Tuesday 16 March 2004
Definition: Hepatoblastoma is a malignant embryonal liver tumor that occurs almost exclusively in infants and very young children.
Hepatoblastoma is composed of epithelial and mesenchymal elements in varying proportions and at various stages of differentiation. The epithelial element recapitulates the stages of hepatocyte development from the primitive blastema through embryonal hepatocytes to fetal hepatocytes. The blastemal or undifferentiated cells have been postulated to represent neoplastic hepatocyte progenitor cells.
world-wide incidence of 0.5-1.5 cases per million children
60 and 85% of all hepatic tumors in children
most common type of pediatric liver tumor
The incidence rate of primary hepatic malignancies in children 0 to 14 years of age is approximately 0.2 per 100,000 chil-dren in the United States, with hepatoblastomas accounting for 47% of the malignancies and nearly 27% of all pediatric hepatic tumors.
By age group, hepatoblastoma accounts for 1% of all pediatric malignancies in children under 15 years age, 1.5% of all malignancies in children younger than 5 years of age, and 3.3% of all malignancies in white and black children under 1 year of age.
The reported incidence is 11.2 cases per million during the 1st year of life; nearly 90% of hepatoblastomas are seen in the ﬁrst 5 years of life, with 68% discovered in the ﬁ rst 2 years and 4% present at the time of birth.
Of 271 primary hepatic malignancies reported in the United States to Surveillance, Epidemiology and End Results (SEER) data between 1973 and 1997 in patients below 20 years of age, 67% and 31% were HB and HCC, respectively.
In the group less than 5 years of age, HB accounted for 91%, whereas among those 15 to 19 years of age, HCC represented 87% of the cases.
The relative frequency of hepatoblastomas in younger children is most apparent when noting that hepatoblastomas account for over 40% of all hepatic tumors (benign and malignant) in children younger than 2 years of age, but only 7.5% of liver tumors in children 5 to 20 years old.
Althougthere is no racial predilection for hepatoblastoma, there is a distinct male predominance from 1.2:1 to 3.6:1.
An increased incidence of hepatoblastoma (from 0.4 to 1.0 per million between 1971 and 1983) has been observed at a Children’s Tumour Registry in Manchester, United Kingdom.
The US National Cancer Institute Surveillance, Epidemiology, and End Result (SEER) program includes approximately 14% of the population; it revealed an average annual increase of 5.2% in the incidence of hepatoblastoma from 1973 to 1992.3 This change might be explained by hepatoblastoma occurring in surviving premature infants.
Hepatoblastomas in Japan accounted for 58% of all malignancies in children who weighed less than 1000 g at birth.
Further analysis of the Japanese Children’s Cancer Registry data revealed that 15 (5%) of 303 hepatoblastomas between 1985 to 1995 occurred in infants with history of prematurity and weight less than 1500 g at birth. This rate was greater than 10 times that for all live births.
The relative risk for hepatoblastoma for children who weighed less than 1000 g at birth was 15.64 compared with 2.53 for those 1000 g to 1499 g and 1.21 for 2000 g to 2499 g.
Of 77 children with hepatoblastoma in the German registry, 3 (4%) were premature infants who required parenteral nutrition, a treatment that has been lifesaving for many small premature infants but has been reported to lead to cirrhosis in many survivors.
It has not previously been associated with hepatoblastoma. The histologic features of hepatoblastoma following prematurity are indistinguishable from other hepatoblastomas.
Toxics and drugs
The Children’s Cancer Group has evaluated environmental or drug exposure. Seventyfive sets of parents of children with hepatoblastoma were compared with those of age-matched controls.
Before and during pregnancy, there was a significant excess of maternal exposure to metals used in welding and soldering, lubricating oils, and protective greases. Paternal exposure to metals was also greater.
At 23 weeks, a congenital hepatoblastoma was found in a stillborn fetus whose mother was an artist exposed to volatile hydrocarbons.
The presenting symptom of virtually all liver tumors in children is abdominal swelling secondary to hepatomegaly. When confronted with this symptom, it is useful to consider the age at which liver tumors tend to occur.
Exceptions are frequent, but age can serve as a guide when the presenting symptoms lack specificity.
In the Pediatric Oncology Group series from 1986 to 2002, 66% of hepatoblastomas were manifest by the second year, and 11% before 6 months of age.
Approximately 50% of those in infants were congenital, given their size when discovered by 2 to 3 months of age; 6% of hepatoblastomas occurred after age 5 years. Hepatocellular carcinomas have been observed as early as 6 months.
Seven examples of mixed hepatoblastomas and hepatocellular carcinomas have been observed at a mean age of 8.5 years; perinatally acquired hepatitis B virus was responsible in 3 instances.
bilobar involvement is seen in 20-30%
multicentric involvement in 15%
elevated serum alpha-fetoprotein (AFP) levels
distant metastases usually occur very late in advanced disease stages
- rare placental localization in fetal forms (#9724342#)
Hepatoblastomas are single masses in approximately 80% of cases.
They occur in the right lobe in 58% of cases, in the left in 15%, and in both lobes in the remaining 27%, either as a large single lesion extending across the midline or as multiple lesions.
Distant metastasis are present in 20% of patients at the time of diagnosis, with the lung as the most common site of metastasis; other common sites are the brain and bone and metastasis occurs more commonly with disease relapse.
Tumors may measure 15 cm or more in diameter and weigh in excess of 1,000 g. Grossly, they are coarsely lobulated and frequently bulge from the surface of the liver.
On cut section, the lesions are tan to light brown to green and display frequent areas of hemorrhageand necrosis. Various types of mesenchymal tissues (e.g., osteoid, cartilaginous, ﬁbrous) in the mixed type of hepatoblastoma may alter the color and consistency of the gross appearance.
Evaluation of margins
The evaluation of margins for total or partial hepatectomy specimens depends on the method and extent of resection. It is recommended that the surgeon be consulted to determine the critical foci within the margins that require microscopic evaluation.
The transection margin of a partial hepatectomy may be large, rendering it impractical for complete examination. In this setting, grossly positive margins should be microscopically confirmed and documented.
If the margins are grossly free of tumor, judicious sampling of the cut surface in the region closest to the nearest identified tumor nodule is indicated.
In selected cases, adequate random sampling of the cut surface may be sufficient. If the neoplasm is found near the surgical margin, the distance from the margin should be reported.
For multiple tumors, the distance from the nearest tumor should be reported.
Histologically, the tumor is traditionally classiﬁed into six patterns.
The epithelial types account for approximately 56% of cases, including pure fetal (31%), embryonal (19%), macrotrabecular (3%), and small cell undifferentiated (3%).
The mixed pattern of epithelial and mesenchymal components accounts for 44% of the cases, including 34% without teratoid features and 10% with such components as squamous epithelium and striated muscle.
The fetal pattern refers to cases in which 100% of the tumor is composed of small, round, uniform cells with abundant cyto-plasm and distinct cytoplasmic membranes. The cells are arranged into thin trabeculae, usually two to three cells thick, with alternating light and dark areas.
The embryonal pattern refers to cases of epithelial hepatoblastoma in which, in addition to fetal cells, part of the tumor has cells arranged into sheets of irregular, angulated cells with a high nucleocytoplasmic ratio, increased nuclear chromatin, and indistinct cytoplasmic membranes.
Pseudorosette and acinar formation are common features.
Foci of extramedullary hematopoiesis (EMH) are seen in both the fetal and embryonal areas.
epithelial hepatoblastoma (56%)
- well differentiated fetal hepatoblastoma (31%)
- hepatoblastoma, epithelial type, purely fetal pattern (mitotically inactive)
- hepatoblastoma, epithelial type, purely fetal pattern (mitotically active)
- embryonal and fetal hepatoblastoma (19%)
- hepatoblastoma, epithelial type, macrotrabecular pattern (3%)
- focal or diffuse small undifferentiated cells in hepatoblastoma (3%) (#11753992#)
- undifferentiated small cell hepatoblastoma (#1384017#)
mixed epithelial and mesenchymal hepatoblastoma (44%)
- hepatoblastoma, mixed epithelial and mesenchymal type, without teratoid features
- hepatoblastoma, mixed epithelial and mesenchymal type, with teratoid features
- hepatoblastoma, rhabdoid type
- hepatoblastoma, other subtype
Approximately 56% of tumors are of the epithelial type, which is subclassified further as pure fetal (31%), embryonal (19%), macrotrabecular (3%), and small-cell undifferentiated (anaplastic; 3%).
Approximately 44% of tumors contain both mixed epithelial and mesenchymal components.
There is no relationship between the age of the child and predominant cell type in hepatoblastoma.
Of all cases at all ages, 85% to 90% contain both fetal and embryonal derivatives in variable proportions; 20% have stromal derivatives.
Because these histologic types tend to be randomly intermingled, both fine-needle aspiration and biopsies may capture a nonrepresentative sample of tumor.
Mesenchymal elements may consist of osteoid, cartilage, or other spindle cells.
Components and patterns
"small cell undifferentiated" or "anaplastic" pattern
The small cell undifferentiated or anaplastic pattern is composed of cells reminescent of neuroblastoma or other small round blue cell tumors, with scanty cytoplasm and hyperchromatic nuclei.
Round or ovoid cells predominate, with the occasional presence of spindle or stellate cells within a mucoid matrix. These cells grow in sheets but lack cohesiveness. Mitoses are occasionally present, but the cells do not produce glycogen, fat droplets, or bile pigment.
Abortive or incompletely formed bile ductules may be present, but electron microscopy or immunohistochemical studies, or both, may be needed to conﬁrm the diagnosis of hepatoblastoma.
Particularly helpful in establishing the diagnosis is the presence of cytoplasmic staining with polyclonal anticytokeratin antibodies.
Gonzalez-Crussi believes that the small cell form represents the subtype with the least differentiation within the highly variable morphologic spectrum of hepatoblastomas.
Medium- and large-sized cells have been reported in undifferentiated hepatoblastomas; some undifferentiated tumors have shown intermediate or large cells, leading to a proposal to subclassify undifferentiated hepatoblastomas as "small cell", "intermediate cell", and "large cell" subtypes.
Immunostains are required to differentiate these from other large undifferentiated cell tumors such as lymphoma, large cell medulloblastoma, large cell neuroblastoma, and Ewing sarcoma family tumors, although some hepatoblastomas may be positive for CD99.
The histogenesis of these undifferentiated tumors is not known. A pathway might involve regression to a primitive cell lineage of the hepatogenic foregut endoderm. These tumors may show loss of INI1. The mixed epithelial and mesenchymal type of tumor contains cells admixed with primitive mesenchyme and various mesenchymally derived tissues.
- Primitive mesenchyme
The highly cellular primitive mesenchyme consists of elongated, spindle-shaped cells with a scanty cytoplasm, and elongated pump nuclei with rounded ends, resembling ﬁbroblastoid/myoﬁ broblastoid tissue.
Some areas may display parallel orientation of cells with deﬁnite collagen ﬁbers and young ﬁbroblasts; other areas may have more loosely arranged cells leading to a myxomatous appearance.
Mature ﬁbrous septa are also seen, along with areas of osteoid and cartilaginous tissue.
Cells within the osteoid foci have an irregular, angular outline and short processes that make them indistinguishable from osteoblasts. Immunohistochemical studies, however, have identiﬁed this osteoid-like material as being produced though a process of epithelial differentiation. Osteoid stromal component is reportedly more prominent following chemotherapy.
The prognostic signiﬁcance of these stromal elements is unclear with studies reporting both improved survival or no effect on survival.
Tumors that show mainly mesenchymal/stromal tissue with apparent lack of an epithelial component have been termed “pediatric hepatic stromal tumors” and have been proposed to resemble similar lesions described in childhood kidney cancer, that is, metanephric stromal tumor/MST.
- Other components
Approximately 20% of the mixed types of hepatoblastomas contain a variety of tissues, including stratiﬁed squamous epithelium, melanin pigment, mucinous epithelium, cartilage, bone, and striated muscle in addition to the epithelial cells, ﬁbrous tissue, and osteoid-like material.
These tumors have been termed "teratoid hepatoblastomas" by.
In view of the histologic heterogeneity of hepatoblastomas, rare tumors with unique morphologies may be seen that are difﬁcult to classify into one of the above categories.
This has led to anecdotal descriptions of “new” variants of hepatoblastomas including:
mucoid anaplastic hepatoblastoma,
hepatoblastomas with endocrine/neuroendocrine differentiation,
hamartoma-like hepatoblastoma/ hepatoblastoma with organoid conﬁguration
It has also led to the allocation of problematic cases into a neutral category, such as "hepatoblastoma, not otherwise speciﬁed".
The macrotrabecular pattern refers to cases in which trabeculae more than 10 cells in thickness are present as a repetitive pattern within the tumor.
The large trabeculae contain either fetal- or embryonal-type cells; a third, larger cell type with cytoplasm that is more abundant than in normal hepatocytes or fetal-type cells; or a combination of all three cell types.
Thus, the term “macrotrabecular” refers more to a growth pattern rather than a distinct subtype and these lesions form a heterogeneous group.
Cases that have embryonal or mesenchymal cells with an isolated macrotrabecular focus are classiﬁed based on the embryonal or mesenchymal cell present and not as macrotrabecular.
Tumors where the third (hepatocyte-like) cell type predominates may be very difﬁcult to distinguish from HCC. The presence of EMH is useful in a diagnosis of hepatoblastoma.
epithelial fetal hepatoblastoma
In completely resected tumors, a pure fetal appearance confers a better prognosis, whereas a small cell undifferentiated or anaplasia is associated with a poor prognosis.
Distinguishing well-differentiated (mitotically inactive) fetal hepatocytic tumor cells from normal liver in an infant can be difficult.
The fetal tumor cells are larger than normal fetal hepatocytes and have a higher nuclear cytoplasmic ratio.
The nuclei are regular and round with little discernible mitotic activity (<2 mitoses per 10 high-power [X40 objective] fields) in the well-differentiated variety.
Fetal tumor cells grow in cords, as in normal liver, or nests or nodules. Clusters of normoblasts (extramedullary hematopoiesis) are commonly seen, as in fetal liver. The cytoplasm of the fetal tumor cells varies from eosinophilic to clear, depending on the amount of glycogen content. Fetal tumor cells may also contain abundant lipid, producing vacuolization. In well-differentiated fetal hepatoblastoma, bile secretion may be observed.
Well-differentiated (mitotically inactive) fetal histology was superior to embryonal differentiation in long-term survival; therefore, the COG study is treating stage I well-differentiated fetal hepatoblastoma (with low mitotic rate) with
epithelial fetal and embryonal hepatoblastoma
The embryonal cellular component of hepatoblastoma is less well-differentiated than its fetal hepatocytic counterpart, with cells that are small and have a high nuclear cytoplasmic ratio with ovoid nuclei and that may assume a tubular or rosette-like configuration. Purely embryonal tumors are almost never encountered and invariably show some fetal areas.
mixed hepatoblastoma (epithelial and mesenchymatous hepatoblastoma)
Often, mixed hepatoblastomas contain epithelial membrane antigen (EMA)-positive nests of squamous epithelium. The osteoid component of mixed hepatoblastomas is found to be a matrix of collagen surrounding cells expressing EMA and having ultrastructural features of epithelium rather than osteoblasts.
Hepatoblastomas may contain other stromal derivatives, including cartilage and rhabdomyoblasts. There is no prognostic significance to the presence of mixed histologic features.
A macrotrabecular pattern of hepatoblastoma growth is one in which fetal or embryonal cells numbered 20 or more within a cord or cluster, as opposed to the usual 2- to 6-cellthick cords or plates.
anaplastic hepatoblastoma and undifferentiated hepatoblastoma
An urgent research need is to identify more effective medical therapy for both small cell undifferentiated hepatoblastoma and rhabdoid hepatoblastoma, the most aggressive forms of this malignancy.
When first distinguished from embryonal epithelium, small undifferentiated cells in hepatoblastoma were noted to resemble neuroblastoma, to have a low mitotic rate, and were called "anaplastic", consistent with the dictionary definition, characterized by imperfect development.
Since “anaplastic” was redefined by Faria et al20 for Wilms tumor as nuclear enlargement to 3 times those of typical tumor cells, hyperchromasia and atypical mitoses, the small cell undifferentiated component no longer is designated as anaplastic.
Beckwith-type anaplasia does occur rarely in hepatoblastoma, and its
significance is unknown. The small cells have been considererd a putative hepatic progenitor cells on the basis of immunohistochemical and electron microscopic studies.
When present in a significant fraction of the hepatoblastoma (75%), or as the sole cell type, the small cell type is typically found in infants younger than 1 year; they have a poor prognosis, with poor response to current therapy.
The prognostic significance of smaller proportions of the small cell undifferentiated type is still undetermined.
Rhabdoid cells have the characteristic eccentric pink cytoplasmic inclusions (periodic acid-Schiff/diastase positive, vimentin or cytokeratin positive) with vesicular nuclei and fibrillar inclusion bodies by electron microscopy.
They may be associated with the small cell component in otherwise typical hepatoblastomas or as the exclusive cell type, in which case they occur in infancy and are associated with a poor prognosis.
multinucleated tumor giant cells
Multinucleated tumor giant cells are found in rare hepatoblastomas, sometimes associated with HCG production and clinical virilization.
Teratoid hepatoblastoma was initially depicted as having intestinal, neural, and melanocytic elements. These are distinguished from true teratomas, which can also occur in the livers of children, on the basis of organoid differentiation and even greater diversity of tissue elements in the teratomas.
Prognostic impact of microscopy
The prognostic impact of histology has been analyzed in a few studies.
Of the ﬁve histologic subtypes (pure fetal, embryonal, mixed epithelial-mesenchymal, macrotrabecular, and small cell undifferentiated), the fetal subtype carries the most favorable prognosis, and small cell undifferentiated the worst.
In general, pure fetal histology is associated with an improved prognosis, while undifferentiated histology is associated with a poor prognosis, with macrotrabecular histology probably having an intermediate prognosis.
In a study of 168 patients, the estimated 24-month survival probability was 50% with the macrotrabecular type in comparison with 92%, 63%, and 0% for the purely foetal, embryonal, and SCUD histologies, respectively.
Small cell undifferentiated histology predicts an increased risk of relapse.
Even a focal (partial or predominant) expression of small cell histology in completely resected HBL may have an unfavour-able effect on outcome, drawing a corollary with focal versus diffuse anaplasia in nephroblastoma.
Small cell undifferentiated tumors appear to be biologically different from tumors with non–small cell histology and have been reported to be similar to rhabdoid tumors at the immunohistochemical (INI1 negative), cytogenetic and molecular level and in terms of their adverse adverse outcomes.
Among the mixed epithelial/mesenchymal type tumors, the presence of mesenchymal elements may be associated with improved prognosis.
Although open biopsy and needle biopsy often are adequate in establishing the diagnosis, the use of FNA may prove difﬁcult, particularly in cases of small cell undifferentiated or embryonal epithelial lesions.
FNA has been reported to be accurate in diagnosing hepatoblastoma in approximately 65% of cases, primarily fetal epithelial and mixed hepatoblastomas.
The diagnosis was most frequently confused with metastatic tumor including Wilms tumor, neuroblastoma, and rhabdomyosarcoma.
Weir et al. have reported the cytologic features of hepatoblastoma in serous cavity ﬂuids. All six specimens examined showed hypercellular smears in a relatively clean background.
Mixed embryonal and fetal subtypes of HBL disclosed three-dimensional clusters of neoplastic cells that formed straight or branched cords and acinus-like struc-tures.
The cells were moderately pleomorphic, had high nuclear-to-cytoplasmic ratios, rare intranuclear inclusions,and numerous mitoses.
The small cell subtype showed tight clusters of small, round, primitive cells with hyperchromatic nuclei, high N/C ratios, and prominent nuclear molding.
In addition, there were numerous single cells with naked nuclei, often in an Indian-ﬁle conﬁguration. Bile pigment, osteoid, and other mesenchymal components were absent in all specimens.
Postchemotherapy resection specimens often show eradication of the embryonal cells and more prevalent osteoid-like foci.
Heifetz et al. reported that vascular invasion, amount of mesenchyme, persistence of embryonal epithelium, extent of tumor necrosis, and mitotic activity of the epithelial component have predictive value in this type of specimen.
This has yet to be confirmed, but the items should be documented, as should the presence of any small undifferentiated cells, which are known to negatively affect prognosis but may have been missed in the initial biopsies of stage III and IV lesions.
Histologic examination of a regional lymphadenectomy specimen usually involves examination of 3 or more lymph nodes.
The regional lymph nodes of the hepatic region include the hilar, hepatoduodenal ligament, and caval lymph nodes.
Nodal involvement of the inferior phrenic lymph nodes or other lymph nodes distal to the hilar, hepatoduodenal ligament, and caval lymph nodes are considered as distant metastasis.
- epithelial, purely fetal hepatoblastoma, mitotically inactive with 2 or fewer mitoses in 10, X40 objective
Less favorable grade
- others type except unfavorable, below.
- May be more favorable if stage I (usually treated with multimodality therapy).
- small cell undifferentiated hepatoblastoma, predominant or sole histopathologic subtype, at any stage or therapy
- rhabdoid hepatoblastoma, predominant or sole histopathologic subtype, at any stage or therapy
Tumors with favorable histopathologic features are purely fetal, well-differentiated
lesions defined as mitotically inactive with a minimal mitotic rate of 2 or fewer mitoses per 10, X40 objective fields. These tumors are also stage I and are usually treated with surgery alone.
Tumors with unfavorable histopathologic features have either undifferentiated small cell or rhabdoid subtypes or both.
When present in a significant fraction of the hepatoblastoma (75%) or as the sole cell type, the small cell, undifferentiated subtype is typically found in infants younger than 1 year with poor prognosis regardless of stage or therapy. When this subtype is present in lesser proportions, the prognostic implications
When the rhabdoid cell type is the exclusive histopathology, it is
also found typically in infants with poor prognosis.
Less favorable histopathologies include all other tumor subtypes not mentioned above, although they may be associated with favorable prognosis if stage I, and are usually treated with multimodal therapy.
Necrosis after neo-adjuvant chemotherapy
Extent of tumor necrosis following neo-adjuvant chemotherapy is an independent prognostic factor in patients with newly diagnosed HB. Histological response may potentially be used in strategies to modify post-surgical therapy to improve survival in HB. (#22190448#)
Staging in the United States combines imaging with surgical judgment about
Computed tomography and magnetic resonance imaging are used
exclusively in the SIOPEL (Societé Internationale D’Oncologie Pediatrique Liver Tumor Study Group) protocol to determine the location and extent of hepatic involvement of hepatoblastoma preoperatively; tumors sparing the left medial and right anterior sectors are primarily resected.
Dissemination of hepatic malignancies occurs within portal veins and follows the
expected ready access of infiltration into hepatic veins, with frequent lung involvement.
Further spread to the brain may occur. Hilar lymph node metastases are relatively
infrequent, but capsular rupture of subcapsular masses either before or during surgery can upstage an otherwise resectable malignancy.
The Children’s Oncology Group staging system is recommended for hepatoblastomas.
Stage I (favorable histologic type) tumors are completely resected and have typical
histologic features of a purely fetal well-differentiated histologic pattern (minimal
mitotic index of 2 mitoses per 10 high-power [X40 objective] fields).
Stage I (other histologic type) tumors are completely resected, with a histologic picture other than purely fetal, well-differentiated pattern.
Stage II tumors are grossly resected with evidence of microscopic residual tumor. Such tumors are rare, and patients with this stage have not fared differently from those with stage I tumors in previous protocols. Resected tumors with preoperative (intraoperative) rupture are classified stage II.
Stage III (unresectable) tumors are those that are considered by the attending surgeon not to be resectable without undue risk to the patient. These include partially resected tumors with measurable tumor left behind. They do not include grossly resected tumors with microscopic disease at the margins or resected tumors with preoperative/intraoperative rupture. Lymph node involvement is considered stage III disease and may require evaluation with second laparotomy after initial 4 courses of chemotherapy.
Stage IV tumors are those that present with measurable metastatic disease to the lungs or other organ. Nodal involvement of the inferior phrenic lymph nodes or other lymph nodes distal to the hilar, hepatoduodenal ligament, or caval lymph nodes are considered as distant metastases.
Resectability is the key prognostic feature for all liver malignancies, with the possible exception of rhabdomyosarcoma (see separate College of American Pathologists protocol for rhabdomyosarcoma).
Unfortunately 67% of hepatoblastomas were not amenable to primary surgery (48% stage III and 19% stage IV) in the 16 years of Pediatric Oncology Group/Children’s Oncology Group accessions.
placental involvement in congenital hepatoblastoma (#9724342#)
The various patterns of hepatoblastoma display differing immunoreactivity, probably based on their degree of differentiation, with the fetal cell areas of an epithelial hepato-blastoma staining positively for a broad range of epithelial markers and small cell undifferentiated hepatoblastomas showing positivity for only a few markers.
Ruck et al. noticed a correlation between the cytokeratin staining of normal biliary epithelium and liver parenchymal cells and the types of epithelial cells in hepatoblastoma, with CK19 more prominent in small cell and embryonal epithelial cell areas (and in biliary epithelium) and CK18 more prominent in fetal epithelial areas (and in normal hepatocytes).
Osteoid areas were positive for both CK18 and CK19, whereas spindle cells areas were not immunoreactive for any of the cytokeratins.
These characteristics have suggested to some authors that the primitive small cells give rise to embryonal hepatoblastoma cells and, after further maturation, fetal hepatoblastomas.
In a study of 12-needle core biopsies in proven hepatoblastomas, Ramsay et al. reported variable antigen expression with positivity for cytokeratins (10/12 cases), alpha-1-antitrypsin (5/12 cases) and AFP (7/12 cases), MIC-2 (CD99) (8/12 cases), NCAM (CD56) (4/12 cases), neuroblastoma marker NB84 (3/12), desmin (2/12 cases), BCL2 (2/12 cases), and one case each for vimentin, NSE, and PGP 9.5.
However, all tumors were negative for CD45, WT1, and S-100.
The authors concluded that hepatoblastoma shows no distinct immunohistochemical proﬁle, and the diagnosis requires a combination of the clinical, imaging, and pathologic ﬁndings, since they can express antigens normally seen in other childhood malignancies.
Insulin-like growth factor 2 and insulin-like growth factor–binding protein expression have been noted in 11 hepatoblastomas, with their expression inversely correlated with the degree of tumor cell differentiation.
Akmal et al. suggest that these markers may be used as an assessment of the degree of differentiation of the tumor.
Interestingly, hypoglycemia has been noted as a rare presenting symptom of hepatoblastoma with the hypoglycemia disappearing after removal of the tumor.
Glypican-3 (GPC3), a heparin sulfate proteoglycan bound to the cell surface, is overexpressed both at the genomic (by microarray studies) and at the protein (by IHC) levels in hepatoblastoma.
In their series, Zynger et al. found that 65 of 65 hepatoblastomas had cytoplasmic immunoreactivity for GPC3 with greater than 90% of cases showing strong, diffuse positivity.
There was no reactivity in benign liver tissue.
Fetal, embryonal, and small cell undifferentiated patterns were diffusely positive in almost all cases, whereas mesenchymal and teratoid patterns were nearly all negative.
Immunohistochemical studies have identiﬁed putative stem cells in hepatoblastomas.
Cells in atypical ducts were found to express simultaneously stem cell markers and hepatocytic or biliary lineage markers, suggesting a direct role for stem cells in the histogenesis of hepatoblastoma.
epithelial membrane antigen + (EMA)
+/- alpha 1-antitrypsin +
+/- ferritin +
+/- vimentin +
Most patients in the United States are staged post-operatively according to the Children’s Cancer Study Group (CCSG) classiﬁcation. Other classiﬁcations include the TNM or variations of the CCSG staging classiﬁcation.
Based on these classiﬁ cations, approximately 38% of hepatoblastomas are stage I at the time of initial diagnosis and before any chemotherapy is administered.
At this same point, about 9% are stage II, 24% stage III, and 29% stage IV.
However, this traditional staging system has been criticized for being rather subjective, depending to a large extent on the surgeon rather than the tumor.
In 1990, the International Society of Pediatric Oncology Liver Study Group (SIOPEL-1) adopted a new preoperative staging system, Pre-treatment Extent of Disease (PRETEXT), based exclusively on images obtained prior to surgery, based on the branching pattern of the portal vein, which divides the liver into eight segments.
The PRETEXT system divides the liver into four sectors:
(a) lateral sector (Couinaud segments 2 and 3);
(b) medial sector (segment 4);
(c) anterior sector (segments 5 and 8);
(d) posterior sector (segment 6 and 7).
Tumors are classiﬁ ed as one of four categories (PRETEXT-I to PRETEXT IV) by determining the number of affected liver sector(s) on imaging.
Extrahepatic growth of the tumor is indicated by adding a letter (V involvement of hepatic vein, P involvement of portal vein, E for extrahepatic extension, M for the presence of distant metastasis).
The PRETEXT system has prognostic value for overall and disease-free survival and is useful in deﬁning treatment.
Although the PRETEXT system was developed mainly to assess the efﬁcacy of neo-adjuvant chemotherapy and to predict surgical resectability, it also had highly prognostic value for both overall survival and event-free survival.
Conceptually, however, both pre-operative and post-operative staging systems use the same parameters for staging, namely, size, vascular invasion, extension and complexity of the primary tumor, and the absence or presence of metastases.
Metastatic spread of hepatoblastoma is seen most frequently to the lung but may also spread to bone, brain, eye, and ovaries. Local extension into hepatic vessels and the inferior vena cava may also occur.
Treatment and Outcomes
Although complete resection of hepatoblastoma is the mainstay of treatment and is the only chance of cure, improvements in survival that have occurred over the last three decades have been a function of standardized chemotherapy regimens that reduce tumor size and enable complete tumor excision, even permitting cure in the presence of initially unresectable or metastatic disease.
Treatment strategies currently combine surgery, transplantation, and chemotherapy (adjuvant and neoadjuvant), as deﬁned by the PRETEXT stage of the tumor.
Surgery remains the mainstay in the treatment of hepatoblastoma, with prognosis directly related to tumor stage.
Small, solitary lesions localized to a single lobe can be adequately treated by lobectomy.
Larger lesions, including those requiring preoperative chemotherapy to allow resection, may require more extensive surgery or transplan-tation.
Surgical complications, particularly hemorrhage, are noted in 14% of primary resections and 29% of second resections.
If resection is possible and safe, an attempt to obtain clear resection margins is essential; positive margins on pathology warrant a re-resection if possible.
Although elevated AFP immediately after resection is common, persistently elevated or rising AFP levels indicate the need to evaluate further for disease recurrence or search for distant metastasis.
Contra-indications to immediate resection include extensive bilateral liver involvement, presence of vascular invasion of major hepatic veins or inferior vena cava, diffuse multifocal disease, and distant metastasis.
At the time of diagnosis, 40% to 60% of hepatoblastomas are considered to be unresectable and 10% to 20% of patients are found to have pulmonary metastases.
Preoperative chemotherapy converts nearly 85% of these “unresectable” lesions to ones that can be entirely grossly removed, that is, to stage I or II lesions with subsequent long-term survival.
Even if unresectable at diagnosis, most hepatoblastomas are unifocal and chemosensitive, with cisplatin and adriamycin being the most commonly used agents.
Chemotherapy has been proven effective in both an adjuvant and neoadju-vant treatment and can shrink tumors.
It makes them less prone to bleed and delineates the tumor from the surrounding nor-mal parenchyma and vascular structures facilitating resection. In rare cases, patients may survive with chemotherapy alone.
On the other hand, some tumors may become resistant to prolonged courses of chemotherapy, and the highest survival rates have historically been observed in patients with initially resected tumors—although these tumors also tend to be the smaller more favorable tumors.
Further, prolonged (>4 cycles) courses of neoadjuvant chemotherapy are discouraged, since this may lead to cumulative chemotherapy toxicity and cause tumor cells to become resistant to chemotherapy.
The 5-year survival rate for hepatoblastomas has improved to over 75% compared with a 5-year survival rateof 35% almost 30 years ago (122).
Preoperative chemotherapy has increased resectability of hepatoblastoma from 40% to 60%, to 90%, with the more extensive tumors requiring transplantation to remove the involved portions of liver.
In fact, primary liver transplantation with neoadjuvant chemotherapy may result in an 80% 5- to10-year disease-free survival rate, whereas the 10-year survival falls to 40% in those undergoing transplantation as “rescue therapy”.
The SIOPEL-1 study showed that, in patients undergoing transplantation, only macroscopic venous inva-sion had a signiﬁcant prognostic effect on survival.
A worldwide electronic registry for liver transplant in childhood liver tumors (hepatoblastoma, HCC, and hemangio-endothelioma) has been established; this “pediatric liver unresectable tumor observatory (PLUTO)” registry (http://transplant.test.cineca.it/).
Traditionally, tumor stage at the time of initial resection has been the key prognostic factor in determining the survival of children and adults with hepatoblastoma.
Data from the recent COG study, 9,645, show 3-year event-free survival of 90% for stage I to II, 50% for stage III, and only 20% for stage IV.
However, as noted earlier, preoperative and postoperative chemotherapy and aggressive treatment of pulmonary and central nervous system metastases have signiﬁ cantly changed the survival rate in patients with stage IV tumors.
Survival is independent of histologic sub-type when adjusted for age, sex, and stage.
Only small cell undifferentiated hepatoblastoma may have a worse prognosis than others, but the small number of cases of this unusual type makes analysis uncertain.
More recently, two broad categories of risk stratiﬁcation have been advocated, namely standard risk and high risk.
Standard risk tumors are PRETEXT I, II, or III.
SIOPEL high-risk tumors are deﬁned as tumors involving all four hepatic sectors (PREVEXT IV), or any tumor with metastasis (m), ingrowth of the vena cava (v), ingrowth of the portal vein (p) or contiguous extrahepatic disease (e), and tumors that fail to express AFP with AFP less than 100 at diagnosis.
Predispostion - Association
- Down syndrome
- trisomy 18 (#9267879#, #9025831#)
- partial trisomy 9p syndrome (#15588861#)
11p15 region deregulation
- Beckwith-Wiedemann syndrome (BWS)
- BWS-associated hepatoblastoma (#16010495#, #14692643#)
- +/- opsoclonus-myoclonus
APC germline mutations
- familial adenomatous polyposis coli (FAP) (#2848134#)
- Gardner syndrome
- Goldenhar syndrome (oculoauriculovertebral dysplasia, absence of portal vein)
- Schinzel-Geidion syndrome
- Beckwith-Wiedemann syndrome
- Beckwith-Wiedemann syndrome with opsoclonus, myoclonus
- Budd-Chiari syndrome
- Prader-Willi syndrome
- Aicardi syndrome
- Sotos syndrome (#19914434#)
- Li-Fraumeni syndrome (TP53 germline mutation)
- Fanconi anemia (Fanconi disease) (#21138478#)
- tyrosinemia type 1 (#20547648#)
- heterotopic lung tissue
- horseshoe kidney
- persistent ductus arteriosus
- renal dysplasia
- right-sided diaphragmatic hernia
- umbilical hernia
- inguinal hernia
- absence of left adrenal gland
- bilateral talipes
- cleft palate
- macroglossia, dysplasia of ear lobes
- malrotation of colon
- meckel diverticulum
- pectum excavatum
- intrathoracic kidney
- single coronary artery
- duplicated ureters
- Absence of left adrenal gland
- bilateral talipes
- duplicated ureters
- dysplasia of ear lobes
- cleft palate
- fetal hydrops
- heterotopic lung tissue
- horseshoe kidney
- inguinal hernia
- intrathoracic kidney
- Meckel diverticulum
- persistent ductus arteriosus
- renal dysplasia
- right-sided diaphragmatic hernia
- single coronary artery
- umbilical hernia
- synchronous Wilms tumor
- heterozygous alpha-antitrypsin deficiency
- type 1a glycogen storage disease
- Glycogen storage disease types Ia, III, and IV
- Heterozygous alpha1-antitrypsin deficiency
- Isosexual precocity
- Total parenteral nutrition
- Very low birth weight
- HIV infection
- HBV infection
- sexual precocity
- Budd-Chiari syndrome
toxics and drugs
- maternal use of clomiphene and pergonal
- oral contraceptive
- alcohol embryopathy
Environmental / Other
- Alcohol embryopathy
- Human immunodeficiency virus or hepatitis B virus infection
- Maternal clomiphene citrate or Pergonal
- Oral contraceptive, mother
- Oral contraceptive, patient
- Synchronous Wilms tumor
- fetal hydrops
- very low birth weight
Neoadjuvant chemotherapy followed by resection has become the mainstay in the treatment of hepatoblastoma (HB). The changes after chemotherapy typically result in tumor necrosis and a fibrohistiocytic response. (#20118773#)
Besides treated HBs having characteristic necrosis and fibrohistiocytic response, two-thirds have areas of cytoarchitectural differentiation ("maturation") mimicking non-neoplastic liver, and a quarter have alterations mimicking hepatocellular carcinoma. (#20118773#)
Nuclear expression of beta-catenin and keratin profiles were useful in distinguishing residual tumor with "maturation" from non-neoplastic liver and therefore in the assessment of surgical margins. (#20118773#)
Statistical analysis reveals that larger pretreatment and posttreatment imaged tumor size, larger tumor size at pathologic examination, and vascular invasion are significant univariate predictors of metastatic disease, whereas pretreatment imaged tumor size and vascular invasion are also significant independent predictors (multivariate logistic regression analysis). (#20118773#)
Multifocality, greater post-treatment necrosis and hepatocellular carcinoma-like morphology were more often associated with metastatic disease, but did not reach statistical significance. (#20118773#)
There appears to be a genetic predisposition to hepatoblastomas, with an increased incidence in a setting of Beckwith-Wiedemann syndrome, hemihypertrophy and familial adenomatous polyposis (FAP).
The relative risk for the development of hepatoblastoma in Beckwith-Wiedemann syndrome is 22.80, while that for FAP is 12.20, suggesting a role for genetic aberrations of chromosomes 11 and 5, respectively, in the pathogenesis of hepatoblastoma.
Inactivation of the APC tumor-suppressor gene (found on chromosome 5) is found in 67% to 89% of sporadic hepatoblastoma. This gene is known to regulate beta-catenin and modulate the wnt signaling pathway, suggesting a role for this signaling pathway in the development of hepatoblastoma.
Additional biologic markers may include trisomies 2, 8, and 20 and translocation of the NOTCH2 gene on chromosome 1.
There is also an association of prematurity/low birth weight and hepatoblastoma, with a relative risk of up to 15.64 in patients weighing less than 1,000 g, compared with patients weighing 2,500 g.
In Japan, Ikeda et al. have noted an increasing incidence of hepatoblastoma in very low birth weight infants from 0.7% of patients with birth weights less than 1,500 g with tumors in 1985 to 1989 to 8.6% of patients with similar low birth weights in 1990 to 1993.
In the United States, Ross and Gurney have observed a similar increasing trend of 5.2% in hepatoblastoma incidence in children 4 years and younger during the most recent two decades, a period corresponding with improved survival for low birth weight children.
Hepatoblastoma has been described in association with trisomy 18, including some cases with abdominal wall defects.
Hepatoblastoma has been noted in a number of sibling pairs including identical male twins and two siblings with GSD type 1a.
There are no known environmental risk factors.
It is likely that these may be related to the so-called "nested epithelial and stromal tumors".
Notwithstanding the morphological differences in mixed HBs between epithelial components of any kind and the stromal components, there is evidence that both have a common lineage.
This is suggested by the observation that beta-catenin mutations visualized by nuclear reactivity occur in epithelial and mesenchymal components.
The pathogenic pathways causing the development of both epithelial and mesenchymal/stromal lineages within the same tumor are not yet known.
However, epithelial-mesenchymal transition or mesenchymal-to-epithelial transition has been hypothesized to play a role in pathogenesis.
See "Hepatoblastoma oncogenetics"
Finegold MJ. Hepatic tumors in childhood. In: Russo P, Ruchelli ED, Piccoli D, eds.
Pathology of Pediatric Gastrointestinal and Liver Disease. New York, NY:
Tumor necrosis predicts survival following neo-adjuvant chemotherapy for hepatoblastoma. Venkatramani R, Wang L, Malvar J, Dias D, Sposto R, Malogolowkin MH, Mascarenhas L. Pediatr Blood Cancer. 2011 Dec 20. PMID: #22190448#
Mutations of PTCH1, MLL2, and MLL3 are not frequent events in hepatoblastoma. Chavan RS, Patel KU, Roy A, Thompson PA, Chintagumpala M, Goss JA, Nuchtern JG, Finegold MJ, Parsons DW, López-Terrada DH. Pediatr Blood Cancer. 2011 Dec 19. PMID: #22183980#
Effects of neoadjuvant chemotherapy on hepatoblastoma: a morphologic and immunohistochemical study. Wang LL, Filippi RZ, Zurakowski D, Archibald T, Vargas SO, Voss SD, Shamberger RC, Davies K, Kozakewich H, Perez-Atayde AR. Am J Surg Pathol. 2010 Mar;34(3):287-99. PMID: #20118773#
Chemotherapy effects on hepatoblastoma. A histological study. Saxena R, Leake JL, Shafford EA, Davenport M, Mowat AP, Pritchard J, Mieli-Vergani G, Howard ER, Spitz L, Malone M, et al. Am J Surg Pathol. 1993 Dec;17(12):1266-71. PMID: #8238734#
Neuroendocrine differentiation in hepatoblastoma. An immunohistochemical investigation. Ruck P, Harms D, Kaiserling E. Am J Surg Pathol. 1990 Sep;14(9):847-55. PMID: #2167615#
Hepatoblastoma. A clinical and pathologic study of 54 cases. Lack EE, Neave C, Vawter GF. Am J Surg Pathol. 1982 Dec;6(8):693-705. PMID: #6301295#
Hepatoblastoma. Attempt at characterization of histologic subtypes. González-Crussi F, Upton MP, Maurer HS. Am J Surg Pathol. 1982 Oct;6(7):599-612. PMID: #6295193#
Cytogenetics and CGH
Cytogenetic and array comparative genomic hybridization analysis of a series of hepatoblastomas. Stejskalová E, Malis J, Snajdauf J, Pýcha K, Urbánková H, Bajciová V, Starý J, Kodet R, Jarosová M. Cancer Genet Cytogenet. 2009 Oct 15;194(2):82-7. PMID: #19781440#
Cytogenetic evaluation of a large series of hepatoblastomas: numerical abnormalities with recurring aberrations involving 1q12-q21. Tomlinson GE, Douglass EC, Pollock BH, Finegold MJ, Schneider NR. Genes Chromosomes Cancer. 2005 Oct;44(2):177-84. PMID: #15981236#
Cytogenetic findings in two new cases of hepatoblastoma. Ali W, Savasan S, Rabah R, Mohamed AN. Cancer Genet Cytogenet. 2002 Mar;133(2):179-82.PMID: #11943350#
Hu J, Wills M, Baker BA, Perlman EJ. Comparative genomic hybridization analysis of hepatoblastomas. Genes Chromosomes Cancer. 2000 Feb;27(2):196-201. PMID: #10612809#
Fletcher JA, Kozakewich HP, Pavelka K et al. Consistent cytogenetic aberrations in
hepatoblastoma: a common pathway of genetic alterations in embryonal liver and
skeletal muscle malignancies? Genes Chromosomes Cancer. 1991;3:37-43.
Steenman M, Tomlinson G, Westerveld A, Mannens M. Comparative genomic
hybridization analysis of hepatoblastomas: additional evidence for a genetic link
with Wilms tumor and rhabdomyosarcoma. Cytogenet Cell Genet. 1999;86:157-161.
Weber RG, Pietsch T, von Schweinitz D, Lichter P. Characterization of genomic
alterations in hepatoblastomas: a role for gains on chromosomes 8q and 20 as
predictors of poor outcome. Am J Pathol. 2000;157:571-578.
Germline APC mutations are not commonly seen in children with sporadic hepatoblastoma. Harvey J, Clark S, Hyer W, Hadzic N, Tomlinson I, Hinds R. J Pediatr Gastroenterol Nutr. 2008 Nov;47(5):675-7. PMID: #18955873#
The spectrum of APC mutations in children with hepatoblastoma from familial adenomatous polyposis kindreds. Hirschman BA, Pollock BH, Tomlinson GE. J Pediatr. 2005 Aug;147(2):263-6. PMID: #16126064#
Jeng YM, Wu MZ, Mao TL, Chang MH, Hsu HC. Somatic mutations of beta-catenin play a crucial role in the tumorigenesis of sporadic hepatoblastoma. Cancer Lett. 2000;152:45-51.
Koch A, Denkhaus D, Albrecht S, Leuschner I, von Schweinitz D, Pietsch T.
Childhood hepatoblastomas frequently carry a mutated degradation targeting box of the beta-catenin gene. Cancer Res. 1999;59:269-273.
MT1G hypermethylation: a potential prognostic marker for hepatoblastoma. Sakamoto LH, de Camargo B, Cajaiba M, Soares FA, Vettore AL. Pediatr Res. 2009 Dec 21. PMID: #20032811#
Signature genes for both hepatoblastoma and hepatocellular carcinoma. Chen F, Li S, Castranova V. Eur J Gastroenterol Hepatol. 2009 Oct;21(10):1220-2. PMID: #19749507#
Altered microRNA Expression Patterns in Hepatoblastoma Patients. Magrelli A, Azzalin G, Salvatore M, Viganotti M, Tosto F, Colombo T, Devito R, Di Masi A, Antoccia A, Lorenzetti S, Maranghi F, Mantovani A, Tanzarella C, Macino G, Taruscio D. Transl Oncol. 2009 Aug 18;2(3):157-63. PMID: #19701500#
Small cell undifferentiated histology in hepatoblastoma may be unfavorable. Haas JE, Feusner JH, Finegold MJ. Cancer. 2001 Dec 15;92(12):3130-4. PMID: #11753992#
Undifferentiated small cell hepatoblastoma with a unique chromosomal translocation: a case report. Hansen K, Bagtas J, Mark HF, Homans A, Singer DB. Pediatr Pathol. 1992;12:457-462.
Weinberg AG, Finegold MJ. Primary hepatic tumors of childhood. Hum Pathol. 1983 Jun;14(6):512-37. PMID: #6303939#
Lack EE, Neave C, Vawter GF. Hepatoblastoma. A clinical and pathologic study of 54 cases. Am J Surg Pathol. 1982 Dec;6(8):693-705. PMID: #6301295#
Gonzalez-Crussi F, Upton MP, Maurer HS. Hepatoblastoma. Attempt at characterization of histologic subtypes. Am J Surg Pathol. 1982 Oct;6(7):599-612. PMID: #6295193#
Schnater JM, Aronson DC, Plaschkes J et al. Surgical view of the treatment of patients with hepatoblastoma: results from the first prospective trial of the International Society of Pediatric Oncology Liver Tumor Study Group. Cancer. 2002;94:1111-1120.