Clinical heterogeneity of pediatric hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is often a chemoresistant neoplasm with a poor prognosis. Pediatric HCC may reflect unique biological and clinical heterogeneity.

fibrolamellar carcinoma that is associated with a specific molecular signature. 7 Notable biological differences between pediatric and adult HCC include risk factors and presence of background liver disease. [5][6][7][8] Unfortunately, despite slightly better chemosensitivity, the 5-year overall survival in children remains low at 13% to 28%. [8][9][10] No large series that define pediatric and adolescent HCC subtype and clinical heterogeneity exist in the literature. A recent publication did comment on the molecular heterogeneity of pediatric HCC by genetically profiling 15 pediatric tumors. 11 Presence of background liver disease, surgical resectability, and response to medical therapy varies between patients. The purpose of this study was to retrospectively characterize HCC in children and adolescents treated at a single institution, to better characterize pediatric HCC heterogeneity.

METHODS
Following institutional review board approval, a retrospective analysis was performed of patients with HCC < 21 years of age treated at our institution between January 1, 2004 and December 31, 2015.
Demographics, preexisting conditions, treatment, and outcomes were reviewed. Abernathy malformation was defined as a congenital portosystemic shunt with the absence of a portal vein. The histopathology, radiographic, and surgical data were reviewed by a pathologist (AG), radiologist (AJT), and surgeons (AB, JDN, MA, and GT).
Histopathology was classified according to WHO Classification of Tumors of the Digestive System. 12 Computed tomography scan and/or magnetic resonance imaging obtained at diagnosis were reviewed to determine the size and the number of liver lesions and the PRETEXT score, based on Children's Oncology Group (COG) modifications to the 2005 PRETEXT guidelines. 13 Evans surgical staging 9 and TNM staging (as per the American Joint Committee on Cancer) were obtained. 14 The Child-Pugh, 15 pediatric end-stage liver disease (PELD) (for children <12 years), 16 and model for end-stage liver disease (MELD) scores (for children ≥ 12 years) 17 were collected. Imaging response were classified as complete response (CR) (100%), partial response (PR) (>30% but less than CR), progressive disease (PD) (≥20% increase in tumor size), and stable disease (SD) (not qualifying for PR or PD).
Second-line therapy and salvage therapy were defined as subsequent therapy for SD or PD or relapsed disease, respectively.
Kaplan-Meier curves were created using GraphPad Prism. A logrank (Mantel-Cox) test was used to calculate P-values.

Histopathology
The primary site of disease was the right lobe (17), left lobe (4)

Surgical treatment
Sixteen patients underwent primary surgical resection and five after neoadjuvant chemotherapy. Surgery included liver transplantation (LT, 11), hemihepatectomy (9), and segmentectomy (1). Nine had a conventional resection and were either PRETEXT I (4) or II (5). Eight had negative parenchymal margins, however three had either metastatic lung disease, regional lymph node disease, or intrabiliary tract and lymphovascular invasion. Other local control treatments included Yttrium-90 radioembolization (4), radiofrequency ablation ( Of the 11 patients who underwent LT ( outside of MC was due to concern for impending liver failure or complications from preexisting liver disease as detailed above.

Medical treatment
Five patients received adjuvant chemotherapy and 13 received neoadjuvant chemotherapy or chemotherapy alone (Table 3).
Three patients discontinued sorafenib due to rash (3) and pancreatitis There was also no difference in median survival based on age <15 years (35.9 months) compared to ≥15 years (20.1 months) ( Figure 2C; Overall survival based on Evans stage and surgical modality. Stage was the most prognostic factor, however in patients with higher stage disease (Evans stages 3 and 4), liver transplantation lead to better survival compared to conventional resection (P < 0.0012) and LT (1). Two of 7 patients (28%) with fibrolamellar HCC and 12 of 25 patients (48%) with conventional HCC remain alive ( Figure 2E; Only 4 of 18 patients (22%) who received medical therapy survived.

DISCUSSION
HCC is an aggressive neoplasm with age-dependent differences in epidemiology, baseline liver function, histology, and response to therapy, suggesting biological differences. A recent publication on the genomic heterogeneity of pediatric HCC did note a molecularly distinct pattern of 15 sequenced tumors. 11 Moreover, there is clinical heterogeneity of HCC comparing younger children to adolescents and young adults, thus far not well described.
Both genetic and anatomic predispositions to HCC are seen in pediatric patients, especially younger children [19][20][21][22] While previous data suggested that 30% of pediatric HCC in the Western world is associated with a predisposition, 19 53% of our cohort had a genetic or anatomic predisposition, reflective of data from a large, tertiary care center. There was no difference in survival between patients with a predisposition versus those without.
The cornerstone for HCC-directed therapy is a complete surgical resection, however two-thirds of pediatric patients with HCC present with unresectable disease. In our cohort, patients with complete surgical resectability and no evidence of regional or distant disease had a favorable outcome, as evidenced by survival based on Evans stage, consistent with prior reports. 9,18 Given frequent chemoresistance and challenges with conventional resection, the role of LT in pediatric HCC continues to evolve. Some studies have shown improved disease-free survival (range 63-89%) with LT, especially in children with background cirrhosis due to an underlying predisposition. 23 30 In our cohort, there was a trend toward improved survival with LT, but not significant ( Figure 1E). It is important to note that both patients who relapsed after LT did so outside of the liver (lung and abdominal cavity) and both who relapsed after conventional resection had local relapses in the liver in the setting of narrow negative margins (0.2-0.3 mm). While the issue of adequate margins remains debatable, [31][32][33][34] these data support LT being a practical method to control local disease in the absence of metastatic disease and in settings where a conventional surgery is challenging.
There is currently no distinct pediatric criteria to determine the There are several limitations of this retrospective review. As a single institutional study from a large tertiary care center, our epidemiological data may not be generalizable to other centers. Additionally, surgical guidelines vary between institutions and may lead to differing outcomes. Our cohort, while the largest single institutional study in the literature to date, remains small and as such, limits extensive analytics.
In summary, pediatric HCC is often a chemoresistant neoplasm in which complete surgical resection by either conventional methods or LT remains the best chance for cure. There is significant heterogeneity in pediatric HCC in regard to risk factors, background liver disease, and resectability, suggesting the need for tailored therapy, rather than a uniform treatment approach to all pediatric and adolescent HCC. The Pediatric Hepatic International Tumor Trial (PHITT), (COG AHEP1531; JCCG -JPLT4; SIOPEL -PHITT), is the first prospective, multicooperative group international trial studying outcomes in patients with HCC using standardized therapy arms based on having resectable or unresectable and/or metastatic disease, as well as the presence or absence of preexisting conditions. Importantly, a major aim of the study will be to analyze the biology of HCC in order to develop biomarkers of clinical heterogeneity in effort to optimize individualized treatment, as well as to advance novel therapeutics.