Overview of hepatoblastoma

INTRODUCTION — Hepatoblastoma is a rare tumor that arises almost exclusively in young children. This topic will provide an overview of the epidemiology, clinical manifestations, diagnosis, and management of hepatoblastoma. The pathology of liver tumors, including hepatoblastoma, is discussed separately. (See "Pathology of malignant liver tumors".)

EPIDEMIOLOGY AND RISK FACTORS — Hepatoblastoma is a rare tumor overall, but it is the most common primary malignant hepatic neoplasm in children, comprising two-thirds of primary liver tumors in this population [1]. While rare overall, the incidence of hepatoblastoma has been increasing worldwide [2,3].

Ninety percent of the cases occur in children less than five years old, primarily before the age of two [4]. In Western countries, the incidence in patients <15 years old is 0.5 to 1.5 per 1 million, making up 1 percent of all pediatric cancers [5,6]. There seems to be a slight predilection for males, with some studies citing a male-to-female ratio of 1.5:1 [7]. Adult cases are very rare; as of 2016, only 68 had been reported in the literature [8].

For yet undetermined reasons, an increased incidence of hepatoblastoma is seen in children born prematurely. The diagnosis of a liver mass presenting in the first several months of life may be a little more challenging as the alpha-fetoprotein (AFP) level may be physiologically elevated and will not accurately reflect the diagnosis and burden of disease. (See 'Laboratory evaluation' below.)

In children and young adults, hepatoblastoma is almost never associated with chronic liver disease [9,10]. Most cases are sporadic. However, in children, there are some inherited syndromes with an increased incidence of hepatoblastoma, which include Beckwith-Wiedemann syndrome, trisomy 18, trisomy 21, Aicardi syndrome, Li-Fraumeni syndrome, Goldenhar syndrome, type 1a glycogen storage disease (von Gierke disease), and familial adenomatous polyposis. For some of these conditions, screening for hepatoblastoma is recommended for affected individuals, which typically involves an abdominal ultrasound and an AFP level every three months until six to eight years of age depending on the underlying condition [11-13]. (See "Beckwith-Wiedemann syndrome", section on 'Neoplasia' and "Clinical manifestations and diagnosis of familial adenomatous polyposis" and "Congenital cytogenetic abnormalities", section on 'Trisomy 18 syndrome' and "Congenital cytogenetic abnormalities", section on 'Trisomy 21 (Down syndrome)' and "Syndromes with craniofacial abnormalities" and "Li-Fraumeni syndrome".)

DEVELOPMENTAL AND MOLECULAR ASPECTS OF HISTOGENESIS — Hepatoblastomas are embryonal tumors that are thought to arise from a primitive hepatocyte precursor cell that has the potential to differentiate along several lines [14]. The following evidence supports an origin from a pluripotent cancer stem cell:

●Hepatoblastoma mimics the developing fetal and embryonal liver, and it is composed of a variety of cell types, giving rise to a morphologically complex tumor that may include fetal and/or embryonal hepatocytes and heterologous tissues, including cartilage, bone, striated muscle fibers, and squamous epithelium. (See 'Histology' below.)

●Most cases arise in infancy and childhood, some are associated with genetic syndromes favoring tissue overgrowth and tumor development, and there is typically the absence of underlying liver disease. (See 'Epidemiology and risk factors' above.)

●Hepatoblastomas can be divided into molecular subclasses by liver differentiation stages that rely entirely on tumor transcriptional profiling. In a study using gene expression profiles and array comparative genomic hybridization, the gene signatures of the fetal and embryonal subtypes of hepatoblastoma differed and were found to mirror the molecular signature of late versus early liver development [15].

●Hepatic progenitor cell marker expression is upregulated in hepatoblastoma [16,17].

●Molecular overlap and the coexistence of hepatoblastoma and hepatocellular cancer, particularly in the rare adult case, support the emerging hypothesis that these two tumors may derive from a common progenitor cell, but perhaps at different stages of liver development.

●Sporadic cases are tightly linked to excessive signaling through the Wnt/beta-catenin signaling pathway, which may be driven by somatic mutations in the beta-1 catenin gene (CTNNB1) [15,16,18-23].

CLINICAL PRESENTATION — In children, hepatoblastoma has no specific presentation and often occurs as an asymptomatic mass in the liver but may occur with nonspecific symptoms such as abdominal pain, nausea, or vomiting; for unclear reasons, they are more often located in the right lobe. It classically arises within a healthy liver, unaffected by underlying disease. Large tumors at diagnosis may be complicated by rupture and hemorrhage. Jaundice is not seen at diagnosis, as liver function is typically normal. Similarly, coagulopathy is unusual at diagnosis.

Sexual precocity may be present due to the synthesis of ectopic beta-human chorionic gonadotropin [24]. Hepatoblastoma may also be associated with severe osteopenia and pathologic fractures in children due to abnormal calcium metabolism [25], but actual bone metastases are extremely rare. Rarely, ectopic production of adrenocorticotropic hormone and/or ectopic production of parathyroid hormone-related protein may lead to Cushing syndrome and hypercalcemia, respectively [26] (see "Epidemiology, pathophysiology, and causes of gynecomastia", section on 'Other rare causes'). Among children presenting with localized disease, approximately two-thirds are initially unresectable [27], and up to 20 percent will have distant metastasis at the time of diagnosis [28]. The most common site of metastasis is the lung, while lymph node involvement is rare [29].

DIAGNOSIS AND INITIAL EVALUATION — The majority of hepatoblastomas are first identified by imaging, but tissue is typically obtained for formal histopathologic diagnosis. Fine needle aspiration or core needle biopsies are the most common methods of tissue sampling, and they can accurately diagnose hepatoblastoma [30]. While open biopsies have commonly been advocated, at least one study from the Children's Oncology Group (COG) noted fewer complications in patients undergoing percutaneous rather than open biopsy [31].

Some, including the Liver Tumour Study Group of the International Society of Pediatric Oncology (SIOP), would make the argument that a liver mass associated with an elevated alpha-fetoprotein (AFP) level in a child less than five years of age with no underlying liver disease is hepatoblastoma, and if the child is too sick to undergo a safe biopsy, they could proceed to treatment urgently without obtaining tissue. However, in general, histologic evaluation is preferred, particularly to rule out other tumors in the differential diagnosis. (See 'Differential diagnosis' below.)

Histology — The rarity of hepatoblastoma has resulted in several international efforts to foster international collaboration and clinical trials. One of the results of this collaboration is an international consensus classification of pediatric liver tumors, including hepatoblastoma, that was proposed in 2014 (table 1) [32], and has since been adopted in the 2019 5^th edition of the World Health Organization Classification of Tumours of the Digestive System (table 2) [33].

In children, hepatoblastoma consists of malignant liver cells at various stages of maturation and a variable mesenchymal component (see "Pathology of malignant liver tumors", section on 'Hepatoblastoma'):

●The epithelial component always predominates and consists of two types of cells: "embryonal"-type cells, which are small, basophilic, darkly stained with uniform, hyperchromatic nuclei and scanty cytoplasm, arranged in sheets, ribbons, rosettes, acini, or tubules; and "fetal"-type cells, resembling hepatocytes with central round to oval nuclei and abundant granular or clear cytoplasm, depending on the amount of glycogen or fat (picture 1). The fetal-type cells are larger, eosinophilic, and lighter stained than the embryonal-type cells and arranged in trabeculae or plates. They are separated by sinusoids and may form bile canaliculi. The fetal pattern evokes the prenatal fetal liver, with sheets of uniform cuboidal cells showing low mitotic activity; the more immature embryonic type is characterized by a higher cell density, enlarged nuclei, and frequent mitoses. AFP is almost always demonstrable in the cytoplasm of the epithelial cell component by immunohistochemical staining (IHC).

●If present, a mesenchymal component, most commonly osteoid tissue and rarely cartilage (picture 1), rhabdomyoblasts, or neural elements, rules out hepatocellular cancer (HCC), at least in children. Numerous hepatoblastoma subtypes have been identified (table 2), some with potential prognostic implications. As an example, patients with the purely fetal subtype and a low mitotic rate (<2 mitoses/10 high-power fields), now termed well-differentiated fetal hepatoblastoma, have a favorable prognosis [34,35]. The small cell undifferentiated (SCU) subtype has been associated with a worse prognosis (and often are associated with normal levels of the tumor marker AFP [36-38]). (See "Pathology of malignant liver tumors", section on 'Hepatoblastoma'.)

However, it has become apparent that some of the cases initially described as SCU hepatoblastoma lacked expression of chromatin, subfamily B, member 1/Integrase interactor-1 (SMARCB1/INI1) by IHC (implying a mutation or deletion in the SWI/SNF-related, matrix-associated, actin-dependent regulator of SMARCB1/INI1 gene) and may be more appropriately classified as rhabdoid tumors, and not true SCU hepatoblastomas which have retained INI-1 expression [32,39,40]. In a study of 35 patients with SCU treated in the AHEP0731 study, the presence of SCU was not associated with an adverse outcome; however, none of these patients had a low AFP (<100 ng/mL) at diagnosis and all but two samples retained INI1 staining [40]. It is important to recognize this variant, as patients whose tumors retain INI1 can be treated according to a risk stratification strategy independent of this factor whereas as those with INI1 loss may benefit from a chemotherapy strategy more appropriate for malignant rhabdoid tumors than hepatoblastoma. In our view, cytogenic evaluation for variants in the SMARCB1/INI1 gene and IHC for INI-1 should be routinely performed in any patient with hepatoblastoma who has an AFP level <100 ng/mL to confirm an accurate diagnosis. (See 'Differential diagnosis' below.)

Differential diagnosis — Even with microscopic examination of tumor tissue, the diagnosis of hepatoblastoma can be challenging, as there are several pathologically similar tumors. It is important to work closely with the pathologist to consider other entities in the differential diagnosis, such as HCC, rhabdoid tumor of the liver, undifferentiated embryonal sarcoma of the liver, ossifying stromal-epithelial tumor, transitional liver cell tumors [41], and in the presence of rhabdomyoblastic differentiation, a teratoma [42].

Differentiating hepatoblastoma from HCC can be particularly challenging as there is both histopathologic and gross overlap between the two tumors, and immunohistochemistry has not aided in differentiation [42]. When the biopsy specimen is small and contains only fetal-type cells, the distinction from well-differentiated HCC may be difficult. Differentiating factors may be the age of the patient, the presence of underlying liver disease in HCC (which is frequently associated with chronic liver disease, such as cirrhosis), extramedullary hematopoiesis in hepatoblastoma, and that surrounding liver tissue is more likely to be normal in hepatoblastoma than in HCC [42].

The differential diagnosis can be especially challenging for the SCU form of hepatoblastoma, which may be a component of an otherwise typical hepatoblastoma, or present as the sole component in the so-called "pure small cell" hepatoblastoma. As noted above, some of these tumors may represent malignant rhabdoid tumors of the liver and not hepatoblastoma, and IHC can help in the distinction [18]. (See 'Histology' above.)

As an example, in an early study of SCU hepatoblastomas, 11 of whom presented with normal or minimally increased AFP, none of whom survived, six lacked nuclear expression of SMARCB1/INI1 by IHC (implying a mutation or deletion in the SMARCB1/INI1 gene), and three had cytogenetic and molecular abnormalities similar to those seen in rhabdoid tumors [38]. A year 2011 consensus panel on pediatric liver tumor classification recommended that for tumors lacking SMARCB1/INI1 nuclear expression, that a diagnosis of a rhabdoid-like tumor should be strongly favored, and that these tumors should be referred for genetic and mutation testing [43]. Rhabdoid tumors of the kidney, soft tissues, and central nervous system have consistently revealed deletions and mutations of the SMARCB1/INI1 gene located in chromosome band 22q11.2 [44,45]. (See "Uncommon brain tumors", section on 'Atypical teratoid/rhabdoid tumor'.)

Radiologic evaluation — High-quality cross-sectional imaging is critical to optimizing treatment for hepatoblastoma because staging and treatment are dependent on the initial imaging analysis. Ultrasound is a particularly useful modality for infants with hepatoblastoma; it typically shows a hyperechoic, solid intrahepatic mass, which occurs in the right lobe 60 to 70 percent of the time [46].

However, although the tumor may be initially visualized on ultrasound, vascular involvement is more accurately assessed by three-dimensional cross-sectional imaging, such as contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI) [28]. CT is usually quicker and easier to get and does not require sedation. However, although it takes longer and requires sedation, MRI may be preferred over CT as it can better discriminate tumor size, margins, and extension into adjacent organs [4].

Radiologic imaging is the basis for designating the Pretreatment Extent of Disease (PRETEXT) stage (figure 1), which denotes the extent of liver involvement, an important aspect of selecting initial treatment [47,48]. (See 'Risk stratification and the pretext staging system' below.)

Cross-sectional imaging of the chest should be obtained in all patients to evaluate for lung metastasis and is preferable to obtain prior to surgery in a patient with a liver mass, as postsedation atelectasis can be difficult to distinguish from pulmonary metastatic disease. Bone metastases are uncommon, and thus, routine radionuclide bone scanning (or fluorodeoxyglucose positron emission tomography) is unwarranted in the absence of suspicion for bone metastases [28].

Laboratory evaluation — Biochemical workup typically reveals nonspecific findings. Liver function tests are often normal or only mildly elevated [49]. Other common laboratory abnormalities include anemia and thrombocytosis [50]. As noted above, rarely patients have ectopic production of adrenocorticotropic hormone and parathyroid hormone-related protein, leading to electrolyte abnormalities and hypercalcemia. (See 'Clinical presentation' above.)

AFP — The primary tumor marker for hepatoblastoma is alpha-fetoprotein (AFP). In most patients, the AFP is elevated [51]. However, AFP is a relatively nonspecific marker, and it cannot fully differentiate hepatoblastoma from conventional HCC although the fibrolamellar variant of HCC can be associated with a low or normal AFP. While not absolute, if the AFP is elevated, the age of the patient (with younger ages favoring hepatoblastoma) and the presence of underlying liver disease (favoring HCC) can be helpful at suggesting the diagnosis. (See "Clinical features and diagnosis of hepatocellular carcinoma", section on 'Alpha-fetoprotein' and "Epidemiology, clinical manifestations, diagnosis, and treatment of fibrolamellar carcinoma", section on 'Tumor markers'.)

Children with hepatoblastoma and a low serum AFP at diagnosis (<100 ng/mL) tend to have a more aggressive course, with widespread disease at presentation, chemoresistance, and a two-year survival rate of only 24 percent [36]. This fact accounts for the inclusion of patients with localized disease but an initial AFP level <100 ng/mL in the high-risk strata of most major hepatoblastoma study group risk stratification models (table 3). (See 'Risk stratification and the pretext staging system' below.)

However, many of these historically described patients with a low serum AFP were more likely to have a malignant rhabdoid tumor than hepatoblastoma. Modern pathologic techniques and cytogenetics, including IHC for nuclear expression of INI1 and gene mutation analysis, can help determine an accurate diagnosis to guide the appropriate therapeutic approach. (See 'Differential diagnosis' above.)

RISK STRATIFICATION AND THE PRETEXT STAGING SYSTEM — The four major study groups for hepatoblastoma, the International Childhood Liver Tumours Strategy Group (SIOPEL), Children's Oncology Group (COG), German Society for Pediatric Oncology and Hematology, and Japanese Study Group for Pediatric Liver Tumors, have historically used disparate risk classification categories, making it difficult to compare outcomes. Fortunately, all groups have increasingly used the radiology-based Pretreatment Extent of Disease (PRETEXT) staging system to refine classification of preoperative tumors. The revised PRETEXT staging system was based on an analysis of data from the Children's Hepatic tumors International Collaboration (CHIC), a shared international database with comprehensive data from over 1600 children treated in eight multicenter hepatoblastoma trials over 25 years [47].

In the PRETEXT system, the extent of liver disease is based on the number of sectors in the liver involved by tumor (figure 1) [32]:

●PRETEXT I – A single sector.

●PRETEXT II – Two sectors involved with two adjoining sections disease free.

●PRETEXT III – Three sectors are involved, and no two adjoining sections are free.

●PRETEXT IV – Four sectors are involved.

●The presence of what are termed "Annotation Factors" is also part of PRETEXT staging and helps guide the surgical approach both at diagnosis and following neoadjuvant chemotherapy. These include the following:

•V – For hepatic/vena cava vein involvement

•P – For portal vein involvement

•E – For extrahepatic extension

•M – For distant metastases

•C – For caudate lobe involvement

•F – For grossly detectable multifocal tumor nodules

•R – Tumor rupture prior to diagnosis

In univariate analysis of the risk groups in the CHIC database, significantly increased risk of an event-free survival event was noted for advanced PRETEXT group, macrovascular venous or portal involvement, contiguous extrahepatic disease, primary tumor multifocality, and tumor rupture at enrollment. Age 8 or older, low alpha-fetoprotein (AFP; <100 ng/mL), and metastatic disease were associated with the worst outcomes [52,53]. Multivariate analysis of these factors led to a risk stratification tree that separates patients into very low-, low-, intermediate-, and high-risk categories [47], and the prognostic ability of this CHIC hepatoblastoma stratification system (CHIC-HS) has been independently validated [54].

Although PRETEXT and the CHIC-HS are now widely used for hepatoblastoma protocols, the different cooperative groups have used the information on disease extent to develop specific risk classification systems that are based on their own adopted staging criteria, and they are not equivalent (table 3). As an example, studies involving high-risk patients within the COG and the Liver Tumour Study Group of the International Society of Pediatric Oncology (SIOP) study groups have not included exactly the same types of patients.

POST-TEXT — For patients whose tumors are not resectable at diagnosis, repeat radiographic evaluation after initial chemotherapy is termed POST-TEXT, and it uses the same radiographic criteria as PRETEXT (figure 1). Downstaging/improvement as compared with PRETEXT stage is used to determine the timing of resection as well as the potential need for liver transplantation or other form of local therapy after upfront chemotherapy. (See 'Role of orthotopic liver transplant' below.)

Of note, tumors can respond to therapy yet not shrink in size. Tumor viability is difficult to discern on follow-up imaging, but a decline in the AFP level may be a useful indicator of chemotherapy sensitivity and response. AFP levels may decrease dramatically without apparent objective tumor shrinkage, and this should not prompt a chemotherapy regimen change, as there are no data that suggest that doing so would induce further objective tumor response. Changes in chemotherapy are also not supported by the rate of AFP decline as long as continuous ongoing decline is occurring [55].

OVERVIEW OF TREATMENT AND OUTCOMES — All children and young adults with hepatoblastoma should be managed by a multidisciplinary team that are familiar with modern treatment principles. Participation in clinical trials is recommended, if feasible.

Prognosis and evolution of treatment over time — The survival rate for hepatoblastoma has increased dramatically over the last 30 years in children, largely paralleling improvements in treatment, particularly chemotherapy, surgical techniques, supportive care, and the introduction of orthotopic liver transplantation (OLT) for locally unresectable disease. Contemporary studies have revealed a five-year event-free survival rate of 75 to 90 percent in children treated with chemotherapy and surgery [56-60]. This represents a dramatic improvement from the prechemotherapy era in the 1980s (five-year survival rate 30 percent) [28,34]. For patients with unresectable tumors but no metastases, five-year survival rates ≥75 percent are reported with chemotherapy plus OLT, although these studies are retrospective and generally report only small numbers of patients [58,61-65]. (See 'Role of orthotopic liver transplant' below.)

Treatment for hepatoblastoma has evolved over time along the following lines:

●Complete tumor resection is necessary for achieving long-term survival; however, there is no uniform consensus on optimal timing. Internationally, the approach to localized hepatoblastoma has differed:

•Most North American, German, and Japanese protocols have favored upfront hepatic resection, when feasible, which is the case in approximately one-third of newly diagnosed patients (PRETEXT I and II tumors with preoperatively radiographically clear venous margins (figure 1)).

•On the other hand, countries that participated in the trials of the Liver Tumor Study Group of the International Society of Pediatric Oncology (SIOP; including some North American institutions) have largely favored neoadjuvant chemotherapy for all patients, regardless of initial assessment of resectability or histology. This typically results in the administration of higher cumulative doses of chemotherapy.

•Surgical morbidity and mortality rates are relatively equivalent with either approach.

●Patients treated with surgery alone for hepatoblastoma have a high recurrence rate, with nearly 40 percent recurring within ten years after excision alone [50]. Because of this, most patients require chemotherapy in combination with surgery for optimal chance for cure. (See 'Histology' above.)

The only exception is patients with resectable tumors at diagnosis who have 100 percent pure-fetal histology and low mitotic activity (well-differentiated fetal [WDF] tumors); these individuals have 100 percent survival with surgery alone and no chemotherapy. This approach was initially described by the Children's Oncology Group (COG) [65].

●For patients initially treated with neoadjuvant therapy and delayed resection, the finding of 100 percent WDF histology is not considered as favorable, and these patients typically continue chemotherapy postresection to optimize outcomes. This is in part due to the inability to know if this histology was present in the entire tumor prior to starting chemotherapy.

●For patients who undergo upfront resection, but do not have 100 percent well-differentiated histology, the COG has reported excellent outcomes using only two cycles of combination chemotherapy (cisplatin, 5-fluorouracil [FU], and vincristine [C5V]) [56]. This diminishes acute and long-term toxicities and improves quality of life as well as the cost of care.

●Patients who are not able to undergo surgical resection either by conventional approach or extreme resections and who do not have metastatic disease are candidates for OLT, which removes local tumor and provides long-term disease-free survival.

Other methods to control disease such as thermal ablation (radiofrequency ablation, microwave ablation) or transarterial chemoembolization, have been used in an attempt to further decrease the tumor to allow for surgical resection or for disease control of unresectable tumors; however, data are limited for long-term survival outcomes with these approaches. (See 'Role of transarterial chemoembolization' below.)

●Patients with metastatic disease have a dismal prognosis. Metastatic disease at diagnosis is considered a high-risk feature in all risk stratification schemes (table 3). (See 'Risk stratification and the pretext staging system' above.)

The SIOPEL-4 trial established intensified upfront cisplatin with doxorubicin as induction therapy for patients with metastatic disease [62]. Those patients who do not have complete resolution of pulmonary disease continue therapy with carboplatin/doxorubicin.

●Evolution of chemotherapy approaches – The introduction of cisplatin and doxorubicin led to significant improvements in the outcomes of children with hepatoblastoma and was clearly advantageous over prior chemotherapy regimens which largely consisted of cyclophosphamide, vincristine, and FU. Other chemotherapeutic agents (vincristine, FU, carboplatin, etoposide, ifosfamide, and irinotecan) have been used to treat newly diagnosed hepatoblastoma in the pediatric population.

Contemporary cisplatin-containing chemotherapy protocols from the four major study groups have evolved along different pathways, based upon published study results and defined risk groups [66]. The different international groups have reported similar five-year event-free survival outcomes of over 70 percent, so debate still continues regarding what regimen is optimal.

•COG – The COG has generally advocated postoperative chemotherapy with a combination of C5V.

Based upon the INT-0098 study [67], COG adopted C5V rather than cisplatin/doxorubicin regimen for patients who undergo upfront resection because of the similar efficacy and more favorable toxicity profile [56].

In the most recent COG study, AHEP0731, 102 patients with newly diagnosed intermediate-risk hepatoblastoma (unresectable, nonmetastatic stage III disease [n = 93] or stage I/II small cell undifferentiated [SCU] histology) were treated with six cycles of doxorubicin, cisplatin, 5-FU, and vincristine (C5VD) [68]. This regimen had tolerable toxicity and excellent cure rates, with a five-year progression-free and overall survival of 88 and 95 percent, respectively. This regimen is currently being tested in a randomized fashion in patients with intermediate-risk hepatoblastoma (NCT03533582).

In addition, data from the AHEP0731 trial have also shown the activity of irinotecan in patients with high-risk metastatic disease or those with alpha-fetoprotein (AFP) <100 ng/mL [69].

•SIOPEL – As noted above, in countries that participated in the Liver Tumor Study Group of the SIOP trials, patients with hepatoblastoma typically undergo induction chemotherapy prior to resection regardless of initial resectability.

The International Childhood Liver Tumours Strategy Group (SIOPEL) has described success in using single-agent cisplatin for patients with resectable disease, while the combination of cisplatin plus doxorubicin has been the SIOPEL standard for patients with advanced unresectable but nonmetastatic disease [70].

The SIOPEL-4 trial established intensified upfront cisplatin plus doxorubicin as induction therapy for patients with metastatic disease [62]. Those patients who do not have complete resolution of pulmonary disease continue preoperative therapy with carboplatin plus doxorubicin before surgery was attempted.

Approaches to minimizing long-term treatment-related toxicity — One of the most important aspects of modern treatment for hepatoblastoma is minimizing long-term treatment-related toxicity, including cardiotoxicity and ototoxicity.

●Cardiotoxicity – The use of dexrazoxane with all doses of doxorubicin in children being treated for hepatoblastoma is encouraged, where available. Current hepatoblastoma trials utilize bolus doxorubicin dosing, which facilitates the use of dexrazoxane, which cannot be used in patients receiving a continuous infusion doxorubicin.

Historically, trials that included doxorubicin utilized continuous infusion for as long as 96 hours which, while less cardiotoxic than bolus dosing, is more cumbersome to administer and requires longer inpatient treatment. While the extent of cardiac toxicity in hepatoblastoma has been relatively low in most trials selected cases do occur in patients treated with anthracycline-containing regimens [59,71]. (See "Cancer survivorship: Cardiovascular and respiratory issues", section on 'Childhood cancer survivors'.)

Dexrazoxane is the only drug that is US Food and Drug Administration (FDA) approved specifically to prevent anthracycline-mediated cardiotoxicity. Although much of the evidence supports the safety of dexrazoxane as a cardioprotectant, whether it has any detrimental impact on cancer-related outcomes remains controversial. As a result, the FDA has restricted approval in adults to patients with metastatic breast cancer who have received at least 300 mg/m² doxorubicin and have an indication for continued doxorubicin.

However, in children, dexrazoxane has been used in some leukemia trials in an attempt to prevent doxorubicin-related cardiomyopathy, with no negative influence on cancer-related outcomes [72,73]. (See "Risk and prevention of anthracycline cardiotoxicity", section on 'Use of dexrazoxane'.)

●Ototoxicity – For patients receiving six cycles of cisplatin monotherapy for hepatoblastoma (cumulative dose 480 mg/m²) we suggest the use of sodium thiosulfate (STS). For patients without metastatic disease who are receiving lower doses of cisplatin or combination therapy, clinicians should consider the risks and benefits before recommending the use of STS to patients and families. We recommend against the use of STS for patients with metastatic disease. The currently accruing Pediatric Hepatic Malignancy International Therapeutic Trial (PHITT) permits the use of STS for treatment groups A, B, and C (very low-, low-, or intermediate-risk disease) at the discretion of the treating center, but does not allow its use for those in group D (high-risk disease). (See 'Our suggested approach to initial treatment' below.)

Hearing loss is one of the most concerning long-term toxicities associated with the use of platinum agents (predominantly cisplatin) for hepatoblastoma therapy. Profound hearing impairment can interfere with subsequent speech and language development and impair neurocognitive function. (See "Overview of neurologic complications of platinum-based chemotherapy", section on 'Incidence and risk factors'.)

Various strategies have been considered to minimize cisplatin ototoxicity:

•Early detection of ototoxicity in children receiving platinum agents may minimize the risk of severe impairment in the frequencies required for speech recognition by providing an opportunity for treatment modification, if possible before auditory damage becomes severe. However, clinical trials have not established the efficacy or safety of reduced cisplatin dosing in response to hearing changes in patients with hepatoblastoma [74], and we do not support this approach without discussions with the families about the risks versus benefits.

•Protection against platinum-induced hearing loss has been achieved in animal models and in patients using STS [75-77]. At least two randomized trials suggest benefit (including one conducted entirely in patients with hepatoblastoma) but questions have been raised as to potential interference with cancer-directed therapy:

-In the COG ACCL0431 trial, 125 children receiving cisplatin for a variety of malignancies (only five patients had hepatoblastoma) who were not registered in a cancer-directed COG therapeutic study were randomly assigned to a single dose of STS six hours after each cisplatin dose or to observation [77]. The use of STS was associated with a significantly smaller proportion of patients with hearing loss four weeks after the final cisplatin dose (29 versus 56 percent, odds ratio of hearing loss 0.31, 95% CI 0.13-0.73). However, in a post hoc analysis of the subgroup of patients (47 patients) with disseminated disease, the three-year overall survival was significantly lower in the thiosulfate group when compared with the control group (45 versus 84 percent, HR 4.10, 95% CI 1.3-13.0).

-On the other hand, in the randomized phase III SIOPEL-6 trial, in which 109 children with standard-risk nonmetastatic hepatoblastoma receiving single-agent cisplatin prior to a planned delayed resection were randomly assigned to receive or not receive STS (a single dose of 20 g/m² administered over 15 minutes, six hours after each of the six cisplatin doses), the addition of STS resulted in a significantly lower incidence of cisplatin-induced hearing loss (hearing loss of grade 1 or higher on the Brock scale [78]) in 33 versus 63 percent (relative risk 0.52, 95% CI 0.33-0.81) [79]. At a median follow-up of 52 months, cancer outcomes were not worse in the STS group (three-year event-free survival 82 versus 79 percent, overall survival 98 versus 92 percent). Expert reviewers considered that this study provided moderate quality of evidence regarding efficacy, but only low quality of evidence regarding potential harms [74].

Largely based on these two studies, sodium thiosulfate has been approved by the FDA to decrease the risk of cisplatin-related ototoxicity in pediatric patients aged one month or older who have received a diagnosis of localized, nonmetastatic solid malignancies [80]. The approved dose is based on actual body weight, as outlined in the United States Prescribing Information [81].

Evidence-based guidelines for prevention of cisplatin-induced ototoxicity in children and adolescents with cancer are available from a multidisciplinary multinational expert panel [82]. Regarding the use of STS, the panel made a strong recommendation for its use in patients treated for nonmetastatic hepatoblastoma and a weak recommendation against its use in patients with metastatic cancer. The current AHEP 1531 trial and the PHITT trial allow the use of STS for groups A, B, and C at the discretion of the treating center but does not allow use for risk groups D, E, and F.

Our suggested approach to initial treatment — Whenever feasible, we recommend that patients with hepatoblastoma be enrolled in clinical trials. The PHITT trial has united the international pediatric liver groups and provided consistent nomenclature and therapeutic approaches to treatment of hepatoblastoma. This trial is a result of the international collaboration that adopts many of the unique therapeutic concepts and approaches from each of the main international groups. It opened to patient enrollment in 2018 and is expected to accrue patients until 2025. Whenever possible, patients should be enrolled in the PHITT trial. (See 'Overview of treatment and outcomes' above.)

Patients are eligible for the trial up to age 30. (See 'Issues in adults' below.)

The PHITT trial risk stratifies patients for treatment based on age, time of tumor resection, PRETEXT group and annotations factors, AFP levels, pathology, and presence of metastases (table 4).

●Very low- and low-risk patients

•Patients who undergo initial resection and who have 100 percent pure fetal histology and low mitotic activity (WDF tumors) are offered observation only following resection (very low-risk, group A1).

•Patients who undergo upfront resection, but do not have 100 percent well-differentiated histology, are treated with two cycles of postoperative cisplatin monotherapy (very low-risk, group A2).

•Patients with unresectable tumors at diagnosis with no PRETEXT annotation factor are treated with two cycles of cisplatin monotherapy. Patients who undergo delayed resection after two cycles of initial cisplatin therapy are randomly assigned to two versus four cycles of lower-dose interval-compressed postoperative cisplatin monotherapy (low-risk, group B1).

•For patients whose tumors are not resected after two cycles of chemotherapy, a total of six cycles of treatment are administered (low-risk, group B2). After the four preoperative chemotherapy cycles, patients are evaluated for attempted resection or liver transplantation.

●Intermediate-risk patients

•Patients with unresectable tumors at diagnosis and no metastatic disease but who have annotation factors are considered to have intermediate-risk disease and they are randomized to either six cycles of interval-compressed cisplatin monotherapy or the standard COG C5VD regimen, as was used for the intermediate-risk arm of the AHEP0731 COG trial [83]; in both groups, patients undergo surgery after cycle 2 or 4 (intermediate-risk, group C).

●High-risk patients

•The definition of high-risk disease is presented in the table (table 4).

•Patients with metastatic disease at diagnosis and other high-risk groups generally have a poor prognosis [47,83]. All such patients will receive induction chemotherapy with a cisplatin-intensive regimen (weekly cisplatin and doxorubicin) as reported by the SIOPEL-4 trial [62] followed by tumor resection. Patients who have complete response (CR) of their lung metastases (chemotherapy alone or with surgical metastasectomy) will receive three cycles of chemotherapy consolidation with carboplatin and doxorubicin (high-risk, group D1).

•Patients who do not achieve lung CR at the end of induction chemotherapy, will be randomized to receive six cycles of either carboplatin plus doxorubicin alternating with carboplatin plus etoposide, or carboplatin plus doxorubicin alternating with vincristine plus irinotecan (high-risk, group D2).

Role of orthotopic liver transplant — Patients who are not able to undergo liver tumor surgical resection and who achieved complete CR of their lung metastases (chemotherapy and/or surgical resection) are candidates for liver transplantation, which may be the only method for removing all local tumor. The success of liver transplantation for hepatoblastoma tumor removal is now well established, and early referral to a liver transplant center for evaluation is recommended for all newly diagnosed unresectable patients rather than awaiting the results of initial chemotherapy. In general, liver transplantation should be performed after induction for patients enrolled on the PHITT study or after two or four cycles of chemotherapy, thus avoiding prolonged chemotherapy administration and additional potential preventable toxicities, and allowing for consolidation chemotherapy to be given after transplant.

In the United States, patients with biopsy-proven hepatoblastoma are considered a special case status 1 for listing for OLT by the Organ Procurement and Transplantation Network (OPTN). Involvement of the portal and hepatic vasculature, and even lung metastasis, are not considered a strict contraindication to liver transplant as long pulmonary lesions can be eradicated with chemotherapy or with an aggressive approach to surgical metastasectomies or wedge resection with wide margins. On the other hand, patients with extrahepatic metastasis not amenable to resection or who lack a response to neoadjuvant chemotherapy are not candidates for liver transplant. The prognostic value of tumor rupture requires further evaluation [84].

Reported long-term survival rates for patients treated with chemotherapy plus liver transplantation are ≥75 percent, although these are all retrospective series, which are based on only small numbers of patients [58,62-65,85-87]. The following represents the range of findings:

●A French retrospective study reported posttransplant survival rates at one and four years of 100 and 83.3 percent for 13 patients who underwent liver transplantation from 2001 to 2009 [85]. Two had pulmonary metastases at diagnosis and seven had extrahepatic vascular involvement.

●The SIOPEL database identified a 10-year posttransplant survival rate of 66 percent (n = 12), and a review of the world literature (n = 106) reported an 82 percent six-year posttransplantation survival rate for primary transplantation, whereas a survival rate of 30 percent was noted for rescue liver transplantation (n = 41) [86].

●In a report of outcomes from the United Network for Organ Sharing database of the OPTN for 783 children undergoing OLT for a primary hepatic malignancy between 2002 and 2020, 688 for hepatoblastoma, five-year overall survival for those with hepatoblastoma was 82 percent and the graft survival rate was 76 percent [86].

Role of transarterial chemoembolization — Transarterial therapies lead to local tumor regression in many cases and might be considered an alternative for patients with unresectable liver tumors who do not respond to primary chemotherapy and are not candidates for liver transplantation for various reasons.

Transarterial chemoembolization (TACE) has been used as a treatment modality in some cases of unresectable disease with good local effect, although there are no randomized or prospective trials [88-92]. The available studies, which are all based on small numbers, demonstrate that TACE is overall well tolerated, and in some cases can induce significant tumor regression, allowing some patients to become amenable to surgical resection.

Transarterial radioembolization using yttrium-90 is commonly used in adults but has only recently been reported in children [93]. This approach is less morbid than TACE and is a candidate for organized clinical study.

Treatment of relapsed disease — Patients may recur with isolated lung disease, isolated local/liver disease, a combination of local and lung relapse, or distant relapse in nonpulmonary sites. For most patients, the outcomes are dismal, but a minority may achieve long-term survival with combination of chemotherapy and surgery.

Recurrent hepatoblastoma is curable in some patients. Pulmonary metastasectomy alone for patients with isolated pulmonary relapse may be effective [94]. On the other hand, the use of a salvage or "rescue" liver transplant has a significantly worse outcome when compared with its use in primary therapy [85]. (See 'Role of orthotopic liver transplant' above.)

For most patients, combined treatment with chemotherapy and surgical removal of the tumor is essential for long-term survival [94]. There is no standard chemotherapy regimen for patients with relapsed disease. Most studies have enrolled few patients, precluding the ability to ascertain true disease responsiveness and there are no standardized response criteria [95]. The following reflects our approach to these patients:

●For those who did not receive upfront doxorubicin, a doxorubicin-containing regimen (eg, ifosfamide/doxorubicin or carboplatin/doxorubicin) is recommended. An analysis of patients enrolled on the randomized COG study INT-0098 (which demonstrated that patients treated upfront with either C5V or cisplatin/doxorubicin had similar survival, but C5V was less toxic), included 55 patients who experienced progression or recurrence after initial treatment [96]. Overall, 11 (31 percent) of the 36 patients treated with C5V were successfully retrieved with a doxorubicin-containing regimen and surgery and remained alive at last contact, while only 1 (18 percent) of 18 patients initially treated with cisplatin/doxorubicin was alive after retrieval therapy.

●Patients who received doxorubicin as part of upfront therapy do worse than those who did not. As patients have typically received significant amounts of cisplatin and doxorubicin as part of primary therapy, re-use of these agents in the treatment of relapsed disease is often limited by cumulative dosing and exposure.

Options include: ifosfamide/carboplatin/etoposide regimen [97], carboplatin plus etoposide [95,98], docetaxel [99], and, if not used upfront, single-agent irinotecan or irinotecan plus vincristine [100]. The efficacy of irinotecan in hepatoblastoma was initially described in relapsed patients [96,100-102], but irinotecan is now being used in some patients with high-risk disease as part of upfront therapy [69]. (See 'Prognosis and evolution of treatment over time' above.)

Issues in adults — As with children, our recommended approach is enrollment into the PHITT trial, if feasible. Enrollment is permitted up to age 30.

Treatment principles are largely extrapolated from the experience in children with hepatoblastoma; however, the benefits of specific treatment strategies are less certain in adults.

Mean survival in the adult population is significantly worse than in the pediatric population, in part related to the later stage at diagnosis. The median survival time after diagnosis was as low as 3.5 months in one series [25]. One-year survival has been estimated to be 30 to 40 percent [49,103]. The longest documented surviving case in the adult population is 151 months in a 21-year-old female treated with resection, with recurrences treated with hepatic TACE, systemic chemotherapy, and 19 percutaneous alcohol injections [104]. In one series, not surprisingly, curative liver resection was an independent prognostic factor for improved survival in Cox multivariate analysis [49]. Even within the adult population, younger individuals tend to do better. In one series, the one-year survival rate was 42 percent in patients under 45 years of age and 0 percent in patients over 45 [42].

SUMMARY AND RECOMMENDATIONS

●Epidemiology and risk factors – Hepatoblastoma is a rare tumor overall, but it is the most common primary malignant hepatic neoplasm in children. Ninety percent of cases arise in children under the age of five. Most cases are sporadic and not associated with chronic liver disease. (See 'Epidemiology and risk factors' above.)

●Clinical presentation – In children, hepatoblastoma often presents as an asymptomatic mass in the liver but there may be nonspecific symptoms such as abdominal pain, nausea, or vomiting. (See 'Clinical presentation' above.)

●Diagnosis and initial evaluation

•Although the majority of hepatoblastomas are first identified by imaging, tissue is typically obtained for formal histopathologic diagnosis. (See 'Diagnosis and initial evaluation' above.)

•Alpha fetoprotein (AFP) is the primary tumor marker for hepatoblastoma and, while elevated in most patients, it does not help to differentiate hepatoblastoma from conventional hepatocellular cancer. (See 'AFP' above.)

•High-quality cross-sectional imaging is critical to optimizing treatment for hepatoblastoma because staging and treatment are dependent on the initial imaging analysis. (See 'Radiologic evaluation' above.)

●Risk stratification and treatment selection – In the PRETEXT staging system, the extent of liver disease is based on the number of sectors in the liver involved by tumor (figure 1). Other factors affecting risk stratification include annotation factors such as hepatic/vena cava or portal vein involvement, extrahepatic extension, metastatic disease at presentation, caudate lobe involvement, multifocality, tumor rupture, and the serum level of AFP (table 3). These risk categories are used for treatment selection. (See 'Risk stratification and the pretext staging system' above and 'Our suggested approach to initial treatment' above.)

●Treatment and outcomes

•The survival rate for hepatoblastoma has increased dramatically over the last 30 years in children, largely paralleling improvements in treatment, particularly the introduction of cisplatin chemotherapy, and increasing use of orthotopic liver transplantation for locally unresectable disease. All children and young adults with hepatoblastoma should be managed by a multidisciplinary team who is familiar with modern treatment principles. Participation in clinical trials is recommended for all patients, if feasible. (See 'Overview of treatment and outcomes' above.)

•Complete tumor resection is necessary for achieving long-term survival; however, there is not uniform consensus on optimal timing. For most patients with PRETEXT I and II tumors (figure 1) with preoperatively radiographically clear venous margins, we suggest upfront surgery rather than initial chemotherapy (Grade 2C). For all other patients, we start with chemotherapy. Other international approaches have favored neoadjuvant chemotherapy for all patients, regardless of initial assessment of resectability or histology. (See 'Overview of treatment and outcomes' above.)

•Early referral to a liver transplant center for evaluation is appropriate for all newly diagnosed patients with unresectable disease. The intent is to perform the necessary surgical procedure by the time that two-thirds to three-fourths of planned chemotherapy has been given. Patients with extrahepatic metastasis not amenable to resection or who lack a response to neoadjuvant chemotherapy are not candidates for liver transplantation. (See 'Role of orthotopic liver transplant' above.)

•Beyond surgery, our general approach for both children and adults up to age 30 entails enrollment into the Pediatric Hepatic Malignancy International Therapeutic Trial (PHITT). The PHITT trial risk stratifies patients for treatment based on age, time of tumor resection, PRETEXT group and annotations factors, AFP levels, pathology, and presence of metastases (table 4). (See 'Our suggested approach to initial treatment' above.)

•Current hepatoblastoma trials utilize bolus rather than infusional doxorubicin dosing, which facilitates the use of dexrazoxane, which cannot be used in patients receiving a continuous infusion doxorubicin. The use of dexrazoxane with all doses of doxorubicin in children being treated for hepatoblastoma should be encouraged, where available. (See 'Approaches to minimizing long-term treatment-related toxicity' above.)

For patients receiving six cycles of cisplatin monotherapy for hepatoblastoma (cumulative dose 480 mg/m²) we suggest the use of sodium thiosulfate (STS) (Grade 2C). For patients without metastatic disease who are receiving lower doses of cisplatin or combination therapy, clinicians should consider the risks and benefits before recommending the use of STS to patients and families. We recommend against the use of STS for patients with metastatic disease (Grade 1B). The currently accruing PHITT trial permits the use of STS for treatment groups A, B, and C (very low-, low-, or intermediate-risk disease) at the discretion of the treating center, but does not allow its use for those in group D (high-risk disease).

•Patients may recur with isolated lung disease, isolated local/liver disease, a combination of local and lung relapse, or distant relapse in nonpulmonary sites. For most patients, the outcomes are dismal, but a minority may achieve long-term survival with combination of chemotherapy and surgery. (See 'Treatment of relapsed disease' above.)

ACKNOWLEDGMENTS — The UpToDate editorial staff acknowledges Howard M Katzenstein, MD, who contributed to earlier versions of this topic review.