Hepatoblastoma (HB) is the most common hepatic malignant tumor of childhood, mostly occurred before 5 years of age. Congenital HB (CHB), those detected in utero or diagnosed within the first 28 days after birth, accounts for less than 10% of all pediatric HB.[1–4] Imaging examinations play a vital role in managing pediatric liver tumors, which can be highly suggestive of a specific diagnosis based on sufficient clinical information. Up to now, however, few studies focusing on the radiological characteristics of CHB has been reported. It has not been studied as thoroughly as HB diagnosed at an older age.
Therefore, the purpose of our study was to review the characteristic clinical and radiologic findings of 7 patients with CHB to facilitate noninvasively preoperative diagnosis and avoid misdiagnosis or unnecessary surgical exploration.
2 Materials and methods
This retrospective study was approved by our institutional review board and patient consent was waived. We identified 7 HB patients diagnosed in the neonatal period based on pathological findings from 2012 to 2019 at our hospital through queries of the radiology and pathology search engines. Patient demographic information, clinical signs and symptoms, laboratory tests, and treatment were obtained from the electronic medical records. All patients underwent abdominal contrast-enhanced computed tomography (CT) examinations before surgical resection. Two pediatric radiologists (L. L., with 5 years of experience, and K.J., with 20 years of experience) reviewed CT scans of the eligible patients by consensus, and evaluated the main imaging features regarding hepatic tumor location, size, morphology, margin, enhancement pattern, presence or absence of cystic necrosis, hemorrhage or calcifications, adjacent anatomical structure involvement, and intra-/extra-hepatic metastasis. The tumor size was indicated by the maximum dimension. The radiologists were blinded to the clinical information and pathological diagnosis.
The demographic and clinical characteristics of included patients in this study are shown in Table 1. All the patients (3 males and 4 females) were born at term, with a median age of 13.5 days (range 2–24 days) at diagnosis. Only 1 patient presented with elevated preoperative alpha-fetoprotein (AFP) of 178539.8 ng/mL, and the remaining 6 patients had normal AFP levels (normal range for neonates: 100∼105ng/mL6). Other blood tests including hemoglobin, white blood cells count, and platelet level were relatively nonspecific. Focal abnormality in the liver was detected by ultrasound in 4 patients antenatally, while the liver tumors were found accidentally in the other 3 patients due to respiratory diseases. No patients had a family history of Beckwith–Wiedemann syndrome, hemihypertrophy, or familial polyposis coli.
At the abdominal CT examinations, all the 7 cases of CHB were single solid and the mean tumor size was 4.7 cm (range: 2.9–10.2 cm). Liver S (V∼VII) were most involved, with 2 cases in S VI, 2 in S VII, 1 in S (V + VI), 1in S (II + III), and 1 in S (IVa + V + VIII). The majority (6/7) of tumors were well-circumscribed. 4/7 tumors presented round while 3/7 tumors with extrahepatic extension irregular or lobulated. Microhemorrhage, cystic necrosis, and nodular calcification was seen in 5/7, 4/7, and 1/7 tumors, respectively. All tumors showed inhomogeneously significant enhancement. All tumors presented marked enhancement with multiple nodular or striped appearance in the center and periphery of the tumors on the arterial phase. The enhancement area showed progressive expansion and fusion filling over time but the attenuation gradually declined on the portal and delayed phases, and finally the majority (6/7) of tumors presented multiple band- or island-like characteristics with prominently peripheral enhancement on the delayed phase (Fig. 1) while the remaining 1 relatively small tumor showed nearly complete but inhomogenous enhancement (Fig. 2). No enhancement was seen in the necrotic area. In addition, only 1/7 tumor had hilar hepatic bile duct and portal vein invasion and secondary intra-hepatic bile duct dilation. The abdominal aorta distal to the coeliac trunk was significantly narrowed in 3/7 patients. No metastatic lesions were identified in all patients at diagnosis.
Pathologically, 6/7 tumors revealed fetal histology with only 1/7 containing mesenchymal elements. Under microscopic examination, the tumor cells were arranged in cord-like pattern and abundant tumor vessels and blood sinuses were present in the stroma. In addition, most tumors contained multifocal extramedullary hematopoiesis. Focal or patchy hemorrhage and necrosis were detected in 6 tumors and calcification in 1. Immunohistochemical staining showed positive expression of Ki-67 in 7 cases, glypican in 5 cases, AFP in 1 case, cytokeratin in 5 cases, vimentin in 3 cases and CD34 in 3 cases.
Most neonatal hepatic neoplasms present as an asymptomatic abdominal mass detected on routine ultrasound, and CHB is also no exception. Thus, the diagnosis might be delayed because of the indolent and nonspecific presentation, resulting in less than 10% of HB diagnosed in the neonatal period.[2,3] There were only eligible 7 CHB cases with pathological confirmation, with a slightly female predominance, which was inconsistent with previous reports mainly due to the relatively small sample size. Moreover, none of the patients in our study had any known risk factors associated with increased incidence of HB, including genetic factors, such as Beckwith Weidemann Syndrome, familial adenomatous polyposis, low birth weight or prematurity, eclampsia or severe pre-eclampsia during pregnancy, and maternal smoking.
Serum AFP has been found to be an useful marker for HB screening.[3,8] However, only 1 patient had an elevated AFP in our series, which did not seem to show a significant association of AFP level with CHB. It is noteworthy that AFP level is normally high at birth and may vary widely among individuals during the neonatal period,[6,8,9] and then decrease rapidly between 6 and 12 months of age, when it attains adult status. Hence, it is a challenge to distinguish tumoral secretion from physiological elevated AFP concentration solely based on 1-time measurement in neonates, which tend to result in a false-negative result.[7,11,12] For this reason, active surveillance of AFP is of great clinical significance for suspected neonates, and CHB should be considered when AFP level increases over time after birth in term neonates and needs to be confirmed by further needle biopsy or resectionas necessary.
From the perspective of pathological origin, HB can be classified into 2 main types:
- (1) epithelial;
- (2) mixed epithelial and mesenchymal.
The former is further divided into multiple subtypes, including fetal, embryonal, small cell undifferentiated, mixed epithelial, and cholangioblastic, according to the degree of cellular differentiation. CHB mainly consists of fetal and embryonal components compared with infants beyond the neonatal period. Patients with pure well-differentiated fetal HB usually have a good outcome with complete excision and no chemotherapy.[14–16] 6/7 tumors in our study showed fetal subtype with only 1 containing mesenchymal elements and all patients had a favorable outcome after excision without additional chemotherapy. The histological immunomarkers that were detected in our series included Ki-67, glypican, alpha-fetoprotein (AFP), cytokeratin, vimentin and B-lymphocyte marker CD34. In recent years, more and more attention has been paid to the significance of glypican-3 in the diagnosis, occurrence and development of HB. Glypican-3 is highly expressed in fetal epithelium, weakly expressed or absent in embryonic epithelial cells, but not expressed in small cell tumor cells. In routine immunohistochemical detection of HB, the intensity and sensitivity of glypican-3 is higher than that of AFP and it has been reported to be an excellent marker for the diagnosis of infantile HB.
Imaging can play a crucial role in diagnosis of hepatic neoplasms in neonates. Most CHB developed in the right lobe of the liver probably resulting from the difference of the circulation to the right and left lobes in fetus. 1/7 tumor among our cases appeared calcification. Calcification is relatively rare and presents usually small and fine in epithelial tumors.[14,19,20] Consistent with the observations seen in previous literature, we also found the occurrence of hemorrhage and cystic necrosis in CHB, and higher incidences in larger tumors. On dynamic contrast-enhanced CT, all the 7 tumors showed early peripheral and intratumoral nodular or striped significant enhancement on the arterial phase, and then the enhancement area showed progressive expansion and fusion filling over time but the attenuation gradually declined, presenting multiple band-like or island-like characteristics with prominently peripheral enhancement on the portal and delayed phases. Our imaging finds validated other 3 previous studys which also discribed hypervascular HB.[5,22,23] This imaging appearrence differs from that of HB in older children generally showing enhance less than surrounding normal liver.[21,24] We infered these imaging features of CHB might be associated with tumor microanatomy that was highly vascularized reflected in abundant tumor vessels and large blood sinuses in the stroma[22,25] and intratumorally multifocal extramedullary hematopoiesis. Decreased aortic caliber below the celiac axis may be present when there is enlargement of the celiac axis or hepatic artery due to vascular shunting into the hepatic mass.[5,23]
The most challenging differential diagnosis of CHB is congenital hemangioma. A single AFP measurement is often unhelpful because the normal range at birth is high. Although congenital hemangioma has some distinctive imaging manifestations—a heterogeneous mass with no pseudocapsule, more common calcification, rarely extrahepatic extension, and gradual centripetal fill-in enhanced pattern till delayed phase, the best strategy to distinguish these 2 hepatic neoplasms is a short period of observation. The congenital hemangioma size may decrease and serum AFP level often falls rapidly after birth, while CHB dose not. In our study, several patients were misdiagnosed as congenital hemangioma at admission, which might result from our lacking recognition of CT characteristics of CHB and neglecting the importance of serial measurements of AFP in the neonatal period.
The main limitations of our study were the retrospective nature and a small sample size, though to our knowledge this is the largest case series reported regarding the CT characteristics of CHB. In addition, it didn’t allow us to adequately understand the imaging difference between different histologic types given that our study subjects were mainly fetal type.
Our study suggested that the clinical and CT characteristics of CHB could differ from those occurred beyond the neonatal age. AFP level should be interpreted with more caution in neonates, and most CHB had fetal components in histopathology. Dynamic contrast-enhanced CT can facilitate recognition and noninvasive diagnosis of CHB, presenting a pattern of progressive expansion and fusion filling but inhomogeneously significant enhancement. However, the estimates of our study are based on too few patients to reach a firmer conclusion. Larger multi-center studies confirming these radiological findings are needed.
Conceptualization: Li Li, Wen Liu, Ke Jin.
Data creation: Li Li, Wen Liu, Rong Wen, Ke Jin.
Formal analysis: Li Li, Wen Liu.
Methodology: Li Li.
Supervision: Li Li, Ke Jin, Wen Liu, Rong Wen.
Validation: Li Li, Ke Jin.
Writing – original draft: Li Li.
Writing – review & editing: Li Li, Wen Liu, Rong Wen, Ke Jin.
. Ammann RA, Plaschkes J, Leibundgut K. Congenital hepatoblastoma
: a distinct entity? Med Pediatr Oncol 1999;32:466–8.
. von Schweinitz D. Neonatal liver tumours. Semin Neonatol 2003;8:403–10.
. Sallam A, Paes B, Bourgeois J. Neonatal hepatoblastoma: two cases posing a diagnostic dilemma, with a review of the literature. Am J Perinatol 2005;22:413–9.
. Trobaugh-Lotrario AD, Chaiyachati BH, Meyers RL, et al. Outcomes for patients with congenital hepatoblastoma
. Pediatr Blood Cancer 2013;60:1817–25.
. Lu M, Greer ML. Hypervascular multifocal hepatoblastoma: dynamic gadolinium-enhanced MRI findings indistinguishable from infantile hemangioendothelioma. Pediatr Radiol 2007;37:587–91.
. Bader D, Riskin A, Vafsi O, et al. Alpha-fetoprotein
in the early neonatal period--a large study and review of the literature. Clin Chim Acta 2004;349:15–23.
. Orbach D, Sarnacki S, Brisse HJ, et al. Neonatal cancer. Lancet Oncol 2013;14:e609–20.
. Ferraro S, Panzeri A, Braga F, et al. Serum alpha-fetoprotein
in pediatric oncology: not a children's tale. Clin Chem Lab Med 2019;57:783–97.
. Blohm ME, Vesterling-Horner D, Calaminus G, et al. Alpha 1-fetoprotein (AFP) reference values in infants up to 2 years of age. Pediatr Hematol Oncol 1998;15:135–42.
. Meyers RL, Maibach R, Hiyama E, et al. Risk-stratified staging in paediatric hepatoblastoma: a unified analysis from the Children's Hepatic tumors International Collaboration. Lancet Oncol 2017;18:122–31.
. Dall’Igna P, Brugieres L, Christin AS, et al. Hepatoblastoma in children aged less than six months at diagnosis: a report from the SIOPEL group. Pediatr Blood Cancer 2018;65(1.):
. Falix FA, Aronson DC, Lamers WH, et al. DLK1, a serum marker for hepatoblastoma in young infants. Pediatr Blood Cancer 2012;59:743–5.
. Roebuck D, Sebire N, Lehmann E, et al. Rapidly involuting congenital haemangioma (RICH) of the liver. Pediatr Radiol 2012;42:308–14.
. Lopez-Terrada D, Alaggio R, de Davila MT, et al. Towards an international pediatric liver tumor consensus classification: proceedings of the Los Angeles COG liver tumors symposium. Mod Pathol 2014;27:472–91.
. Marcio HM, Howard MK, Rebecka LM, 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.
. 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.
. Xiao-Li X, Huan Q, Su-Qi Y, et al. Expression of glypican-3 is highly associated with pediatric hepatoblastoma: a systemic analysis. Asian Pac J Cancer Prev 2015;16:1029–31.
. Feng-hua W, Jian-ming W, Heong-ting V, et al. Glypican 3 expression in hepatoblastoma and its diagnostic implication. Zhonghua bing li xue za zhi 2013;42:806–9.
. Feng-sen B, Xin-yu Y, Yu-chun Y, et al. CT findings of pediatric hepatoblastoma and correlation with pathological subtypes. Chin J Med Imaging Technol 2017;33:1297–300.
. Makin E, Davenport M. Fetal and neonatal liver tumours. Early Hum Dev 2010;86:637–42.
. Xin L, Jian-bo S, et al. Chinese medical imaging. Vol Pediatric Imaging. Beijing: People's medical Publishing House; 2019.
. Shih JC, Tsao PN, Huang SF, et al. Antenatal diagnosis of congenital hepatoblastoma
in utero. Ultrasound Obstet Gynecol 2000;16:94–7.
. Ingram JD, Yerushalmi B, Connell J, et al. Hepatoblastoma in a neonate: a hypervascular presentation mimicking hemangioendothelioma. Pediatr Radiol 2000;30:794–7.
. McCarville MB, Derek JR. Diagnosis and staging of hepatoblastoma: imaging aspects. Pediatric blood & cancer 2012;59:793–9.
. Lin Z, Xiu-hui Z, Shi-min L, et al. Clinicopathological characteristics and prognosis of hepatoblastoma in children. Journal of Practical Oncology 2000. 102–4.
. Soini T, Haveri H, Elo JM, et al. Transcription factor GATA-4 is abundantly expressed in childhood but not in adult liver tumors. J Pediatr Gastroenterol Nutr 2012;54:101–8.