Childhood primary tumors of the liver are rare. Of these, about two thirds are benign and one third malignant (1). The incidence varies from geographic region to region. According to the Surveillance, Epidemiology and End Results (SEER) data, between 1973 and 1977 and 1993 and 1997, hepatoblastoma rates increased (from 0.6 to 1.2/1,000,000), while hepatocellular carcinoma rates decreased (from 0.45 to 0.29/1,000,000) (2). The most common tumors according to histopathology are hepatoblastoma (HBL), hepatocellular carcinoma (HCC), benign vascular tumor, mesenchymal hamartoma, and sarcoma (Table 14.1) (1). Benign tumors are usually found incidentally as an asymptomatic abdominal mass. Malignant tumors usually present with abdominal distension with or without a palpable abdominal mass. Children with both HBL and HCC may also present with weight loss, fever, and anorexia. Anemia, thrombocytopenia, and leukocytosis are sometimes associated findings.

Complete surgical resection is the standard curative treatment for malignant tumors. Chemotherapy, radiotherapy, and local ablation treatment are adjunctive to surgery or palliative for incurable patients.

Computed tomography (CT) and magnetic resonance imaging (MRI) are the most definitive diagnostic tests and can often help to differentiate benign from malignant tumors preoperatively.

Children with primary malignant liver tumors (PMLTs) often present with an abdominal mass or generalized abdominal enlargement. The child may have pain localized to the right upper quadrant, fever, anorexia, weight loss, jaundice, or vomiting. The first presentation may, on occasion, be an acute abdominal crisis caused by tumor rupture and hemoperitoneum. Paraneoplastic syndrome, such as precocious sexual development, has been associated with PMLTs (10). Anemia and thrombocytosis often occur. This is probably related to the ability of HBL cells to secrete gonadotropin, interleukin-1B, which induces IL-6 production in fibroblasts of endothelial cells (11,12).

If a hepatic or biliary tumor produces biliary obstruction, there may be elevations of serum hepatic enzymes or bilirubin (36,37). An abdominal ultrasound examination helps to establish the presence of a hepatic mass and differentiates cystic from solid lesions. Ultrasound permits evaluation of the adrenals and the kidneys, helps to exclude these organs as possible primary sites of tumor metastatic to the liver, and allows evaluation of the inferior vena cava for tumor thrombus. Color Doppler ultrasound may be performed to demonstrate the relationship of the tumor to hepatic vessels (38). The technetium-90m sulfur colloid scan is of some value in localizing the tumor to the liver and defining its boundaries. It is a sensitive test, but not specific. CT and MRI provide superior delineation of the mass and show evidence of multifocality (39). On nonenhanced CT, epithelial HBL is a homogeneous low-attenuation mass. The mixed type is inhomogeneous. Calcifications can be present. Contrast enhancement of the periphery or septa may be seen. On MRI, the epithelial type is homogeneous and hypointense on T1 and hyperintense on T2. The mixed type may be more inhomogeneous and hypointense on T1 and T2, and hemorrhage will show high signal intensity on T1 and T2 (Fig. 14.4). On CT, HCC will show solitary or multiple masses. Calcification can occur. The tumor rim may show enhancement with contrast. On MRI, the T1 images may be isointense or hyperintense. On T2 images, a mosaic pattern (caused by necrosis, hemorrhage, septa, and fatty metaplasia) is common (29). The operating surgeon may request an angiogram to obtain information about the origin and distribution of the right and left hepatic arteries to provide additional information on the vascular supply of the tumor (10). Routine chest radiographs and thoracic CT scans are necessary because 10-20% of patients with HBL and 30% of patients with HCC have lung metastasis at the time of diagnosis (12,18,40,41). Positron emission tomography CT holds promise for better surgical outcomes and greater sensitivity for residual or relapsed disease (42).

A valuable test in evaluating PMLT is the serum AFP. The clinician should note that AFP, synthesized by the fetal liver, may be elevated in pregnant women. AFP levels are high in the normal newborn infant but drop rapidly by the age of 1 month. AFP should be barely detectable by the age of 2 years (18). Approximately two thirds of children with HBL and HCC have
an elevated AFP (37). However, the clinician should be cautioned that AFP is not always elevated in HBL or HCC and that this protein may be slightly elevated in benign hamartomas. Absence of AFP elevation may be a poor prognostic sign in HBL; it is associated with the small-cell (anaplastic) histologic type and responds poorly to therapy. For patients with elevated AFP, it may be of value in monitoring the course of therapy (38). Complete resection of HBL or HCC should result in a normal AFP by 2 months postoperatively (17). No change in an elevated AFP after surgery indicates either residual tumor or regenerating normal liver. The diagnosis of growing tumor depends on confirmatory imaging studies. An elevation in the AFP during the follow-up period generally heralds a local tumor recurrence or distant metastasis (37).

Genetic abnormalities are found in HBL. Trisomies 2, 8, and 20 have been reported in tumor samples as well as t (1;4) (q12;34) translocation. The loss of heterozygosity of 11p15 is seen in about one third of cases (12). There is a clear association between HBL and familial adenomatous polyposis (FAP). Germline mutations in APC tumor suppressor gene occur in patients with FAP. One study estimated that 1 in 20 HBLs is probably associated with FAP (43). Increasing data support a role for aberrant transduction of the Wnt/beta-catenin signaling pathway and its molecular targets in HBL tumorigenesis (25,44,45). A study in a mouse model of HBLs induced by toxin exposure detected mutations of the beta-catenin protooncogene in 100% of the tumors analyzed (46). Increasing evidence suggests that HBL is derived from a pluripotent stem cell (19). This supports the hypothesis that such tumor may arise from a developmental error during hepatogenesis. Research particularly focused on these developmental processes governing liver maturation and growth may ultimately lead to more effective targeted therapy for this disease.

There are two major staging systems for PMLT, one is from CCSG and POG (Table 14.3), and the other is from 2002 The American Joint Committee on Cancer (AJCC) TNM staging system (Table 14.4) (48). All are based on resectability and, to a lesser extent, microscopic residual tumor (12,49). Figure 14.5 shows that the International Society of Pediatric Oncology Liver Tumor Study Group (SIOPEL) Pretreatment Extent of Disease (PRETEXT) grouping system divides the liver into four parts (40). The clinician must be wary of the pitfalls in the existing staging systems. There are at least two important drawbacks of some of the systems shown. First, the systems are applied at initial diagnosis. If upfront chemotherapy is given and if it produces a substantive tumor response before surgery, then the initial stage may no longer be pertinent. Second, a surgeon’s determination of resectability is somewhat subjective and varies among practitioners (50).