Abstract
BACKGROUND:
Hepatocellular carcinoma (HCC) is a rare cancer in children, with various histologic subtypes and a paucity of data to guide clinical management and predict prognosis.
METHODS:
A multi-institutional review of children with hepatocellular neoplasms was performed, including demographic, staging, treatment, and outcomes data. Patients were categorized as having conventional HCC (cHCC) with or without underlying liver disease, fibrolamellar carcinoma (FLC), and hepatoblastoma with HCC features (HB-HCC). Univariate and multivariate analyses identified predictors of mortality and relapse.
RESULTS:
In total, 262 children were identified; and an institutional histologic review revealed 110 cHCCs (42%; 69 normal background liver, 34 inflammatory/cirrhotic, 7 unknown), 119 FLCs (45%), and 33 HB-HCCs (12%). The authors observed notable differences in presentation and behavior among tumor subtypes, including increased lymph node involvement in FLC and higher stage in cHCC. Factors associated with mortality included cHCC (hazard ratio [HR], 1.63; P = .038), elevated α-fetoprotein (HR, 3.1; P = .014), multifocality (HR, 2.4; P < .001), and PRETEXT (pretreatment extent of disease) stage IV (HR, 5.76; P < .001). Multivariate analysis identified increased mortality in cHCC versus FLC (HR, 2.2; P = .004) and in unresectable tumors (HR, 3.4; P < .001). Disease-free status at any point predicted survival.
CONCLUSIONS:
This multi-institutional, detailed data set allowed a comprehensive analysis of outcomes for children with these rare hepatocellular neoplasms. The current data demonstrated that pediatric HCC subtypes are not equivalent entities because FLC and cHCC have distinct anatomic patterns and outcomes in concert with their known molecular differences. This data set will be further used to elucidate the impact of histology on specific treatment responses, with the goal of designing risk-stratified algorithms for children with HCC.
Keywords: fibrolamellar, hepatobiliary, hepatocellular carcinoma (HCC), hepatocellular neoplasm, not otherwise specified (HCN-NOS), pediatric oncology
LAY SUMMARY:
This is the largest reported granular data set on children with hepatocellular carcinoma.
The study evaluates different subtypes of hepatocellular carcinoma and identifies key differences between subtypes.
This information is pivotal in improving understanding of these rare cancers and may be used to improve clinical management and subsequent outcome in children with these rare malignancies.
INTRODUCTION
Recent analysis of collaborative data regarding the most common liver tumor in children, hepatoblastoma (HB), has led to an improved understanding of the factors driving HB prognosis.1,2 Other pediatric liver cancers are rare, and little is known about their biologic behavior.3 Although it is the third most common cancer in adults, hepatocellular carcinoma (HCC) is so rare in children that our current clinical understanding comes from small case series or reviews of national cancer databases, the latter of which lack sufficient granularity to guide management.4–7 This gap in the published literature has resulted in uncertainty regarding the true incidence of rare pediatric liver tumors and how best to treat them.
One of the historic challenges in this field has been the lack of a consensus classification system. In an attempt initially designed to standardize histologic subtypes of HB, the Children’s Hepatic Tumor International Collaboration (CHIC)2 convened an international panel of pediatric liver pathologists to develop a common histologic classification for all pediatric liver tumors. On the basis of this work, published by Lopez-Terrada et al in 2014,8 we currently believe that there are 5 distinct subtypes of pediatric HCC: conventional HCCs (cHCCs) in patients with metabolic liver disease, cHCCs arising in chronically injured liver, de novo cHCCs in normal livers, fibrolamellar carcinomas (FLCs), and tumors that have mixed features of HCC and HB (HB-HCCs) (also referred to as hepatocellular neoplasm, not otherwise specified [HCN-NOS]).
Although small numbers of children with HCC have been included in prior collaborative trials9,10 and in a pilot study of sorafenib in Germany,11 they were not included in the most recently completed Children’s Oncology Group (COG) liver tumor study (AHEP0731). To bridge these knowledge gaps, we leveraged the Pediatric Surgical Oncology Research Collaborative infrastructure12,13 to better understand current outcomes for children with non-HB hepatocellular neoplasms and to determine how histologic subtype predicts outcome for these patients. Contrary to prior pediatric studies, we found clear differences between cHCCs and FLCs and suggest that these groups warrant distinct categorization and management in future trials.
MATERIALS AND METHODS
Patients from 19 hospitals participating in the Pediatric Surgical Oncology Research Collaborative, a multi-institutional consortium of North American pediatric surgeons dedicated to advancing the surgical care of children with cancer, were included in the study. Central Institutional Review Board (IRB) approval was obtained at Cincinnati Children’s Hospital Medical Center (IRB no. 2019–0703C), and each institution obtained IRB approval, either individually or through reliance on the central IRB. Study data were collected and managed using REDCap electronic data-capture tools hosted at Cincinnati Children’s Hospital Medical Center, and data use agreements between each institution and the lead institution (Utah) were established.
Patients younger than 20 years who had hepatocellular neoplasms between 1990 and 2017 were identified; demographic data, radiographic14 and surgical stage, and other comorbidities were noted. Treatment modalities, extent and completeness of resection, and histologic subtype were gleaned from direct review of patient records. Although termed HCN-NOS in the consensus classification of histologic subtypes from Lopez-Terrada et al,8 because we do not yet have central expert pathologist review in this study, tumors with features of both HB and HCC referenced in the institutional pathology report are herein called HB-HCC. Outcome measures included the presence or absence of surgical complications, development of a secondary malignancy, achievement of disease-free status, progression, relapse, and death.
Demographic, imaging, and treatment characteristics were then compared between different tumor (cHCC, FLC, HB-HCC) and background liver (inflammatory and noninflammatory) histologic groups. Baseline numeric characteristics were compared using an analysis of variance or the Kruskal-Wallis test, depending on distribution skew. Comparisons of numeric variables with parenchymal characteristics were conducted with a Wilcoxon rank-sum test. Categorical variables were compared using χ2 tests or the Fisher exact test if any expected cell counts were <5.
Univariable Cox proportional hazard regression analyses were performed to examine the effect of histology on relapse or death. Separate multivariable Cox regression models were constructed for relapse or death using backward, stepwise model selection to select 4 or 5 predictors, so that there were a minimum of 5 events per predictor in each of these models.15 Hazard ratios (HRs) with 95% confidence intervals (CIs) and P values are reported; Kaplan-Meier analyses were used to demonstrate the survival and relapse functions for each histologic group. HB-HCCs were excluded from the Cox proportional hazards regression model due to violation of the proportional hazards assumption.16 Comparisons across all subgroups were conducted using a weighted log-rank test.17 Statistical significance was assessed at the .05 level using 2-tailed tests. All analyses were conducted in Stata MP/15.1 (StataCorp, 2017).
RESULTS
Patient Demographics and Tumor Staging
We identified 262 patients (see Supporting Table 1) who had hepatocellular neoplasms, including 110 (42%) cHCCs (69 de novo tumors with normal background liver, 34 inflammatory/cirrhotic, 7 unknown), 119 (45%) FLCs, and 33 (12%) HB-HCCs. There were differences in the mean age of presentation, depending on histology, ranging from 105.9 months (HB-HCC) to 182 months (FLC; P = .002) (Table 1). There were also notable differences in patient race and ethnicity, with an increased proportion of White patients who had FLC and more Hispanic patients who had cHCC and HB-HCC (P = .001). Comorbid conditions were less common in children with FLC, and differences in α-fetoprotein (AFP) levels were seen among tumor types, with AFP elevations >100 ng/mL reported in 90.9% of children with HB-HCC versus 4.2% of children with FLC (P < .001). Patients who had cHCC with underlying inflammatory liver disease were more likely to be younger (119.7 ± 75.1 vs 150.3 ± 58.0 months; P = .02) and to have associated comorbid conditions (51.1% vs 20.9%; P < .001), and they were less likely to present with metastatic disease (14.9% vs 34.9%; P = .012) than patients who had cHCC with de novo tumors and normal background liver. We compared pretreatment extent of disease (PRETEXT) stage and annotation factors between groups and found notable differences in radiographic findings (Table 1). These data highlight the distinct anatomic patterns of FLC versus cHCC, with the former more likely to spread to lymph nodes and the latter more likely to spread more extensively within the liver.
TABLE 1.
Patient Demographics and Tumor Staging
| No. of Patients (%) | No. of Patients (%) | |||||||
|---|---|---|---|---|---|---|---|---|
| Variable | Total, N = 262 | FLC, N = 119 | HB-HCC, N = 33 | Conventional HCC, N = 110 | P | Inflammatory Liver Disease, N = 47 | Normal Liver, N = 215 | P |
| Age: Mean ± SD, mo | 144.8 ± 62.4 | 182.0 ± 36.8 | 105.9 ± 51.9 | 115.9 ± 65.4 | <.001 | 119.7 ± 75.1 | 150.3 ± 58.0 | .002 |
| Males | 148 (56.5) | 58 (48.7) | 22 (66.7) | 68 (61.8) | .062 | 28 (59.6) | 120 (55.8) | .64 |
| Race/ethnicity | .001 | .75 | ||||||
| White | 171 (65.0) | 94 (79.0) | 21 (64.0) | 56 (51.0) | 28 (60.0) | 143 (67.0) | ||
| Black | 20 (8.0) | 7 (6.0) | 2 (6.0) | 11 (10.0) | 5 (11.0) | 15 (7.0) | ||
| Hispanic | 33 (13.0) | 6 (5.0) | 6 (18.0) | 21 (19.0) | 7 (15.0) | 26 (12.0) | ||
| Other | 38 (15.0) | 12 (10.0) | 4 (12.0) | 22 (20.0) | 7 (15.0) | 31 (14.0) | ||
| Other comorbidity | .002 | <.001 | ||||||
| Yes | 69 (26.3) | 20 (16.8) | 11 (33.3) | 38 (34.5) | 24 (51.1) | 45 (20.9) | ||
| Unknown | 16 (6.1) | 4 (3.4) | 1 (3.0) | 11 (10.0) | 2 (4.3) | 14 (6.5) | ||
| AFP at diagnosis, ng/mL | <.001 | .48 | ||||||
| <100 | 103 (39.3) | 76 (63.9) | 3 (9.1) | 24 (21.8) | 15 (31.9) | 88 (40.9) | ||
| 100–1,000,000 | 89 (34.0) | 4 (3.4) | 28 (84.8) | 57 (51.8) | 20 (42.6) | 69 (32.1) | ||
| >1,000,000 | 13 (5.0) | 1 (0.8) | 2 (6.1) | 10 (9.1) | 3 (6.4) | 10 (4.7) | ||
| Unknown | 57 (21.8) | 38 (31.9) | 0 (0.0) | 19 (17.3) | 9 (19.1) | 48 (22.3) | ||
| PRETEXT group | .002 | .18 | ||||||
| I | 21 (8.0) | 15 (12.6) | 1 (3.0) | 5 (4.5) | 4 (8.5) | 17 (7.9) | ||
| II | 58 (22.1) | 35 (29.4) | 5 (15.2) | 18 (16.4) | 11 (23.4) | 47 (21.9) | ||
| III | 53 (20.2) | 21 (17.6) | 8 (24.2) | 24 (21.8) | 4 (8.5) | 49 (22.8) | ||
| IV | 48 (18.3) | 12 (10.1) | 12 (36.4) | 24 (21.8) | 8 (17.0) | 40 (18.6) | ||
| Unknown | 82 (31.3) | 36 (30.3) | 7 (21.2) | 39 (35.5) | 20 (42.6) | 62 (28.8) | ||
| Major vascular involvement | .016 | .22 | ||||||
| Yes | 98 (37.4) | 39 (32.8) | 15 (45.5) | 44 (40.0) | 16 (34.0) | 82 (38.1) | ||
| Unknown | 19 (7.3) | 4 (3.4) | 1 (3.0) | 14 (12.7) | 1 (2.1) | 18 (8.4) | ||
| Lymph node metastasis | <.001 | .24 | ||||||
| Yes | 77 (29.7) | 59 (50.0) | 1 (3.1) | 17 (15.6) | 11 (23.4) | 66 (31.1) | ||
| Unknown | 29 (11.2) | 5 (4.2) | 3 (9.4) | 21 (19.3) | 3 (6.4) | 26 (12.3) | ||
| Distant metastasis | .25 | .012 | ||||||
| Yes | 82 (31.3) | 42 (35.3) | 10 (30.3) | 30 (27.3) | 7 (14.9) | 75 (34.9) | ||
| Unknown | 11 (4.2) | 3 (2.5) | 0 (0.0) | 8 (7.3) | 1 (2.1) | 10 (4.7) | ||
Abbreviations: AFP, α-fetoprotein; FLC, fibrolamellar carcinoma; HB-HCC hepatoblastoma with hepatocellular carcinoma features; HCC, hepatocellular carcinoma; PRETEXT, pretreatment extent of disease; SD, standard deviation; mo, months.
Treatment and Outcomes
Open or laparoscopic biopsies were more commonly performed than percutaneous approaches during our study period (Table 2). More children with HB-HCC received neoadjuvant chemotherapy (63.6% vs 45.5% in cHCC and 52.1% in FLC; P < .001) or adjuvant chemotherapy (60.6% vs 21.8% in cHCC and 37.8% in FLC; P < .001). cHCCs arising in the setting of inflammatory liver disease were less likely to be treated with neoadjuvant chemotherapy (31.9% vs 54.9%) or adjuvant chemotherapy (6.4% vs 40%) compared with those arising in normal livers.
TABLE 2.
Treatment and Outcomes
| No. of Patients (%) | No. of Patients (%) | |||||||
|---|---|---|---|---|---|---|---|---|
| Variable | Total, N = 262 | FLC, N = 119 | HB-HCC, N = 33 | Conventional HCC, N = 110 | P | Inflammatory Liver Disease, N = 47 | Normal Liver, N = 215 | P |
| Dx, biopsy | .13 | .85 | ||||||
| Open or laparoscopic | 123 (46.9) | 54 (45.4) | 20 (60.6) | 49 (44.5) | 21 (44.7) | 102 (47.4) | ||
| Percutaneous | 91 (34.7) | 48 (40.3) | 6 (18.2) | 37 (33.6) | 18 (38.3) | 73 (34.0) | ||
| Unknown | 48 (18.3) | 17 (14.3) | 7 (21.2) | 24 (21.8) | 8 (17.0) | 40 (18.6) | ||
| Neoadjuvant chemotherapy | 133 (50.8) | 62 (52.1) | 21 (63.6) | 50 (45.5) | <.001 | 15 (31.9) | 118 (54.9) | .017 |
| Adjuvant chemotherapy | 89 (34.0) | 45 (37.8) | 20 (60.6) | 24 (21.8) | <.001 | 3 (6.4) | 86 (40.0) | <.001 |
| Transarteriai Y90 radioembolization | 12 (4.6) | 5 (4.2) | 2 (6.1) | 5 (4.5) | .005 | 3 (6.4) | 9 (4.2) | .76 |
| Transarterial chemoembolization | 29 (11.1) | 10 (8.4) | 4 (12.1) | 15 (13.6) | .002 | 7 (14.9) | 22 (10.2) | .60 |
| Surgery at Dx | <.001 | <.001 | ||||||
| Partial hepatectomy | 90 (80.4) | 64 (95.5) | 7 (87.5) | 19 (51.4) | 7 (29.2) | 83 (94.3) | ||
| Liver transplantation | 22 (19.6) | 3 (4.5) | 1 (12.5) | 18 (48.6) | 17 (70.8) | 5 (5.7) | ||
| Surgery after chemotherapy | .10 | .006 | ||||||
| Partial hepatectomy | 44 (67.7) | 18 (81.8) | 9 (50.0) | 17 (68.0) | 2 (25.0) | 42 (73.7) | ||
| Liver transplantation | 21 (32.3) | 4 (18.2) | 9 (50.0) | 8 (32.0) | 6 (75.0) | 15 (26.3) | ||
| No surgical resection | 81 (30.9) | 29 (24.4) | 6 (18.2) | 46 (41.8) | .004 | 15 (31.9) | 66 (30.7) | .87 |
| Extent of surgery | .013 | .56 | ||||||
| R0 | 155 (59.2) | 77 (64.7) | 23 (69.7) | 55 (50.0) | 27 (57.4) | 128 (59.5) | ||
| R1 | 16 (6.1) | 7 (5.9) | 4 (12.1) | 5 (4.5) | 1 (2.1) | 15 (7.0) | ||
| R2 | 5 (1.9) | 4 (3.4) | 0 (0.0) | 1 (0.9) | 1 (2.1) | 4 (1.9) | ||
| Unknown or none | 86 (32.8) | 31 (26.1) | 6 (18.2) | 49 (44.5) | 18 (38.3) | 68 (31.6) | ||
| Surgical complications within 30 d | .18 | .14 | ||||||
| Any | 66 (25.2) | 35 (29.4) | 10 (30.3) | 21 (19.1) | 17 (36.2) | 49 (22.8) | ||
| Surgical death | 2 (0.8) | 0 (0.0) | 0 (0.0) | 2 (1.8) | 0 (0.0) | 2 (0.9) | ||
| Ever considered disease-free | 146 (55.7) | 71 (59.7) | 22 (66.7) | 53 (48.2) | .11 | 26 (55.3) | 120 (55.8) | .88 |
| Relapse | 56 (21.5) | 36 (30.3) | 11 (34.4) | 9 (8.2) | .001 | 2 (4.3) | 54 (25.2) | .003 |
| Time of first relapse, N = 51: Median [IQR], mo | 11.3 [5.2–22.8] | 11.3 [4.4–22.8] | 12.2 [5.2–106.6] | 15.0 [9.8–20.4] | — | 10.6 [7.6–13.5] | 11.3 [5.2–22.8] | — |
| Relapse in liver | 41 (53.2) | 24 (44.4) | 9 (75.0) | 8 (72.7) | .060 | 2 (40.0) | 39 (54.2) | .54 |
| Relapse in lung | 32 (41.6) | 22 (40.7) | 6 (50.0) | 4 (36.4) | .78 | 2 (40.0) | 30 (41.7) | .94 |
| Relapse in other location | 40 (51.9) | 35 (64.8) | 3 (23.1) | 2 (20.0) | .002 | 2 (40.0) | 38 (52.8) | .58 |
| Progression | 58 (22.3) | 34 (28.8) | 4 (12.5) | 20 (18.2) | .029 | 10 (21.3) | 48 (22.5) | .56 |
| Time of first progression after Dx: Median [IQR], mo | 6.0 [3.9–11.4] | 9.8 [3.9–12.4] | 9.2 [5.2–11.1] | 5.4 [1.6–6.0] | — | 9.8 [4.0–19.4] | 5.8 [3.9–11.1] | — |
| Death | 128 (49.8) | 61 (51.7) | 14 (42.4) | 53 (50.0)a | .021 | 22 (48.9) | 106 (50.0) | .55 |
| Relapseb | 32 (25.0) | 19 (31.1) | 7 (50.0) | 4 (7.5) | 1 (4.5) | 30 (28.3) | ||
| Progressionb | 47 (36.7) | 29 (47.5) | 1 (7.1) | 17 (32.0) | 8 (36.3) | 39 (36.8) | ||
| Otherb | 1 (0.7) | 0 (0.0) | 0 (0.0) | 1 (1.9) | 0 (0.0) | 1 (0.9) | ||
| Unknownb | 48 (37.5) | 13 (21.3) | 6 (42.8) | 31 (58.4) | 13 (59.0) | 36 (33.9) | ||
| Time from Dx to death: Median [IQR], mo | 29.3 [11.4–73.4] | 31.2 [18.1–71.6] | 50.1 [12.1–102.1] | 21.6 [6.7–66.8] | — | 29.9 [10.6–62.0] | 29.1 [11.5–75.7] | — |
Abbreviations: AFP, α-fetoprotein; Dx, diagnosis; FLC, fibrolamellar carcinoma; HB-HCC hepatoblastoma with hepatocellular carcinoma features; HCC, hepatocellular carcinoma; PRETEXT, pretreatment extent of disease; R0, microscopically margin-negative resection; R1, resection with microscopic residual tumor; R2, resection with macroscopic residual tumor; Y90, Yttrium 90.
Four patients had unknown status at last follow-up (total N = 106).
These were associated with mortality and/or death.
In total, 177 (67.6%) patients had their tumors resected at some time in their treatment. One hundred twelve (42.4%) children underwent resection at diagnosis. Of these, 90 underwent a partial hepatectomy, and 22 underwent orthotopic liver transplantation (OLT); OLT was most commonly performed in patients with cHCC (P = .002). Sixty-five patients underwent resection after neoadjuvant therapy, including 44 partial hepatectomies and 21 OLTs. Eighty-one (30.9%) children never underwent resection, and more of these children had cHCC than HB-HCC or FLC (P = .004). We were unable to determine the surgical management in 4 patients. An R0 resection was more commonly achieved in FLC (64.7%) and HB-HCC (69.7%) than in cHCC (50%; P = .013).
Surgical complications did not differ by histologic subtype, but 25.5% of all patients experienced a perioperative complication, and death within 30 days was reported in 2 patients (0.8%) (Table 2). One hundred forty-six (55.7%) patients were able to achieve disease-free status; these rates were similar between tumor types. Patterns of relapse differed because patients who had cHCC and HB-HCC trended toward relapsing more frequently in the liver (P = .06), and those who had FLC relapsed more frequently outside the liver and lung (P = .02), with 16 of 35 patients relapsing in lymph nodes.
One hundred twenty-eight (49.8%) children in our data set died; long-term mortality rates were similar between those who had FLC (51.7%) versus cHCC (50%) but were lower in those who had HB-HCC (42.4%; P = .021). Fifty-three deaths occurred in children with cHCC, 42 of whom never achieved disease-free status. Eleven deaths occurred in children who achieved disease-free status but later relapsed (n = 9), developed a secondary neoplasm (n = 1), or died from an unknown reason (n = 1). Overall, deaths were observed in 4 of 9 patients who had cHCC with a documented relapse and in 17 of 20 who had progression of disease. One child with progressive disease had unknown status (alive vs dead) at last follow-up. There were 63 deaths in children with FLC. Mortality occurred in 19 of the 36 patients who relapsed and in 29 of 34 who had evidence of progressive disease at last follow-up. In some patients, we were unable to determine the cause of death (eg, relapse, progress, etc) because of the limitations of our retrospective review. Nonetheless, there were significant differences in the median time from diagnosis to death between histologic subtypes (21.6 months in cHCC vs 31.2 months in FLC vs 50.1 months in HB-HCC; P = .005). Disease progression was more common in FLC (28.8%) versus cHCC (18.2%) and HB-HCC (12.5%; P = .029). Relapse was least common in cHCC (8.2%) compared with FLC (30.3%) and HB-HCC (34.4.%; P = .001), coincident with the higher percentage of kids with cHCC who were unable undergo resection.
Patients who had cHCC with inflammatory liver disease were more likely to undergo OLT, both at diagnosis (70.7% vs 5.7%; P < .001) and after neoadjuvant chemotherapy (75% vs 26.3%; P = .006). Patients who had cHCC with underlying inflammatory liver disease achieved disease-free status at a rate similar to those without underlying liver disease (55.3% vs 55.8%), but tumors arising in noninflammatory livers had higher rates of relapse (25.5% vs 4.3%; P = .003). No differences in mortality were observed in these 2 groups, however (48.9% vs 50%, respectively), suggesting that earlier detection and increased use of OLT in children with underlying liver disease balanced out their increased attendant operative risk. Across different tumor subtypes, lower mortality rates were observed among those who underwent resection. For example, 42 of 64 children with cHCC who did not achieve disease-free status died (65.6%), whereas only 11 of 53 (20.8%) who underwent resection died. There were 7 patients with cHCC who did not achieve disease-free status despite resection. Similar trends were seen in patients with FLC, such that achieving disease-free status was identified as a statistically significant predictor of outcome in our univariate analyses (Table 3).
TABLE 3.
Univariable Cox Model for Death: Conventional Hepatocellular Carcinoma and Fibrolamellar Carcinoma
| Variable | Level | HR (95% CI) | P |
|---|---|---|---|
| Pathology | FLC | Reference | |
| Conventional HCC | 1.63 (1.03, 2.58) | .038 | |
| Non-inflammatory | Inflammatory liver disease | Reference | |
| Normal liver | 0.74 (0.43, 1.26) | .27 | |
| Age (month) | <98 | Reference | |
| 99 to 156 | 1.48 (0.74, 2.94) | .27 | |
| 157 to 197 | 1.14 (0.57, 2.27) | .71 | |
| ≥198 | 1.37 (0.71, 2.66) | .35 | |
| Other comorbidity | No | Reference | |
| Yes | 1.60 (0.97, 2.66) | .067 | |
| Unknown | 1.97 (0.72, 5.38) | .19 | |
| AFP at Dx, ng/mL | AFP <100 | Reference | |
| AFP 100–1,000,000 | 1.72 (1.05, 2.81) | .032 | |
| AFP >1,000,000 | 3.11 (1.26, 7.71) | .014 | |
| Unknown | 0.90 (0.54, 1.52) | .70 | |
| Extent of surgery | R0 | Reference | |
| R1 | 1.53 (0.66, 3.58) | .32 | |
| R2 | 2.79 (0.81, 9.59) | .10 | |
| Unknown or None | 5.77 (3.60, 9.25) | <.001 | |
| PRETEXT Group | I | Reference | |
| II | 1.64 (0.60, 4.49) | .34 | |
| III | 2.59 (0.92, 7.28) | .072 | |
| IV | 5.76 (2.02, 16.41) | .001 | |
| Unknown | 3.62 (1.33, 9.83) | .012 | |
| Major Vascular Involvement | No | Reference | |
| Yes | 3.17 (1.81, 5.55) | <.001 | |
| Unknown | 1.05 (0.40, 2.76) | .92 | |
| Multifocal | Unifocal | Reference | |
| Multifocal | 2.40 (1.52, 3.80) | <.001 | |
| Unknown | 0.82 (0.27, 2.45) | .72 | |
| Caudate lobe involvement | No | Reference | |
| Yes | 2.02 (1.18, 3.48) | .011 | |
| Unknown | 0.99 (0.50, 1.96) | .98 | |
| Lymph node metastasis | No | Reference | |
| Yes | 1.45 (0.93, 2.26) | .10 | |
| Unknown | 2.30 (1.14, 4.62) | .020 | |
| Distant metastasis | No | Reference | |
| Yes | 2.72 (1.72, 4.31) | <.001 | |
| Unknown | 3.78 (1.44, 9.92) | .007 | |
| Surgical complications | None | Reference | |
| Any <30 d | 0.68 (0.41, 1.12) | .13 | |
| Surgical death <30 d | 10.69 (1.91, 59.87) | .007 | |
| Ever considered disease free | No | Reference | |
| Yes | 0.10 (0.06, 0.18) | <.001 | |
| Unknown | 0.36 (0.14, 0.94) | .036 | |
| Time of surgical resection | At diagnosis | Reference | |
| After chemotherapy | 1.71 (0.93, 3.14) | .086 | |
| Resect (Time unknown) | 0.64 (0.07, 5.41) | .68 | |
| No resection | 10.37 (5.69, 18.90) | <.001 | |
| Treatment | Resection at Diagnosis:Partial Hepatectomy | Reference | |
| Resection at Diagnosis:OLT | 0.67 (0.19, 2.41) | .54 | |
| Resection after chemotherapy: Partial Hepatectomy | 1.78 (0.92, 3.46) | .088 | |
| Resection after chemotherapy: OLT | 1.04 (0.32, 3.36) | .95 | |
| No Surgical Resection | 9.62 (5.18, 17.87) | <.001 |
Abbreviations: AFP, α-fetoprotein; CI, confidence interval; Dx, diagnosis; FLC, fibrolamellar carcinoma; HCC, hepatocellular carcinoma; HR, hazard ratio; OLT, orthotopic liver transplantation; PRETEXT, pretreatment extent of disease; R0, microscopically margin-negative resection; R1, resection with microscopic residual tumor; R2, resection with macroscopic residual tumor.
Univariable Cox Models for Relapse and Death in cHCC and FLC
We designed Cox models to assess the impact of clinical and treatment factors on relapse; because of the low number and for statistical purposes, patients with HB-HCC were excluded from these analyses (Table 4). The only predictor of relapse in cHCC and FLC was the absence of surgical resection (P = .009), although the presence of lymph node metastases approached significance (P = .062).
TABLE 4.
Univariable Cox Model for Relapse of Hepatocellular Carcinoma and Fibrolamellar Carcinoma
| Variable | Level | HR (95% CI) | P |
|---|---|---|---|
| Pathology | FLC | Reference | |
| Conventional HCC | 0.44 (0.17, 1.13) | .087 | |
| Non-inflammatory | Inflammatory liver | Reference | |
| Non-inflammatory liver | 3.32 (0.79, 14.05) | .10 | |
| Age (month) | <98 | Reference | |
| 99 to 156 | 1.24 (0.29, 5.24) | .77 | |
| 157 to 197 | 2.62 (0.69, 9.93) | .16 | |
| ≥198 | 1.72 (0.41, 7.16) | .45 | |
| Other comorbidity | No | Reference | |
| Yes | 1.07 (0.44, 2.58) | .89 | |
| Unknown | 1.81 (0.26, 12.57) | .55 | |
| AFP at Dx, ng/mL | AFP <100 | Reference | |
| AFP 100–1,000,000 | 0.67 (0.27, 1.69) | .40 | |
| AFP >1,000,000 | 0.00 (0.00, .) | 1.00 | |
| Unknown | 0.51 (0.24, 1.11) | .091 | |
| Extent of surgery | R0 | Reference | |
| R1 | 0.49 (0.14, 1.65) | .25 | |
| R2 | 0.78 (0.09, 6.38) | .82 | |
| Unknown or None | 0.15 (0.04, 0.63) | .009 | |
| PRETEXT Group | I | Reference | |
| II | 2.01 (0.55, 7.33) | .29 | |
| III | 2.77 (0.72, 10.67) | .14 | |
| IV | 0.00 (0.00, .) | 1.00 | |
| Unknown | 2.38 (0.64, 8.81) | .20 | |
| Major Vascular Involvement | No | Reference | |
| Yes | 0.22 (0.03, 1.69) | .15 | |
| Unknown | 0.52 (0.05, 5.11) | .58 | |
| Multifocal | Unifocal | Reference | |
| Multifocal | 1.16 (0.54, 2.51) | .70 | |
| Unknown | 0.75 (0.16, 3.54) | .71 | |
| Caudate lobe involvement | No | Reference | |
| Yes | 0.89 (0.31, 2.52) | .82 | |
| Unknown | 0.24 (0.03, 1.98) | .19 | |
| Lymph node metastasis | No | Reference | |
| Yes | 1.83 (0.97, 3.46) | .062 | |
| Unknown | 0.22 (0.03, 1.79) | .15 | |
| Distant metastasis | No | Reference | |
| Yes | 1.21 (0.59, 2.48) | .61 | |
| Unknown | 1.94 (0.20, 19.11) | .57 | |
| Surgical complications | None | Reference | |
| Any <30 d | 1.76 (0.92, 3.39) | .090 | |
| Surgical death <30 d | 0.00 (0.00, .) | 1.00 |
Abbreviations: AFP, α-fetoprotein; CI, confidence interval; Dx, diagnosis; FLC, fibrolamellar carcinoma; HCC, hepatocellular carcinoma; HR, hazard ratio; PRETEXT, pretreatment extent of disease; R0, microscopically margin-negative resection; R1, resection with microscopic residual tumor; R2, resection with macroscopic residual tumor.
Univariate analyses identified several factors predictive of mortality; patients with HB-HCC were again excluded. cHCC was an independent predictor of mortality (HR, 1.63; 95% CI, 1.03–2.58; P = .038) because of the earlier time from diagnosis to death. Other factors associated with mortality included elevated AFP, unknown or gross residual disease at the time of resection, imaging factors such as PRETEXT stage IV or the presence of distant metastasis, and lack of surgical resection (Table 3). Several of these anatomic factors also predicted a poor outcome in HB, although low AFP was historically thought to portend a worse prognosis in HB,1 contrary to our findings in HCC. The primary factor associated with decreased mortality was attaining disease-free status at any time (HR, 0.10; 95% CI, 0.06–0.18; P < .001).
Multivariate Cox Models and Overall Rates of Relapse and Death in cHCC and FLC
Multivariable regression analyses identified multifocal disease as an independent predictor of relapse (HR, 3.47; 95% CI, 1.38–8.74; P = .008) (Table 5), and cHCC (HR, 2.20; 95% CI, 1.29–3.76; P = .004) and the absence of surgical resection (HR, 3.45; 95% CI, 1.70–7.02; P < .001) were identified as strong predictors of mortality. Attaining disease-free status at any point was associated with survival (HR, 0.15; 95% CI, 0.07–0.29; P < .001). These data strongly support an aggressive surgical approach for all children with HCC.
TABLE 5.
Multiple Cox model for Relapse and Death
| Variable | Level | HR (95% CI) | P |
|---|---|---|---|
| Relapse | |||
| Pathology | FLC | Reference | |
| Conventional HCC | 0.39 (0.13, 1.14) | .084 | |
| Multifocal | Unifocal | Reference | |
| Multifocal | 3.47 (1.38, 8.74) | .008 | |
| Unknown | 1.24 (0.18, 8.77) | .83 | |
| Distant metastasis | No | Reference | |
| Yes | 2.23 (0.92, 5.43) | .077 | |
| Unknown | 8.65 (0.44, 170.50) | .16 | |
| Lymph node metastasis | No | Reference | |
| Yes | 1.60 (0.75, 3.41) | .22 | |
| Unknown | 0.20 (0.01, 5.85) | .35 | |
| Death | |||
| Pathology | FLC | Reference | |
| Conventional HCC | 2.20 (1.29, 3.76) | .004 | |
| Ever considered disease free | No | Reference | |
| Yes | 0.15 (0.07, 0.29) | <.001 | |
| Unknown | 0.41 (0.15, 1.18) | .10 | |
| Time of surgical resection | At diagnosis | Reference | |
| After chemotherapy | 1.02 (0.51, 2.05) | .95 | |
| Resect (Time unknown) | 0.47 (0.05, 4.29) | .50 | |
| No resection | 3.45 (1.70, 7.02) | <.001 | |
| Surgical complications | None | Reference | |
| Any <30 d | 1.75 (0.92, 3.32) | .089 | |
| Surgical death <30 d | 9.34 (1.43, 61.16) | .02 |
Abbreviations: CI, confidence interval; FLC, fibrolamellar carcinoma; HCC, hepatocellular carcinoma; HR, hazard ratio.
Figure 1 illustrates Kaplan-Meier curves for survival. Although more patients who had cHCC died in the first 1 or 2 years, at later times, the survival curves for those who had cHCC and FLC nearly overlapped; patients with cHCC had 50% mortality at 50 months and those with FLC had 50% mortality at 60 months. Patients with HB-HCC had higher long-term survival than the other 2 histologic groups.
FIGURE 1.
Kaplan-Meier curves illustrate survival in patients who have fibrolamellar carcinoma (FLC), conventional hepatocellular carcinoma (cHCC), and hepatoblastoma with hepatocellular carcinoma features (HB-HCC).
DISCUSSION
In this multicenter, retrospective review of pediatric patients with hepatocellular neoplasms, including cHCC, FLC, and HB-HCC, we found that these histologic subtypes have distinct clinical patterns. In addition to these differences, we identified several predictors of mortality in these rare cancers, including PRETEXT stage IV tumors, involvement of the caudate lobe or major vascular structures, and multifocal or metastatic disease. Interestingly, elevated AFP was associated with mortality in our data set, which is in contrast to the higher mortality seen in patients who have HB with low AFP. We demonstrated that metastatic patterns differ between FLC and cHCC, lending support to the growing body of literature demonstrating that FLC is distinct from cHCC and should be treated as such.18 FLCs have a high frequency of lymphatic spread, suggesting that periportal lymph node sampling should be performed in all children undergoing resection, even if nodes are grossly normal, and all nodal stations that are enlarged by imaging criteria should be aggressively cleared. We also found that perioperative mortality was rare in our overall data set and that resection, which is an independent predictor of survival (HR, 0.15; P < .001), was critical to achieving disease-free status.
Although it is not yet clear which specific biologic changes portend the differential clinical behavior observed in FLC compared with cHCC, it is evident that they have distinct molecular profiles. FLCs express a fusion transcript, including a heat-shock protein (DNAJ, DNAJB1)19 and the catalytic subunit of protein kinase A (PKAc, PRKACA).20 This fusion event has not been reported in cHCC, pediatric or otherwise, and is believed to be pathognomonic for FLC. The DNAJB1-PRKACA fusion changes PKA activity and phosphorylation targets,21 in addition to subsequent downstream transcriptomic changes.22 Conversely, the genomic landscape of pediatric cHCC is heterogeneous and includes molecular alterations in genes that are frequently mutated in adult HCC, including CTNNB1 and TERT, although TP53 mutations appear to be less common in pediatric versus adult cHCCs.23,24
Analyzing differences in outcomes between these types of hepatocellular neoplasms is complex, because both resection rates and transplantation rates differ between cell types. Early studies suggested improved survival in patients who had FLC compared with those who had other HCCs.7,9,25,26 Later work supported the notion that improved outcome in FLC is because of the absence of cirrhosis, thus patients with FLC can tolerate extensive resections.18,27 The majority of patients with pediatric cHCC in our data set had normal background liver (Table 1) and would also be expected to tolerate extensive resections. Differences in survival curves between these 2 types were apparent early in follow-up and disappeared later. Patients who had cHCC had a decreased frequency of surgical resection and a shorter time interval to progression (Table 2) compared with those who had FLC. The observation that patients who developed cHCC in the setting of underlying liver disease had lower rates of relapse compared with cHCCs arising in normal liver may be due to the higher rate of OLT in these patients or the finding that they were less likely to have metastases at presentation. We do not know which patients in our cohort underwent screening, which is more common in children who have underlying liver disorders and includes liver ultrasound and AFP measurement every 6 months. Screening may be a key contributor to the observed improved survival from <10% to >50% that has been observed in pediatric HCC over the last 4 decades.28,29
One limitation of our study is the lack of histologic review by a central expert panel. These rare and histologically challenging tumors can be difficult to classify as cHCC versus FLC versus HB-HCC. However, unlike some prior pediatric studies that analyzed national deidentified databases (eg, Surveillance, Epidemiology, and End Results Program) of pediatric HCC,6 we were able to review the granular details of each individual patient’s pathology report. We suspect that some of the patients with FLC who had elevated AFP may in fact have had cHCC because FLCs are generally not believed to secrete AFP.18,30 A recent publication reviewed a series of FLC cases from the National Cancer Database and concluded that patients with FLC who had elevated serum AFP had a poorer prognosis.31 However, the National Cancer Database lacks confirmation and/or central review of histologic diagnosis, so we suspect that some conclusions could have arisen because of tumor misclassification. Our study is also limited by the lack of molecular characterization of the majority of tumors because non-HB tumors were not included in the most recently completed study from COG (AHEP0731), and the DNAJB1-PRKACA fusion that defines FLC was not yet known during the majority of our study period, thus this molecular testing was not performed on all tumors. Other limitations include inherent challenges of accumulating a retrospective data set from a large number of institutions over a long-time interval. Although it is necessary to enroll sufficient numbers, this approach results in several potential limitations that warrant discussion. In some cases, there were gaps in the data set, such as the ability to distinguish or delineate the cause of mortality (progression, relapse, or other). Despite best efforts to fill in these gaps by surgeons at each institution, these data were in some instances unavailable. Consequently, several relevant data points include an unknown category, limiting the strength of our findings.
Despite overall improvements in outcomes for children with cancer over the past decades, hepatocellular neoplasms remain particularly fatal. Adult patients with unresectable cHCC and adequate performance status are currently managed with targeted therapies (eg, atezolizumab, lenvatinib, others), with or without immune checkpoint inhibition; the management of these patients is rapidly evolving.32 However, patients with FLC have shown minimal to no response to targeted agents that have efficacy in cHCC.33 On the basis of molecular analyses, clinical trials of mTOR inhibition and estrogen suppression34 or aurora A kinase inhibition35 in FLC have been performed; unfortunately, neither trial demonstrated a significant clinical benefit. The ongoing Pediatric Hepatic International Tumor Trial (PHITT) (COG-AHEP1531) is enrolling children with HCC (including FLC) from North America (COG), Europe (the International Society of Paediatric Oncology), and Japan (the Japan Children’s Cancer Group) and will enhance our understanding of chemotherapeutic responses in these patients. PHITT is limited in that all histologic types of HCC are grouped together; this issue is ameliorated by a companion biologic study in which molecular profiles of all of these tumors will be interrogated. Another current trial (ClinicalTrials.gov identifier NCT04248569) examines the efficacy of a peptide vaccine against the DNAJB1-PRKACA fusion combined with checkpoint inhibition in FLC; results from that trial are pending. A trial of checkpoint inhibition alone in pediatric HCC is enrolling (ClinicalTrials.gov identifier NCT04134559); we expect that these pending data, in conjunction with ongoing bench research, will better inform future treatment algorithms.20,36 We hope that increased awareness of FLC and pediatric cHCC as distinct diseases will lead to additional efforts to develop novel therapies for these patients because repeat, aggressive resection (including metastatic disease in FLC) is currently their most effective option.
In summary, using a robust data set containing detailed clinical information on 262 children with hepatocellular cancers, we noted different patterns of metastatic spread in cHCC and FLC and identified multifocality as a novel predictor of relapse. Inability to achieve disease-free status is the most important predictor of mortality in these patients. This data set will next be used to facilitate a central pathology review and molecular analyses of the tumors, which we anticipate will ultimately improve tailored treatment and subsequent outcomes for children with these challenging liver tumors.
Supplementary Material
FUNDING SUPPORT
This investigation was supported by the University of Utah Population Health Research Foundation, with funding in part from the National Center for Advancing Translational Sciences of the National Institutes of Health under award UL1TR002538.
CONFLICT OF INTEREST DISCLOSURES
Max R. Langham reports travel support from the Children’s Oncology Group outside the submitted work. Elizabeth A. Fialkowski reports a research grant from Oregon Health and Science University outside the submitted work. Kimberly J. Riehle reports a grant from the US Department of Defense and travel support from the Fibrolamellar Foundation outside the submitted work. The remaining authors made no disclosures.
Footnotes
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Additional supporting information may be found in the online version of this article.
REFERENCES
- 1.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;18:122–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Czauderna P, Haeberle B, Hiyama E, et al. The Children’s Hepatic Tumors International Collaboration (CHIC): novel global rare tumor database yields new prognostic factors in hepatoblastoma and becomes a research model. Eur J Cancer. 2016;52:92–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Aronson DC, Meyers RL. Malignant tumors of the liver in children. Semin Pediatr Surg. 2016;25:265–275. [DOI] [PubMed] [Google Scholar]
- 4.Shi Y, Rojas Y, Zhang W, et al. Characteristics and outcomes in children with undifferentiated embryonal sarcoma of the liver: a report from the National Cancer Database. Pediatr Blood Cancer. 2017;64:e26272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.McAteer JP, Goldin AB, Healey PJ, Gow KW. Hepatocellular carcinoma in children: epidemiology and the impact of regional lymphadenectomy on surgical outcomes. J Pediatr Surg. 2013;48:2194–2201. [DOI] [PubMed] [Google Scholar]
- 6.McAteer JP, Goldin AB, Healey PJ, Gow KW. Surgical treatment of primary liver tumors in children: outcomes analysis of resection and transplantation in the SEER database. Pediatr Transplant. 2013;17:744–750. [DOI] [PubMed] [Google Scholar]
- 7.Allan BJ, Wang B, Davis JS, et al. A review of 218 pediatric cases of hepatocellular carcinoma. J Pediatr Surg. 2014;49:166–171; discussion 171. [DOI] [PubMed] [Google Scholar]
- 8.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;27:472–491. [DOI] [PubMed] [Google Scholar]
- 9.Weeda VB, Murawski M, McCabe AJ, et al. Fibrolamellar variant of hepatocellular carcinoma does not have a better survival than conventional hepatocellular carcinoma—results and treatment recommendations from the Childhood Liver Tumour Strategy Group (SIOPEL) experience. Eur J Cancer. 2013;49:2698–2704. [DOI] [PubMed] [Google Scholar]
- 10.Murawski M, Weeda VB, Maibach R, et al. Hepatocellular carcinoma in children: does modified platinum- and doxorubicin-based chemotherapy increase tumor resectability and change outcome? Lessons learned from the SIOPEL 2 and 3 studies. J Clin Oncol. 2016;34:1050–1056. [DOI] [PubMed] [Google Scholar]
- 11.Schmid I, Haberle B, Albert MH, et al. Sorafenib and cisplatin/doxorubicin (PLADO) in pediatric hepatocellular carcinoma. Pediatr Blood Cancer. 2012;58:539–544. [DOI] [PubMed] [Google Scholar]
- 12.Polites SF, Heaton TE, La Quaglia MP, et al. Pneumonectomy for pediatric tumors—a Pediatric Surgical Oncology Research Collaborative study. Ann Surg. 2022;274:e605–e609. [DOI] [PubMed] [Google Scholar]
- 13.Overman RE, Kartal TT, Cunningham AJ, et al. Optimization of percutaneous biopsy for diagnosis and pretreatment risk assessment of neuroblastoma. Pediatr Blood Cancer. 2020;67:e28153. [DOI] [PubMed] [Google Scholar]
- 14.Towbin AJ, Meyers RL, Woodley H, et al. 2017 PRETEXT: radiologic staging system for primary hepatic malignancies of childhood revised for the Paediatric Hepatic International Tumour Trial (PHITT). Pediatr Radiol. 2018;48:536–554. [DOI] [PubMed] [Google Scholar]
- 15.Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol. 2007;165:710–718. [DOI] [PubMed] [Google Scholar]
- 16.Cox DR. Regression models and life-tables. J R Stat Soc B. 1972;34: 187–220. [Google Scholar]
- 17.Bouliotis G, Billingham L. Crossing survival curves: alternatives to the log-rank test. Trials. 2011;12:A137. [Google Scholar]
- 18.Riggle KM, Turnham R, Scott JD, Yeung RS, Riehle KJ. Fibrolamellar hepatocellular carcinoma: mechanistic distinction from adult hepatocellular carcinoma. Pediatr Blood Cancer. 2016;63:1163–1167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Honeyman JN, Simon EP, Robine N, et al. Detection of a recurrent DNAJB1-PRKACA chimeric transcript in fibrolamellar hepatocellular carcinoma. Science. 2014;343:1010–1014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kastenhuber ER, Lalazar G, Houlihan SL, et al. DNAJB1-PRKACA fusion kinase interacts with beta-catenin and the liver regenerative response to drive fibrolamellar hepatocellular carcinoma. Proc Natl Acad Sci U S A. 2017;114:13076–13084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Turnham RE, Smith FD, Kenerson HL, et al. An acquired scaffolding function of the DNAJ-PKAc fusion contributes to oncogenic signaling in fibrolamellar carcinoma. Elife. 2019;8:e44187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Simon EP, Freije CA, Farber BA, et al. Transcriptomic characterization of fibrolamellar hepatocellular carcinoma. Proc Natl Acad Sci U S A. 2015;112:E5916–E5925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Haines K, Sarabia SF, Alvarez KR, et al. Characterization of pediatric hepatocellular carcinoma reveals genomic heterogeneity and diverse signaling pathway activation. Pediatr Blood Cancer. 2019;66:e27745. [DOI] [PubMed] [Google Scholar]
- 24.Nagae G, Yamamoto S, Fujita M, et al. Genetic and epigenetic basis of hepatoblastoma diversity. Nat Commun. 2021;12:5423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Eggert T, McGlynn KA, Duffy A, Manns MP, Greten TF, Altekruse SF. Fibrolamellar hepatocellular carcinoma in the USA, 2000–2010: a detailed report on frequency, treatment and outcome based on the Surveillance, Epidemiology, and End Results database. United European Gastroenterol J. 2013;1:351–357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Groeschl RT, Miura JT, Wong RK, et al. Multi-institutional analysis of recurrence and survival after hepatectomy for fibrolamellar carcinoma. J Surg Oncol. 2014;110:412–415. [DOI] [PubMed] [Google Scholar]
- 27.Lim IIP, Farber BA, LaQuaglia MP. Advances in fibrolamellar hepatocellular carcinoma: a review. Eur J Pediatr Surg. 2014;24:461–466. [DOI] [PubMed] [Google Scholar]
- 28.Khanna R, Verma SK. Pediatric hepatocellular carcinoma. World J Gastroenterol. 2018;24:3980–3999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kelly D, Sharif K, Brown RM, Morland B. Hepatocellular carcinoma in children. Clin Liver Dis. 2015;19:433–447. [DOI] [PubMed] [Google Scholar]
- 30.Berman MA, Burnham JA, Sheahan DG. Fibrolamellar carcinoma of the liver: an immunohistochemical study of nineteen cases and a review of the literature. Hum Pathol. 1988;19:784–794. [DOI] [PubMed] [Google Scholar]
- 31.McDonald JD, Gupta S, Shindorf ML, et al. Elevated serum alpha-fetoprotein is associated with abbreviated survival for patients with fibrolamellar hepatocellular carcinoma who undergo a curative resection. Ann Surg Oncol. 2020;27:1900–1905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gordan JD, Kennedy EB, Abou-Alfa GK, et al. Systemic therapy for advanced hepatocellular carcinoma: ASCO guideline. J Clin Oncol. 2020;38:4317–4345. [DOI] [PubMed] [Google Scholar]
- 33.Ang CS, Kelley RK, Choti MA, et al. Clinicopathologic characteristics and survival outcomes of patients with fibrolamellar carcinoma: data from the fibrolamellar carcinoma consortium. Gastrointest Cancer Res. 2013;6:3–9. [PMC free article] [PubMed] [Google Scholar]
- 34.El Dika I, Mayer RJ, Venook AP, et al. A multicenter randomized three-arm phase II study of (1) everolimus, (2) estrogen deprivation therapy (EDT) with leuprolide + letrozole, and (3) everolimus + EDT in patients with unresectable fibrolamellar carcinoma. Oncologist. 2020;25:925–e1603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Abou-Alfa GK, Mayer R, Venook AP, et al. Phase II multicenter, open-label study of oral ENMD-2076 for the treatment of patients with advanced fibrolamellar carcinoma. Oncologist. 2020;25: e1837–e1845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Lalazar G, Requena D, Ramos-Espiritu L, et al. Identification of novel therapeutic targets for fibrolamellar carcinoma using patient-derived xenografts and direct-from-patient screening. Cancer Discov. 2021;11:2544–2563. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.