of HCV, often
suspected to cause the development of a sizeable
subset of intra-hepatic CCA cases elsewhere [
HCVassociated biliary cells transformation is apparently still
infrequent in Sub Saharan Africa. Our results indicate
that HCC biopsies occur predominantly among 20 to
39-years-old adults in Cameroon. This age distribution is
much younger than what is commonly reported in SSA
for HCC as a whole. This phenomenon is presumably
due to the feasibility of liver biopsy in younger patients.
Indeed, liver tissues and functions including clotting
factors production are often deteriorating with age making
biopsies more at risk of complications and thus
]. A similar distortion, attributable to
similar reasons, was also conspicuous regarding the
prevalence of liver cirrhosis (20%) in the present series.
The present work provides data on the frequency of
the histological sub-types of HCC. Trabecular variants
(micro- or macro-) were the most commonly observed
patterns in the present series (56.3% when take together).
These results are consistent with studies conducted in
South-Eastern Nigeria (49.3%) [
], in Zambia (52.9%) [
in Tanzania (47.9%) [
] and Zaire (31.4%) [
]. Our results
contrast however with a study conducted in Bangladesh
where pseudo glandular subtype of HCC was the most
]. The prevalence of
acinar/pseudoglandular morphotype in our series (32%) is grossly similar to that
recently observed in Tanzania (25%) or Zambia (23%) but
exceeds largely the proportion observed in North-Central
Nigeria (13%). The causes such major differences in pattern
distribution are unknown.
In the current report, 58.8% of HCC patients were
presenting a certain degree of liver steatosis with more than
70% of these cases infected with HCV. The concomitance
of a fatty liver and HCC used to be rather rare both in
East or West Africa in reports published several decades
]. Publications reporting sizeable proportions
of fatty metamorphosis in liver tissues from Africans tend
to appear recently though [
]. The presence of fatty
degeneration to such extent in a series of HCC from the
XXIst century might therefore represent a major evolution
of liver pathophysiology and HCC epidemiology in SSA
Our study reports for the first time the incidence of the
different types of morphologies taken by HCC in
Cameroon. Our results showed that the trabecular subtype is
the predominant presentation but acinar/pseudo-glandular
pattern is another important presentation of HCC in the
country. Our study also reports a remarkable binary
presentation of HCC and corresponding liver tissues with
trabecular tumors arising mostly from fibrotic livers whereas
acinar specimens were primarily observed in fatty livers.
Further investigations needs to be conducted to identify
micro-environmental mechanisms that promote these
The first one is that biological data, clinical or
pathological stages of the tumors and detailed risk factors
associated to the different types of HCC such as HBV DNA and
HCV RNA loads were not available. Finally our survey is
characterized by shortcomings inherent to cross-sectional
observational studies and as such cannot compete with
true case–control studies. However, it has the merit to
provide an up-to-date landscape on the distribution and
Microtrabecular Macrotrabecular Acinar Scirrhous
Fig. 2 The distribution of the morphological variants of Hepatocellular carcinoma according to histological liver feature (steatosis, fibrosis and liver
activity). F0 no fibrosis, F1 portal fibrosis without septa, F2 portal fibrosis and few septa, F3 numerous septa without cirrhosis, F4 cirrhosis
epidemiology associated with the different types of HCC in
Cameroon. Further prospective studies are needed to
confirm our results.
Additional file 1. Microphotographs of the A: (H&E × 10) liver
parenchyma within the limits of normal but showing a slight dilation of
sinusoids; B: (H&E × 10) and C: (H&E × 20) chronic hepatitis with moderate
activity characterized by ballonnisation, cellular clarification of moderate
intensity and the presence of macrovascular steatosis less than 20%; the
morphological variants of HCC. D: (H&E × 4) and E: (H&E × 20) Moderately
differentiated HCC with trabecular/acinar pattern; F: Moderately
differentiated HCC with moderately to severe steatosis pattern (H&E × 40). H&E
HCC: hepatocellular carcinoma; HBV: hepatitis B virus; HCV: hepatitis C virus;
SSA: Sub Saharan Africa; PLC: primary liver cancer; F0: no fibrosis; F1: portal
fibrosis without septa; F2: portal fibrosis and few septa; F3: numerous septa
without cirrhosis; F4: cirrhosis; A0: none; A1: mild; A2: moderate; et A3: severe;
ns: not significant; na: not available.
RN PP and PFM designed the study; MAA reviewed medical records, collected
data and drafted the manuscript; PP performed statistical analysis; PJAA and
AGNA provided the collection of all the data and re-evaluated preserved
biopsy specimens. PP, RN and PFM were involved in editing the manuscript. All
authors read and approved the final manuscript.
We would like to thank Vital Atangana for his assistance in finding medicals
record and the International Network of Pasteur Institutes and DEDONDER
Projects for financial support of this research.
The authors declare that they have no competing interests.
Availability of data and materials
The datasets used and/or analyzed during this study are available from the
corresponding author on reasonable request. Some data generated in the
current study are included as Additional file 1.
Consent for publication
Ethics approval and consent to participate
This study was the continuity of another project on primary liver Cancers
approved by the National Ethics Commitee (Number 199/CNE/SE/2011) and
the Ministry of Health (Number 631–01.12).
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
1. Ferlay J , Soerjomataram I , Dikshit R , Eser S , Mathers C , Rebelo M , et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012 . Int J Cancer . 2015 ; 136 : E359 - 86 .
2. El-Serag HB . Epidemiology of viral hepatitis and hepatocellular carcinoma . Gastroenterology . 2012 ; 142 ( 1264 -1273): e1 .
3. Arzumanyan A , Reis HMGPV , Feitelson MA . Pathogenic mechanisms in HBV- and HCV-associated hepatocellular carcinoma . Nat Rev Cancer . 2013 ; 13 : 123 - 35 . http://www.nature.com/nrc/journal/v13/n2/full/nrc34 49.html. Accessed Feb 9 2016 .
4. Mbakop A , Yomi J , Yankeum J , Nkegoum B , Mouelle Sone A. Cancer localisation in men and women aged over 50 in Cameroon . Bull Cancer . 1997 ; 84 : 1119 - 22 .
5. Srivatanakul P , Sriplung H , Deerasamee S . Epidemiology of liver cancer: an overview . Asian Pac. J Cancer Prev (APJCP) . 2004 ; 5 : 118 - 25 .
6. El-Serag HB , Rudolph KL . Hepatocellular carcinoma: epidemiology and molecular carcinogenesis . Gastroenterology . 2007 ; 132 : 2557 - 76 .
7. Torbenson MS . Morphologic subtypes of hepatocellular carcinoma . Gastroenterol Clin North Am . 2017 ; 46 : 365 - 91 .
8. Anwar WA , Khaled HM , Amra HA , El-Nezami H , Loffredo CA. Changing pattern of hepatocellular carcinoma (HCC) and its risk factors in Egypt: possibilities for prevention . Mutat Res . 2008 ; 659 : 176 - 84 .
9. Forner A , Llovet JM , Bruix J . Hepatocellular carcinoma . Lancet Lond Engl . 2012 ; 379 : 1245 - 55 .
10. Kew MC . Epidemiology of hepatocellular carcinoma in sub-Saharan Africa . Ann Hepatol. 2013 ; 12 : 173 - 82 .
11. Kew MC . Hepatocellular carcinoma in African Blacks: recent progress in etiology and pathogenesis . World J Hepatol . 2010 ; 2 : 65 - 73 .
12. Kew MC . Hepatocellular carcinoma with and without cirrhosis. A comparison in southern African blacks . Gastroenterology . 1989 ; 97 : 136 - 9 .
13. Ludwig J , Viggiano TR , McGill DB , Oh BJ . Nonalcoholic steatohepatitis: Mayo Clinic experiences with a hitherto unnamed disease . Mayo Clin Proc . 1980 ; 55 : 434 - 8 .
14. Brunt EM , Janney CG , Di Bisceglie AM , Neuschwander-Tetri BA , Bacon BR . Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions . Am J Gastroenterol . 1999 ; 94 : 2467 - 74 .
15. Byron D , Minuk GY. Clinical hepatology: profile of an urban, hospitalbased practice . Hepatology . 1996 ; 24 : 813 - 5 .
16. Farrell GC . Non-alcoholic steatohepatitis: what is it, and why is it important in the Asia-Pacific region? J Gastroenterol Hepatol . 2003 ; 18 : 124 - 38 .
17. Bedossa P , Poynard T. An algorithm for the grading of activity in chronic hepatitis C. The METAVIR Cooperative Study Group . Hepatology . 1996 ; 24 : 289 - 93 .
18. Hirohashi S , Ishak K , Kojiro M , Wanless I , Theis N. Tumours of the liver and intrahepatic bile ducts . In: Hamilton SR , Aaltonen LA , editors. Pathology and genetics of tumours of the digestive system . France: Lyon; 2000 . p. 159 - 72 . https://www.iarc.fr/en/publications/pdfs-online/pat-gen/bb2/ BB2.pdf.
19. Poynard T , Ratziu V , McHutchison J , Manns M , Goodman Z , Zeuzem S , et al. Effect of treatment with peginterferon or interferon alfa-2b and ribavirin on steatosis in patients infected with hepatitis C. Hepatology . 2003 ; 38 : 75 - 85 .
20. Razumilava N , Gores GJ . Classification, diagnosis, and management of cholangiocarcinoma . Clin Gastroenterol Hepatol . 2013 ; 11 : 13 - 21 .
21. Ugiagbe EE , Udoh MO . The histopathological pattern of liver biopsies at the University of Benin Teaching Hospital . Niger J Clin Pract . 2013 ; 16 : 526 - 9 .
22. Abdulkareem FB , Banjo AA , Elesha SO , Daramola AO . Histopathological study of liver diseases at the Lagos University Teaching Hospital, Nigeria ( 1989 - 2000 ). Niger Postgrad Med J. 2006 ; 13 : 41 - 6 .
23. Kashala LO , Conne B , Kalengayi MM , Kapanci Y , Frei PC , Lambert PH . Histopathologic features of hepatocellular carcinoma in Zaire . Cancer . 1990 ; 65 : 130 - 4 .
24. Seleye-Fubara D , Jebbin NJ . Hepatocellular carcinoma in Port Harcourt, Nigeria: clinicopathologic study of 75 cases . Ann Afr Med. 2007 ; 6 : 54 - 7 .
25. Mudenda S , Kaile T , Simakando M , Ngwira Z , Peter J , Marimo C , et al. https ://www.johronline.com/issue/20161229- 172205 .461.pdf. Accessed 18 Aug 2018 .
26. Jaka H , Mshana SE , Rambau PF , Masalu N , Chalya PL , Kalluvya SE . Hepatocellular carcinoma: clinicopathological profile and challenges of management in a resource-limited setting . World J Surg. Oncol . 2014 ; 12 : 246 .
27. Vhriterhire RA , Ngbea JA , Ojo BA , Jegede OO , Manasseh AN , Ayuba MD , et al. Hepatocellular carcinoma sub-types in North-Central Nigeria: a histological review of liver biopsies . http://www.sciencedomain.org/ abstract/12165. 2016 . http://imsear.hellis.org/handle/123456789/182207. Accessed 17 Aug 2017 .
28. Khan M , Haq SA , Ahmed N , Matin MA . Etiology and clinical profile of hepatocellular carcinoma in Bangladesh . Bangladesh Med Res Counc Bull . 1997 ; 23 : 16 - 24 .
29. Steiner PE , Davies JNP . Cirrhosis and primary liver carcinoma in Uganda Africans . Br J Cancer . 1957 ; 11 : 523 - 34 .
30. Steiner PE , Camain R , Netik J. Observations on cirrhosis and liver cancer at Dakar, French West Africa . Cancer Res . 1959 ; 19 : 567 - 80 .
31. Pavlica D , Samuel I . Primary carcinoma of the liver in Ethiopia. A study of 38 cases proved at post-mortem examination . Br J Cancer . 1970 ; 24 : 22 - 9 .
32. Kruger FC , Daniels C , Kidd M , Swart G , Brundyn K , Van Rensburg C , et al. Non-alcoholic fatty liver disease (NAFLD) in the Western Cape: a descriptive analysis . South Afr Med J. 2010 ; 100 : 168 - 71 .
33. Onyekwere CA , Ogbera AO , Balogun BO . Non-alcoholic fatty liver disease and the metabolic syndrome in an urban hospital serving an African community . Ann Hepatol. 2011 ; 10 : 119 - 24 .
34. Ntagirabiri R , Cikomola J , Baransaka E , Ndirahisha E , Ndabaneze E , Niyongabo T. Stéatose hépatique lors du syndrome métabolique chez l'adulte noir africain : cas du Burundi . J Afr Hépato-Gastroentérologie . 2014 ; 8 : 195 - 9 . https://link.springer.com/article/10.1007/s12157-014 -0553-9. Accessed 22 Feb 2018 .
35. Steyn NP , McHiza ZJ . Obesity and the nutrition transition in Sub-Saharan Africa . Ann N Y Acad Sci . 2014 ; 1311 : 88 - 101 .
36. McGlynn KA , Petrick JL , London WT. Global epidemiology of hepatocellular carcinoma: an emphasis on demographic and regional variability . Clin Liver Dis . 2015 ; 19 : 223 - 38 .