Hepatoblastoma: recent developments in research and treatment

Hepatoblastoma: recent developments in research and treatment

Seminars in Pediatric Surgery (2012) 21, 21-30 Hepatoblastoma: recent developments in research and treatment Dietrich von Schweinitz, MD, PhD From Dr...

1MB Sizes 0 Downloads 23 Views

Seminars in Pediatric Surgery (2012) 21, 21-30

Hepatoblastoma: recent developments in research and treatment Dietrich von Schweinitz, MD, PhD From Dr. von Hauner Children’s Hospital, University of Munich, Munich, Germany. KEYWORDS Hepatoblastoma; Surgery; Chemotherapy; Risk stratification; Molecular biology

Hepatoblastoma is the most common liver tumor of early childhood. According to recent studies its incidence seems to be increasing in North America and Europe. Since new histological variants have been described recently the formerly clear-cut distinction of hepatoblastoma and hepatocellular carcinoma may not be valid anymore and a new histological classification will be inaugurated by an international working group. Recent research identified prognostically relevant gene signatures as well as potential molecular targets for therapy of hepatoblastoma. The multicentric study groups in the USA, Europe and Japan recommend cisplatin based chemotherapy for neoadjuvant and adjuvant treatment. However, their risk stratification systems and general treatment strategies differ substantially. Therefore the four groups agreed to pool their patients’ data for an analysis of prognostic criteria which can be used for defining common risk groups. While 90% of standard risk and 65% of high risk hepatoblastomas can be cured, the still dismal outcome of multifocal disseminated and metastasising tumors warrants the investigation of new cytotoxic drugs and substances against specific molecular targets. © 2012 Elsevier Inc. All rights reserved.

Hepatoblastoma is a highly malignant embryonal liver tumor that almost exclusively occurs in infants and toddlers. Its clinical importance is based on the fact that it is the most common primary liver tumor of childhood and after neuroblastoma and nephroblastoma the third most common abdominal neoplasm in this age group.1 Because current knowledge on hepatoblastoma has been extensively summarized very recently in a textbook chapter1 as well as in a monograph on pediatric liver tumors,2 this review will summarize advances of greatest importance for the clinician and include findings published during the last 12 months.

Epidemiology The incidence of hepatoblastoma in the United States and Europe seems to be slowly increasing with now 1.2 to 1.5 Address reprint requests and correspondence: Dietrich von Schweinitz, MD, PhD, Dr. von Hauner Children’s Hospital, University of Munich, Lindwurmstrasse 4, D-80337, Munich, Germany. E-mail: [email protected]

1055-8586/$ -see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1053/j.sempedsurg.2011.10.011

cases per million population. The rates in boys are higher (1.57) than in girls (1.09). In the spectrum of pediatric tumors, it still is a rare tumor, comprising approximately 1% of all pediatric malignancies.3,4 According to recent analytical epidemiologic studies, there are several environmental risk factors associated with hepatoblastoma. Premature birth and very low birthweight have been found to be associated with the later appearance of hepatoblastoma in several developed countries. Increase in these patient cohorts may be the cause for the increased frequency of this tumor. Oxygen therapy, medication such as furosemide, total parenteral nutrition, radiation, plasticizers, and other toxins are thought to play a role, but the exact mechanisms are not yet understood.1,4 Still, very low birthweight infants should be monitored for hepatoblastoma. Also, the authors of recent studies have shown maternal and paternal preconceptional and gestational tobacco smoking to be a risk factor for hepatoblastoma, which accounts especially for regular smoking of both parents.3,4 Therefore, the International Agency for Research on Cancer has classified tobacco smoke (via the parents) as a carcinogen for hepatoblastoma in

22 Table 1

Seminars in Pediatric Surgery, Vol 21, No 1, February 2012 Histologic classification of tumors6

Hepatoblastoma, wholly epithelial type Fetal subtype Mixed embryonal/fetal subtype Macrotrabecular subtype Small cell undifferentiated subtype Hepatoblastoma, mixed epithelial, and mesenchymal type Without teratoid features With teratoid features Hepatoblastoma, not otherwise specified

2009. The association of hepatoblastoma with fetal alcohol syndrome has been reported anecdotally.4 A large epidemiologic study is in progress at the University of Minnesota (HOPE study ⫽ hepatoblastoma origins and pediatric epidemiology; http://www.cancer.umn.edu/hopestudy) to systematically collect more data. A large number of congenital syndromes have been described in patients with hepatoblastoma, but only trisomy 18 (Edward’s syndrome), familial adenomatous polyposis, and Beckwith–Wiedemann syndrome have been clearly shown to increase the risk of hepatoblastoma. The age at presentation in conjunction with these syndromes is equivalent to that of the sporadic cases.4 Therefore, follow-up investigations of these children should continue through the preschool age period.

Histopathology and molecular biology Histology Hepatoblastoma is obviously derived from stem cells or early progenitor cells of the different epithelial components of the liver.5 The main histologic features, such as fetal, embryonal, small cell undifferentiated, macrotrabecular, and teratoid, found in pure epithelial and mixed epithelial and mesenchymal hepatoblastomas are well described and have been used for classification for many years (Table 1).1,6 It is known that these features are associated with different prognoses; pure fetal histology is favorable, and small cell undifferentiated and macrotrabecular histology are unfavorable. It became evident recently that the picture is more complex with more oncogenetic information, leading to the concept of a “family” of hepatoblastoma tumors.7 This family includes a subset of small cell undifferentiated hepatoblastomas with rhabdoid features and a lack of INI1 expression and their possible genetic connection to rhabdoid tumors of the liver.5,8 Other variant histologic features of different cell lineages have been described, leading to a proposed extended classification,1,6 which presently is being considered by an international working group of experts to establish a common language. Furthermore, the clear-cut distinction between hepatoblastoma and hepatocellular carcinoma is no longer clear because so-called transitional liver cell tumors have been identified which contain elements of both

hepatoblastoma as well as hepatocellular carcinoma.1,6,9 After chemotherapy, hepatoblastomas show necrosis and a fibrohistiocytic response. Mesenchymal components can now predominate,6 areas of cytoarchitectural differentiation can mimic normal liver, and other tumor components may also resemble hepatocellular carcinoma.10

Cytogenetics In recent conventional cytogenetic analysis performed on hepatoblastomas investigators have found several alterations with chromosomal gains occurring more frequently than losses.11 The most common alterations are trisomy 2, 8, and 20. Comparative genomic hybridization has enabled researchers to further define chromosomal alterations and to detect regions with proposed hepatoblastoma relevant genes. Recently, the implementation of whole-genome DNA chip-based technologies led to the detection of many altered genomic regions, of which gain of material on chromosome 2q 13-22, 2q 36-37 and deletions of 2p and 4q were associated with advanced tumors and poor prognosis.11,12 In analyses, investigators have identified loss of heterozygosity on chromosome 1p 32, 1p 36.3, and 11p 15.5, the latter of which is often associated with Beckwith–Wiedemann syndrome and leads to a loss of the tumor suppressor gene H19 and activation of both IGF-2 alleles.11

Molecular biology Several molecular signaling pathways that are known to be essential for embryonal development have been found to be altered in hepatoblastoma cells. Thus, the WNT (wingless) pathway is activated in many hepatoblastomas, leading to a stabilization of the key molecule beta-catenin, which then accumulates in the tumor cell nuclei and activates several proliferation -associated target genes, such as myc and cyclin D1. The WNT activation in hepatoblastomas can be caused by mutation of beta-catenin itself but also by the mutation of other important pathway molecules, such as the APC gene (in patients with familial adenomatous polyposis coli), AXIN1, and conduction (AXIN2).11 Another developmental signaling pathway aberrantly activated in a large portion of hepatoblastomas is the hedgehog (HH) pathway. High mRNA or protein levels of HH ligands have been found, as well as up-regulation of the downstream targets GLI1, BCL2, and PTCH. Along with the detection of a down-regulation of the negative regulator hedgehog interacting protein (HHIP) in some hepatoblastomas, an autocrine activation of HH signaling can be anticipated in these tumors.11 Immunohistochemically, a correlation between the expression of the HH pathway components Smo and Gli1 and tumor grade, size, stage, and prognosis was recently found.13 The genetic and epigenetic alterations of the IGF2/H19 locus on chromosome 11p 15.5 lead to an overexpression of the growth factor IGF-2, which again results in

von Schweinitz

Hepatoblastoma: Recent Developments in Research and Treatment

an activation of the IGF/phosphoinositide 3-kinase/AKT pathway in the majority of hepatoblastomas. It was convincingly shown that the downstream targets of the IGF axis, namely AKT and mammalian target of rapamycin, are strongly expressed and activated in these tumors, which depend on this activated pathway for their survival.11 The hepatocyte growth factor/scatter factor (HGF) stimulates several different cellular activities, including growth, invasion, motility, differentiation, and angiogenesis and is active in embryonal development, wound healing, and organ regeneration. The HGF/c-Met receptor signaling can be transmitted via the PI3K/AKT as well as the mitogen-activated protein kinase pathways. It has been shown that patients with hepatoblastoma have elevated HGF serum levels, which even increase after surgical resection, and that hepatoblastoma cells react to this growth factor with migration and invasion as well as with prolonged survival.11 Very recently, the receptor tyrosine kinase erbB2, which is up-regulated in many adult carcinomas, was also found to be overexpressed in many hepatoblastomas and may also have an important functional role for this tumor.14 Furthermore, recent experimental work showed an up-regulation of the activated pathway regulator NOTCH2,15 of the antiapoptotic molecule serpin B3,16 several multiple-drug resistance associated genes,1 as well as of microRNA-49217 in hepatoblastomas, which all may contribute to the malignant behavior. Investigations on epigenetically altered tumor suppressor genes have shown in the last few years that promoter hypermethylation is a common event in hepatoblastomas, which leads to activation of signaling pathways (WNT and HH, see above in this section) or suppression of antiproliferative and apoptosis regulators.11 The authors of several recent studies were concerned with the alternative approach by DNA microarray analysis for molecular phenotyping of hepatoblastomas. In the most promising one, Cairo et al18 identified a 16-gene signature discriminating tumors with fairly well histological differentiation and a favorable prognosis against advanced and poorly differentiated tumors with a dismal outcome. Many of the genes in this signature group are those, which have also been found to be functionally important in the different pathways delineated above. With this gene signature molecular classification of hepatoblastomas should become possible after thorough clinical testing.


ratory evaluations for suspected hepatoblastoma include blood count, liver function tests, lactate dehydrogenase, alpha-fetoprotein (AFP), beta-human chorionic gonadotropin, ferritin, carcinoembryonic antigen, neuron-specific enolase, and catecholamines, as well as hepatotrophic viral titers.1,9 AFP is the most important marker for hepatoblastoma; it is increased in 90% of patients with the tumor. Several investigators have shown that most hepatoblastomas with low AFP levels (⬍100 ng/mL) are aggressive and associated with a poor prognosis. In neonates the interpretation of AFP measurements is more difficult because of the naturally high serum levels in infants. Also, AFP may be elevated in some patients with benign liver tumors, including mesenchymal hamartoma and adenoma.1

Imaging Abdominal ultrasound is the technique of choice as the initial diagnostic procedure for suspected liver tumors. It can be used to identify the liver as the organ of tumor origin and it is particularly useful to show the relationship of the hepatic vessels to the tumor, as well as vessel invasion. Computed tomography (CT) is essential for the evaluation of pulmonary metastasis and can further assess the primary liver tumor and the lymph node status. Here, a CT angiogram with intravenous contrast can visualize arterial, venous, and portal systems in the liver. Despite improvements of the technique in CT, however, the best modality in many centers for revealing morphologic details and to differentiate liver tumors is the magnetic resonance imaging (MRI). Although in most cases of hepatoblastoma sedation or anesthesia is needed for the procedure, we routinely recommend MRI for optimal demonstration of the tumor (Figure 1). With suitable MRI techniques, a clear picture of the tumor’s morphology, its relation to the surrounding structures and vessels, and the lymph node status can be achieved. In our own recent experience, most hepatoblastomas are positive on fludeoxyglucose-positron emission tomography (PET) before therapy and therefore this nuclear scan may be helpful for detection of metastasis and assessment of tumor viability. However, there currently exist no large clinical studies on the use of fludeoxyglucose-PET in hepatoblastoma, so the value of this technique for diagnosis of childhood liver tumors is not yet established. The combination with CT as PET-CT may prove to be more helpful in the future.19

Diagnosis and staging Tumor biopsy Clinical symptoms The clinical presentation of hepatoblastoma as an asymptomatic abdominal mass, sometimes accompanied with fever, fatigue, anorexia, and weight loss, is well established. Severe symptoms, including obstructive jaundice and tumor rupture, are relatively rare occurences.1 Appropriate labo-

For histologic confirmation of the diagnosis, a biopsy is usually taken in patients who do not undergo primary tumor resection. A biopsy can be accomplished by laparotomy, laparoscopy, or by an image-guided percutaneous core needle biopsy in most cases. Fine-needle aspirate with cytologic diagnosis is often not diagnostic. Because core needle biopsies can have compli-

24 cations, such as bleeding and tumor rupture in rare cases, tumors in children 6 months to 3 years of age with a highly elevated serum-AFP (⬎1000 ng/mL and ⬎3⫻ normal range for age) may be clinically diagnosed and treated as hepatoblastoma in the German GPOH group.9 Following this policy, by 2003 57 children treated since 1994 had been diagnosed successfully via the use of these guidelines. This approach must be strictly limited, however, to the small group of children who meet these criteria.

Seminars in Pediatric Surgery, Vol 21, No 1, February 2012 Table 2

COG (Evans) staging system for hepatoblastomas1

Stage I: complete gross resection at diagnosis with clear margins Stage II: complete gross resection at diagnosis with microscopic residual disease at the margins of resection Stage III: biopsy only at diagnosis, or gross total resection with nodal involvement or tumor spill or incomplete resection with gross intrahepatic disease Stage IV: metastatic disease at diagnosis COG, Children’s Oncology Group.

Tumor staging Precise tumor staging is essential for accurate risk stratification and planning of therapy of hepatoblastoma. The Children’s Oncology Group (COG; http://www.childrensoncologygroup. org) in the United States has historically used Evans stages

Figure 1 MRI of an extended hepatoblastoma (PRETEXT III) in an 8-month-old child before (A) and after (B) neoadjuvant chemotherapy.

I—IV (Table 2) for risk stratification, which are based on the findings and results of initial surgery.1 In contrast, the PRETEXT (pretreatment extent) staging system was devised by the European SIOPEL group (Société Internationale Oncologie Pédiatrique/International Society for Pediatric Oncology—Epithelial Liver Tumors; http://www.siopel. org) to categorize the tumor by imaging before initiation of therapy. In the meantime, also the Japanese Pediatric Liver Tumor (JPLT) and German (GPOH; http://www.kinderkrebsinfo.de) study groups have adapted this system and in COG it is used to define surgical resectability.1 In the PRETEXT system, a tumor is classified according to the number of anatomic sections involved by the tumor at the time of diagnosis.20 Thus, PRETEXT I tumors involve 1 section and 3 contiguous sections are free, PRETEXT II tumors involve 1 or 2 sections and leave free 2 contiguous sections; PRETEXT III tumors involve 2 or 3 sections, leaving free only one section; and PRETEXT IV tumors involve all 4 sections (Figure 2). Macrovascular involvement can be documented for portal veins with a “P” (P1 for

Figure 2

PRETEXT staging system.

von Schweinitz

Hepatoblastoma: Recent Developments in Research and Treatment

one branch, P2 for both branches) and for hepatic veins and vena cava with “V” (V1— one, V2—2, V3—3 veins and/or vena cava). Involvement of the caudate lobe is documented with “C,” multifocal tumor with “F” (F0 — unifocal, F1— multifocal), extrahepatic tumor with “E” (E1— direct extension, E2—peritoneal nodes), tumor rupture with “H,” and additional ascites with “a.” Finally, distant metastasis are documented with “M” (p—lungs, s—skeleton, c— brain, m— bone marrow) and positive lymph nodes with “N” (N1—abdominal, N2— distant).21 Over the years, the PRETEXT staging system has proven to be practical for individual tumor classification and prognostically highly relevant. The same system can be used after neoadjuvant chemotherapy as POSTTEXT reclassification.

Table 3 Current risk stratification of hepatoblastoma staging groups COG Very low risk Low risk

Stage I, PFH Stage I non pure fetal and non small cell undifferentiated histology or stage II non small cell undifferentiated histology (SCU) Intermediate-risk Stage I or II ⫹ small cell undifferentiated histology or stage III High risk Any stage IV or any stage ⫹ serumAFP ⬍100 ng/mL SIOPEL ⫹ GPOH Standard-risk PRETEXT I, II or III, AFP ⬎100 ng/mL, no additional PRETEXT criteria High-risk PRETEXT IV, AFP ⬍100 ng/mL, any PRETEXT and additional PRETEXT: criteria positive (E, H, M, N, P2, V3) JPLT Stratification according to PRETEXT PRETEXT I, no additional criteria PRETEXT II, no additional criteria PRETEXT III/IV or any PRETEXT ⫹ additional criteria (E, H, M, N, P, V)

Treatment Risk stratification and treatment strategies Marked differences exist in risk stratification and treatment strategies for hepatoblastoma between the different study groups. In the United States, the protocol of the current COG study AHEP 0731 recommends initial surgery for all children with a liver tumor. A primary resection should be undertaken for limited PRETEXT I and II tumors with at least 1 cm of clear margins, whereas those tumors with larger extension (PRETEXT III, IV), vascular invasion or distant metastasis should be treated with neoadjuvant chemotherapy. The result of primary surgery determines the tumor’s stage according to the Evans system. All patients are treated with adjuvant chemotherapy according to POSTTEXT classification.1 The only exemption are patients with a completely resected (stage I) hepatoblastoma with pure fetal histology who do not receive any chemotherapy, because these were found to have a 100% cure rate with surgery alone.22 Thus, the COG study AHEP 0731 intends to differentially adapt treatment by stratifying patients into 4 different risk groups reducing the intensity of chemotherapy in approximately 30% of the patients. These risk groups are (Table 3): ● ●

● ●

very low risk group: COG-stage I, pure fetal histology; low risk: COG-stage II, non pure fetal histology and COG-stage II, all non small cell undifferentiated histology; intermediate risk: stage I/II small cell undifferentiated histology or stage III any histology; and high risk: any stage IV or any stage and serum-AFP ⬍100 ng/mL.

Thus, the known poor prognostic factors, small cell undifferentiated histology1 and low serum-AFP ⬍100 ng/mL, are also included in this stratification. In contrast to the United States, in Europe and attached regions, the international SIOPEL group and the German


AFP, alpha-fetoprotein; COG, Children’s Oncology Group; PFH, pure fetal histology; SIOPEL ⫹ GPOH, Société Internationale Oncologie Pédiatrique/International Society for Pediatric Oncology—Epithelial Liver Tumors and German Society for Pediatric Oncology, JPLT, Japanese Pediatric Liver Tumor Group.

GPOH group do not recommend primary surgery and resection of hepatoblastoma; this recommendation is based on the response rate of approximately 90% for these tumors to neoadjuvant therapy, which not only makes the tumors smaller and more solid but potentially can also suppress occult micrometastases without delay. When one uses this strategy, resections of hepatoblastoma become easier and safer in a setting in which many centers with different expertise treat these patients. Finally, the GPOH group observed that hepatic surgery alone with induction of liver regeneration may also promote the growth of residual tumor and metastases not pretreated with chemotherapy by secretion of so-called liver growth factors.9 Therefore, neoadjuvant chemotherapy is recommended for more or less all hepatoblastomas. The SIOPEL and GPOH groups currently stratify hepatoblastomas into 2 risk groups (Table 4)20: ● ●

standard risk: PRETEXT I, II or III, no additional PRETEXT criteria, serum-AFP ⬎100 ng/mL; and high risk: PRETEXT IV, PRETEXT I, II or III with P⫹, V⫹, E⫹, H⫹, M⫹, N⫹, respectively and all patients with serum-AFP ⬍100 ng/mL.

In the Japanese JPLT group, PRETEXT I hepatoblastomas without other PRETEXT criteria are primarily resected, and all others pretreated with 2-3 courses of chemotherapy according to tumor extension and response.23 A further risk group stratification has not been published by this group.

26 Table 4

Seminars in Pediatric Surgery, Vol 21, No 1, February 2012 Current chemotherapy recommendations of the hepatoblastoma study groups

Study group

Patient/SR group




Very low risk Low risk Intermediate risk

None Adj. CDDP–5-FU–VCR ⫻ 2 Neoadj. CDDP–5FU–VCR –Doxo ⫻ 4-6 Adj. CDDP–5FU–VCR –Doxo ⫻ 2 Neoadj. VCR–Irinot. ⫻ 2 (window) ⫹ CDDP–5FU–VCR–Doxo ⫻ 6 ⫹ If response to window: ⫹ VCR–Irino ⫻ 2 Neoadj. CDDP ⫻ 4 Adj. CDDP ⫻ 2 Randomized ⫾ STS Neoadj. CDDP ⫻ 4 alternating with Carbo/Doxo ⫻ 3 Adj. Carbo/Doxo ⫻ 2 alternating with CDDP ⫻ 1 Neoadj. CDDP–Doxo ⫻ 2-3 Adj. CDDP–Doxo ⫻ 1 Neoadj. CDDP ⫻ 4 alternating with Carbo/Doxo ⫻ 3 Adj. Carbo/Doxo ⫻ 2 alternating with CDDP ⫻ 1 Adj. CDDP/Pira ⫻ 4 (50% dose) Neoadj. CDDP/Pira ⫻ 2 Adj. CDDP/Pira ⫻ 4 (50% dose) Neoadj. CDDP/Pira ⫻ 3-4 or CDDP/Pira ⫻ 2 ⫹ Ifo/ Carbo/Pira/Eto ⫻ 1-2 Adj. CDDP/Pira ⫻ 2 or Ifo/Carbo/Pira/Eto ⫻ 2 Adj additional high-dose Eto/Carbo/melphalan

Study AHEP-0731

High risk


Standard risk

High risk GPOH

Standard risk High risk



Study SIOPEL-6

Interim phase recommendation (from SIOPEL-3 HR) Interim phase recommendation Interim phase recommendation (from SIOPEL-3 HR) Study JPLT-2 (1999-2008)

Carbo, carboplatin; CDDP, cisplatin; COG, Children’s Oncology Group; Doxo, doxorubicin; Eto, etoposide; 5-FU, 5-fluorouracil; GPOH, German Society for Pediatric Oncology; Ifo, ifosfamide; Irinot,irinotecan; Pira, pirarubicin; SIOPEL, Société Internationale Oncologie Pédiatrique/International Society for Pediatric Oncology—Epithelial Liver Tumors; STS, sodium-thiosulfate; VCR, vincristine.

Tumor resection A complete surgical resection remains the mainstay of hepatoblastoma treatment. The critical aspects of surgical techniques for hepatoblastoma resection are well described (Figure 3)9 and where recently described in detail.24 Although the techniques have not changed, increasing evidence suggests that resection margins of only a few millimeters for hepatoblastoma after good response to chemotherapy are sufficient and that microscopic residual tumor in these cases does not necessarily warrant a reresection.24 With this finding, conventional resections of chemotherapeutically pretreated hepatoblastomas with central localization25 or POSTTEXT III and IV extension26 became feasible and have rendered excellent treatment results. This again renews the discussion on what are the indications for orthotopic liver transplantation (OLT).

Liver transplantation Recent experience has shown that OLT is a good option for treatment of unresectable hepatoblastoma resulting in cure for the majority of patients. Recently, the study groups COG, SIOPEL, and GPOH have developed common guidelines for OLT in hepatoblastoma.1,27,28 According to these, multifocal hepatoblastoma in all 4 liver sections (PRETEXT IV) is a clear indication for OLT because total eradication of

all intrahepatic tumor metastases by chemotherapy is unlikely and recurrence rates after conventional resection are high. Patients with solitary PRETEXT IV hepatoblastomas that are not clearly downstaged to PRETEXT III are also recommended to receive OLT. Hepatoblastoma with portal vein involvement may be treated best with OLT, especially in cases of involvement of both branches and the bifurcation, as micrometastasis may persist after neoadjuvant chemotherapy. A comparative study of OLT versus conventional resection has not been performed. Patients with portal vein invasion in the recent GPOH trial HB99 had the same prognosis as other high-risk patients.29 Hepatoblastoma with involvement of all 3 hepatic veins (V3) can also be an indication for OLT because the venous resection will be more extensive than with conventional resection. Few clear data exist to direct therapy in these patients. Central hepatoblastoma may be appropriate for OLT if a conventional resection25 does not seem feasible. Finally, intrahepatic residual or relapsed hepatoblastoma can be an indication for OLT, although gross residual tumor at resection should be avoided because rescue OLT has a worse outcome than primary OLT.1,27,28 The indication for OLT in nonresectable hepatoblastoma with pulmonary metastases remains controversial. Good results have been achieved in patients, with lung nodes entirely eradicated by chemotherapy according to imaging

von Schweinitz

Hepatoblastoma: Recent Developments in Research and Treatment

Figure 3 Resection of a hepatoblastoma of an 11- month-old child in the right liver lobe after meticulous preparation of vessels using an electric knife without clamping of the left vessels. A, before resection; B, resection almost finished. (Color version of figure is available online.)

with multislice thoracic CT or surgery, but the recurrence rate of metastases in these patients may be greater than anticipated. Therefore, posttransplantation chemotherapy should be administrated in cases with pulmonary metastases, but valid studies of the results are still missing.28 To learn more about children with OLT for liver tumors, an international electronic registry, the Pediatric Liver Unresectable Tumor Observatory (PLUTO, http://pluto.cineca. org) has been established, which collects detailed clinical data of these patients. Currently, 49 patients with hepatoblastoma have been registered, 43 of whom (87%) are tumor-free.30 Further registration of patients is encouraged and it is hoped that valid data can be derived from this registry soon.

Chemotherapy Experience demonstrates that most hepatoblastomas respond well to chemotherapy and that cisplatin (CDDP) is the most effective agent against this tumor. This agent has been used in the large studies of all the cooperative group trials.1,9,31 Combination regimes that include CDDP result in response rates of up to 93%. The SIOPEL group could demonstrate in a randomized trial that results of standard risk hepatoblastoma treated with 6 courses of CDDP alone


are equal to those treated with CDDP and doxorubicin.32 A further intensification of the alternating CDDP and carboplatin/doxorubicin chemotherapy by weekly application in the last SIOPEL-4 study resulted in another improvement of short-term survival of high-risk patients (event-free survival 77%, overall survival 84%),33 but a longer follow-up of 3-5 years is lacking. The GPOH experience showed that the application of high-dose therapy with carboplatin and etoposide with autologous stem cell transplantation produces equal but not superior results in high-risk patients than the condensed application of conventional dose CDDP containing chemotherapy given over 7 courses (publication in preparation). On the basis of their own and reported experience, the study groups developed their present treatment recommendations (Table 4). In the COG study AHEP 0731, very low-risk patients do not receive any chemotherapy, and low-risk patients receive only 2 courses of CDDP (100 mg m⫺2), 5-fluorouracil (600 mg/m2), and vincristine (1.5 g m⫺2), ie, C5V, postoperatively to reduce toxicity for these favorable outcome patients because the addition of amifostine did not reduce CDDP-induced hearing loss in the previous COG study P9645.1 Intermediate-risk patients receive neoadjuvant chemotherapy with 2 to 4 courses C5V plus doxorubicin (30 mg/m2, C5VD) and 2 further courses after resection or OLT. In high-risk patients, an upfront window therapy with 2 courses of vincristine (1.5 g/m2) and irinotecan (50 mg/m2) is applied and in case of response followed by 6 courses of C5VD with 1 course of vincristine and irinotecan in between each 2 courses block, whereas in nonresponders to the up-front window therapy only 6 courses of C5VD are given. In these patients, the tumor resection or OLT should be performed after 4 courses of C5VD. The SIOPEL group currently conducts the SIOPEL-6 study for standard-risk hepatoblastoma. The chemotherapy in this protocol is identical to the 6 courses of the CDDP (80 mg/m2) monotherapy arm of the previous trial SIOPEL-3 standard-risk study, but the patients are randomized for the additional administration of sodium thiosulfate (STS). This study shall assess the efficacy of STS to reduce hearing loss which is the most frequent toxic side effect of CDDP and will monitor any potential impact of STS on the response to CDDP.31 Because the results of the recently closed SIOPEL-4 study on high-risk hepatoblastoma are not yet valid and the intensified weekly application of alternating CDDP and carboplatin/doxorubicin courses proved to be very toxic, the SIOPEL group decided to recommend chemotherapy according to the SIOPEL-3 high-risk study34 for an interim phase until a new high-risk hepatoblastoma trial can be inaugurated. Therefore, high-risk patients are treated with CDDP (80 mg/m2), alternating every 2 weeks with carboplatin (500 mg/m2) plus doxorubicin (60 mg/m2) for 7 neoadjuvant and 3 adjuvant courses. After closing the study HB99, the GPOH group decided to join the SIOPEL group for future multiinstitutional stud-


Seminars in Pediatric Surgery, Vol 21, No 1, February 2012

ies and gave recommendations for hepatoblastoma treatment in GPOH centers for the present interim phase. For standard-risk hepatoblastoma, the group did not want to adapt the regime of the recent SIOPEL studies with CDDP as a monotherapy over 6 courses but rather suggested the well-tested combination of CDDP (100 mg/m2) and doxorubicin (60 mg/m2) for a maximum of 4 courses to reduce CDDP toxicity while fully preserving the therapeutic effect and avoiding the development of drug resistance. For highrisk patients, the GPOH group also proposes the regime of the SIOPEL-3 high-risk study (see above in this section),34 thus avoiding the side effects of high-dose chemotherapy with autologous stem cell transplantation of the former GPOH HB99 regime.9 The JPLT group treated patients with a PRETEXT I tumor and no additional PRETEXT criteria with 4 courses of adjuvant chemotherapy containing CDDP (40 mg/m2) and pirarubicin (15 mg/m2) in the study JPLT-2. In addition, PRETEXT II tumors received 2 of these courses in double dosage as neoadjuvant therapy. The same accounts for PRETEXT III and IV tumors, as well as hepatoblastomas with other PRETEXT criteria for 3 neoadjuvant courses. If the tumor was not responsive, the chemotherapy was switched to the combination of ifosfamide (3 g/m2), carboplatin (400 mg/m2), pirarubicin (30 mg/m2), and etoposide (100 mg/m2) to achieve resectability. After resection, the neoadjuvant regime was given for another 2 courses, and patients with small cell undifferentiated histology tumors and/or distant metastasis were additionally treated with high-dose chemotherapy with etoposide (200 mg/m2) carboplatin (400 mg/m2), and melphalan (90 mg/m2).23 This study was terminated in December 2008, but a new strategy and chemotherapy regime has not yet been published by the JPLT group. New cytotoxic drugs are required for successful treatment of resistant or recurrent heptatoblatoma. New drugs have been tested preclinically and in single patients, but options are limited.35 A recent COG phase 2 trial with oxaliplatin did not reveal an advantage over the conventionally used CDDP.36 In preclinical tests topotecan and paclitaxel have shown some activity against hepatoblastoma, but they have not been thoroughly tested in clinical trials.35 The same is true for substances to overcome drug resistance (ie, verapamil) or agents to enhance the efficacy of cytotoxic drugs. The latter may be achieved using acetaminophen’s toxicity with delayed N-acetylcysteine rescue, which has been used in selected patients, but still has to be tested in larger trials.37 Another approach for high-risk hepatoblastoma patients to prevent recurrent disease would be to apply long-term maintenance chemotherapy over 1-2 years after achieving remission. Irinotecan has been successfully used for this in single patients,38 so this approach should be evaluated in future multiinstitutional trials.

possible to influence tumor growth with substances precisely targeted at growth factor receptors or key molecules of altered intracellular signal transduction pathways. Thus, several tyrosine-kinase targeting drugs may have an effect on hepatoblastoma, such as sorafenib, which was shown to inhibit growth of hepatocellular carcinoma in children,39 trastuzumab (Herceptin) against the erbB2-receptor, and sunitinib (Sutent) against a variety of growth factors and their receptors, including HGF/c-met and IGF-2. The altered WNT and HH pathways seem to be good targets for new substances. In early experiments, we found agents that have an effect against the WNT pathway (epigallocatechingaleate) and against the HH pathway (cyclopamine, betulinic acid) to inhibit hepatoblastoma growth in vitro and in vivo.11 Further research on drugs against the WNT pathway molecules is in progress,40 whereas the HH pathway antagonists are already in clinical trials.41 We recently also found sirolimus (rapamycin) to act against hepatoblastoma cells (publication in preparation); therefore, this drug may be a good candidate for a long-term maintenance therapy, especially in patients after OLT. Finally, already identified differentiating substances (retinoic acid) and apoptotic agents (arsenic trioxide; ABT-737) have very recently been shown to inhibit hepatoblastoma growth.42,43

Targeted agents The recent advances in molecular research on hepatoblastoma (see molecular biology) indicate that it should be

Alternative modalities for local treatment The developments concerning alternative techniques to achieve local control of hepatoblastoma recently have been summarized in detail.44 The most promising approach is hepatic artery chemoembolization (HACE, also called transarterial chemoembolization—TACE), which has been performed in the last several decades in single patients or small series. Different cytotoxic drugs, mostly CDDP and doxorubicin, have been mixed either with water-soluble radiographic contrast medium or with ethiodized oil (Lipoidol). The procedure is completed by embolization of the feeding arteries of the tumor with gelatine foam or stainless steel coils. Hepatoblastomas often respond well to this therapy, in addition to systemically pretreated tumors. However, this technique is feasible only in cases in which both branches of the hepatic artery are not involved. The rate of complications is substantial, with pain, nausea, and fever in most patients, sometimes tumor lysis syndrome or Lipoidol embolization into the lungs, which may be fatal. Taken together, HACE is a good technique for localized hepatoblastoma for increasing respectability or for bridging time until OLT. Other transarterial techniques, such as embolization without chemotherapy and radio-embolization with yttrium-90 microspheres (selective internal radiation) have been applied in very few cases. The same is true for portal vein embolization of the ipsilateral portal vein branch to induce growth in the remnant liver before an extended resection, which normally is not necessary in children and might also promote tumor growth.44 Also,

von Schweinitz

Hepatoblastoma: Recent Developments in Research and Treatment

percutaneous tumor ablation with radiofrequency ablation, ethanol injection, cryoablation, laser, or microwave ablation, which are commonly applied in adults, are rarely indicated for hepatoblastoma. These techniques work only in small lesions and often do not completely eradicate all tumor cells, so they usually are only appropriate in a palliative situation or in case of limited local recurrence in combination with contraindications against a reresection.44

Prognosis and prognostic factors The prognosis of children with hepatoblastoma was improved substantially during the past few years through the efforts of all the cooperative study groups. Thus, limited hepatoblastomas, such as Evans stage I and II (COG), PRETEXT I/II (JPLT), or standard-risk (SIOPEL, GPOH) tumors have an event-free survival (EFS) of 80%-90% and an overall survival (OS) of 85%-100%.1,22,23 More extended but localized hepatoblastomas of COG stage III or PRETEXT III without further negative prognostic factors (intermediate or standard-risk) currently can reach an EFS of 65%-75% and an OS of 70%-80%.1 However, high-risk hepatoblastoma with extension over the complete liver (PRETEXT IV), especially with multifocal lesions, those with extrahepatic tumor, macroscopic invasion of large vessels, or distant or lymph node metastases, have a much worse prognosis, ie, an EFS of 45%-65% and an OS of 50%-65%.1,23,29,34 Among these children there seems to exist a group of very high-risk tumors with a serum-AFP ⬍100 mg/mL,1,45 small cell undifferentiated histology,1,7 distant metastases, and macroscopic invasion of the major hepatic veins and/or the vena cava.29 Thus, all study groups have stratified the patients into risk groups to tailor therapy (see Risk stratification and treatment strategies). Because analyses of patients groups have found other prognostic criteria, which may also be associated with poor outcome,9 it would be helpful to analyze these in a large number of patients with the goal to better refine treatment. Therefore, the 4 study groups initiated the cooperative project CHIC (Childhood Hepatic Tumors International Collaboration) in which the relevant data from the past cooperative trials will be analyzed for identification of further prognostic factors. This will hopefully lead to a more refined and internationally uniform risk stratification of hepatoblastoma patients in combined treatment trials.1

References 1. Meyers RL, Aronson DC, von Schweinitz D, et al. Pediatric liver tumors, in Pizzo PA, Poplack DG (eds): Principles and Practice in Pediatric Oncology. Philadelphia, PA, Wolters Kluwer, Lippincott. Williams Wilkins, 2011, 838-60. 2. Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011.


3. McLaughlin CC, Baptiste MS, Schymura MJ, et al. Maternal and infant birth characteristics and hepatoblastoma. Am J Epidemiol 2006; 163:818-28. 4. Birch JM. Epidemiology of pediatric liver tumors, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 15-26. 5. Zimmermann A. Ontogenetic aspects of liver tumors, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 43-51. 6. Zimmermann A, Lopez-Terrada D. Pathology of pediatric liver tumors, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 83-112. 7. Zimmermann A. The emerging family of hepatoblastoma tumours: From ontogenesis to oncogenesis. Eur J Cancer 2005;41:1503-14. 8. Trobaugh-Lotrario AD, Tomlinson GE, Finegold MJ, et al. Small cell undifferentiated variant of hepatoblastoma: Adverse clinical and molecular features similar to rhabdoid tumors. Pediatr Blood Cancer 2009;52:328-34. 9. von Schweinitz D. Management of liver tumors in childhood. Semin Pediatr Surg 2006;15:17-24. 10. Wang LL, Filippi RZ, Zurakowski D, et al. Effects of neoadjuvant chemotherapy on hepatoblastoma: A morphologic and immunohistochemical study. Am J Surg Pathol 2010;34:287-99. 11. Kappler R, von Schweinitz D. Molecular aspects of hepatoblastoma, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 27-42. 12. Arai Y, Honda S, Haruta M, et al. Genome-wide analysis of allelic imbalances reveals 4q deletions as a poor prognostic factor and MDM4 amplificatiopn at 1q32.1 in hepatoblastoma. Genes Chrom Cancer 2010;49:596-609. 13. Li YC, Deng YH, Guo ZH, et al. Prognostic value of hedgehog signal component expressions in hepatoblastoma patients. Eur J Med Res 2010;15:468-74. 14. Pietsch T, Steiger B, Leuschner I, et al. Expression and function of the receptor tyrosine kinase erbB2 in hepatoblastomas. Monatsschr Kinderheilkd 2011;159:400. 15. Litten J, Chen T, Schultz R, et al. Activated NOTCH2 is overexpressed in hepatoblastomas: An immunohistochemical study. Pediatr Dev Pathol 2011; Epub ahead of print. 16. Turato C, Buendia MA, Fabre M, et al. Over-expression of SERPINB3 in hepatoblastoma: A possible insight into the genesis of this tumour? Eur J Cancer 2011; Epub ahead of print. 17. Von Frowein J, Pagel P, Kappler R, et al. MicroRNA-492 is processed from the keratin 19 gene and up-regulated in metastatic hepatoblastoma. Hepatology 2011;53:833-42. 18. Cairo S, Armengol C, De Reyniès A, et al. Hepatic stem-like phenotype and interplay of Wnt/beta-catenin and Myc signaling in aggressive childhood liver cancer. Cancer Cell 2008;14:471-84. 19. Franzius C. FDG-PET/CT in pediatric solid tumors. Q J Nucl Med Mol Imaging 2010;54:401-10. 20. Roebuck DJ. Imaging and staging of pediatric liver tumors, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 65-82. 21. Roebuck DJ, Aronson DC, Clapuyt P, et al. PRETEXT: A revised staging system for primary malignant liver tumours of childhood developed by the SIOPEL group. Pediatr Radiol 2005;2007:123-32. 22. Malogolowkin MH, Katzenstein HM, Meyers RL, et al. Complete surgical resection is curative for children with hepatoblastoma with pure fetal histology: A report from the Children’s Oncology Group. J Clin Oncol 2011;29:3301-6. 23. Hishiki T, Matsunaga T, Sasaki F, et al. Outcome of hepatoblastomas treated using the Japanese Study Group for Pediatric Tumor (JPLT) protocol 2: Report from the JPLT. Pediatr Surg Int 2011;27:1-8. 24. Czauderna P, von Schweinitz D. Surgical treatment, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 113-32.


Seminars in Pediatric Surgery, Vol 21, No 1, February 2012

25. Guérin F, Gauthier F, Martelli H, et al. Outcome of central hepatectomy for hepatoblastomas. J Pediatr Surg 2010;45:555-63. 26. Lautz TB, Ben-Ami T, Tantemsapya N, et al. Successful nontransplant resection of POST-TEXT III and IV hepatoblastoma. Cancer 2011; 117:1976-83. 27. Meyers RL, Otte J-B. Liver transplantation for unresectable liver tumors in children, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 133-52. 28. Gupta AA, Gerstle JT, Ng V, et al. Critical review of controversial issues in the management of advanced pediatric liver tumors. Pediatr Blood Cancer 2011;56:1013-8. 29. Haeberle B, Schmid I, von Schweinitz D. Is there a very high risk group in childhood hepatoblastoma? A subgroup analysis of the GPOH study HB99. Pediatr Blood Cancer 2010;55:815. 30. Otte JB, Meyers R. PLUTO first report. Pediatr Transplant 2010;14: 830-5. 31. Perilongo G, Morland B, Malogolowkin M. Chemotherapy for childhood hepatoblastoma and hepatocellular carcinoma, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 153-64. 32. Perilongo G, Maibach R, Shafford E, et al. Cisplatin versus cisplatin plus doxorubicin for standard-risk hepatoblastoma. N Engl J Med 2009;361:1662-70. 33. Zsiros J, Brock P, Brugières L, et al. High cure rate in high risk hepatoblastoma with dose intensive cisplatin based chemotherapy— Results of the SIOPEL-4 trial. Pediatr Blood Cancer 2010;55:816. 34. Zsiros J, Maibach R, Shafford E, et al. Successful treatment of childhood high-risk hepatoblastoma with dose-intensive multiagent chemotherapy and surgery: Final results of the SIOPEL-3HR study. J Clin Oncol 2010;28:2584-90.

35. Zsiros J, Brugieres L. Salvage strategies, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 165-76. 36. Beaty O, Berg S, Blaney S, et al. A phase II trial and pharmacokinetic study of oxaliplatin in children with refractory solid tumors: A Children’s Oncology Group study. Pediatr Blood Cancer 2010;55:440-5. 37. Neuwelt AJ, Wu YJ, Knap N, et al. Using acetaminophen’s toxicity mechanism to enhance cisplatin efficacy in hepatocarcinoma and hepatoblastoma cell lines. Neoplasia 2009;11:1003-11. 38. Qayed M, Powell C, Morgan ER, et al. Irinotecan as maintenance therapy in high-risk hepatoblastoma. Pediatr Blood Cancer 2010;54: 761-3. 39. Schmid I, Haeberle B, Corbacioglu S, et al. First experience with sorafenib and cisplatin/doxorubicin (PLADO) in pediatric hepatocellular carcinoma. Pediatr Blood Cancer 2011; Epub ahead of print. 40. Takahashi-Yanaga F, Kahn M. Targeting Wnt signaling: Can we safely eradicate cancer stem cells? Clin Cancer Res 2010;16:3153-62. 41. Ng JM, Curran T. The Hedgehog’s tale: Developing strategies for targeting cancer. Nat Rev Cancer 2011;11:493-501. 42. Aktas S, Zadeoglulari Z, Ercetin P, et al. The effect of differentiating and apoptotic agents on notch signalling pathway in hepatoblastoma. Hepato Gastroenterol 2010;57:891-8. 43. Lieber J, Eicher C, Wenz J, et al. The BH13 mimetic ABT-737 increases treatment efficiency of paclitaxel against hepatoblastoma. BMC Cancer 2011;11:362. 44. Roebuck DJ. Alternative approaches for treatment, in Zimmermann A, Perilongo G, Malogolowkin M, von Schweinitz D (eds): Pediatric Liver Tumors. Heidelberg, Springer, 2011, 177-88. 45. De Ioris M, Brugières l, Zimmermann A, et al. Hepatoblastoma with a low serum alpha-fetoprotein level at diagnosis: The SIOPEL group experience. Eur J Cancer 2008;44:545-50.