Fibrolamellar hepatocellular carcinoma (FL-HCC) is a rare variant of hepatocellular carcinoma, has distinct pathologic features, and typically occurs in young patients without underlying hepatitis or cirrhosis.
Fibrolamellar hepatocellular carcinoma (FL-HCC) is a rare variant of hepatocellular carcinoma, has distinct pathologic features, and typically occurs in young patients without underlying hepatitis or cirrhosis.
Forty-one patients with the pathologic diagnosis of FL-HCC evaluated at our institution between 1986 and 2003 were identified from a prospective database.
Median age of all patients was 27 years. None of these patients had underlying hepatitis or cirrhosis, and only 3 (7%) patients had an α-fetoprotein level > 200 ng/mL. Twenty-eight patients with primary disease underwent complete gross resection, and 13 patients were unresectable. In patients treated with resection, median tumor size was 9 cm (range, 3–17), 9 (36%) had vascular invasion, and 14 (50%) had lymph node metastases. There were no perioperative deaths. With a median follow-up of 34 months, 5-year overall survival for resected patients was 76%. However, 5-year recurrence-free survival was only 18%, and of the 9 resected patients with more than 5 years of follow-up, 7 had recurrences. Lymph node metastasis was the only significant negative prognostic factor. Seventeen (61%) patients underwent a second operation for recurrent disease. Median survival for unresected patients with FL-HCC was only 12 months, and no patient survived beyond 5 years.
FL-HCC occurs in a distinctly different population of patients than common HCC, and patients with FL-HCC generally fare better after complete resection. These tumors have a relatively indolent tumor biology, and late recurrences are common. Repeat resections for recurrence should be considered given the lack of other effective treatment options. Cancer 2006. © 2006 American Cancer Society.
In 2005, there were an estimated 18,920 new primary liver (including hepatocellular carcinoma) and intrahepatic bile duct cancers in the United States.1 El-Serag et al. recently reviewed Fibrolamellar hepatocellular carcinoma (FL-HCC) cases in the Surveillance, Epidemiology, and End Results data, which captures tumors that occur in about 14% of the United States population.2 They reported 68 cases of FL-HCC over a 15-year period. Over the same time period, 7896 cases of common HCC were recorded, so FL-HCC cases represented 0.9% of all HCC cases reported. The mean age of FL-HCC patients was 39 years, and there was an equal gender distribution. Five-year survival was 31.8% for FL-HCC compared with 6.8% for common HCC.
FL-HCC is a rare variant of hepatocellular carcinoma (HCC) initially described by Edmondson in 1956.3 Tumors are usually well circumscribed masses characterized by well differentiated polygonal hepatic cells with eosinophilic and granular cytoplasm surrounded by thick, fibrous stroma arranged in bands.4 In addition to histologic differences to common HCC, FL-HCC occurs in younger patients, with a median age of about 25 years.5 Most series report an equal number of men and women,6 unlike common HCC, which occurs 4–8 times more often in men.2 Common HCC usually occurs in the setting of chronic hepatitis or cirrhosis,7 whereas FL-HCC usually occurs in normal livers.8 In addition, elevations in alpha-fetoprotein (α-fetoprotein [AFP]) levels are uncommon with FL-HCC, but elevations in neurotensin, vitamin B12 binding capacity, and Des-gamma-carboxy prothrombin have been reported.5, 8–10
Initial series of FL-HCC from the 1980s described this tumor as more indolent than common HCC with a slower rate of growth and a more favorable prognosis.4, 11, 12 Other subsequent studies found that survival after resection was similar in patients with FL-HCC and common HCC, and these studies suggested that the improved prognosis of patients with FL-HCC was related to a higher rate of resectability.13, 14 More recent surgical series have again reported significantly better survival for patients after resection of FL-HCC than after resection of common HCC.5, 15 In the present study, we examined the outcome of patients with FL-HCC who were referred to a tertiary-care cancer center over an 18-year period.
By using the prospective hepatobiliary database at the Memorial Sloan-Kettering Center Cancer, 41 patients with a pathologic diagnosis of FL-HCC were identified over an 18-year period between 1986 and 2003. Of this group, 28 patients with primary disease underwent complete gross resection, and 13 patients were deemed unresectable. During the same period, 193 patients with common HCC who underwent curative resection were also identified. Clinical information was collected from the database, office charts, and patient interviews. Information collected included patient demographics, results of imaging studies, serology, surgical procedures, pathology, hospital course, and outcome. The presence of hepatitis C antibody and/or hepatitis B surface antigen was considered evidence for parenchyma disorder. Follow-up was obtained by personal contact with the patients, patients' families, or family physicians.
All patients were reviewed at a twice-weekly multidisciplinary hepatobiliary conference. The extent of resection was defined according to Goldsmith and Woodburne.16 Couinaud nomenclature was followed: an extended right hepatectomy is resection of segments 4 through 8; an extended left hepatectomy is resection of segments 2, 3, 4, 5 and 8; a right hepatectomy is resection of segments 5 through 8; and a left hepatectomy is resection of segments 2 through 4.17 Clinical staging of disease was performed using the new American Joint Committee on Cancer (AJCC) staging criteria.18
Survival analysis was calculated with the Kaplan–Meier method19 and compared by the log-rank test using SPSS 11.0.1 (SPSS Inc., Chicago, IL). Differences were considered significant at P < 0.05.
Over an 18-year period, 41 patients with FL-HCC were identified in our prospective hepatobiliary database with a median age of 27 years (Table 1). There were 17 males and 24 females. None of the FL-HCC patients had underlying hepatitis or cirrhosis, and AFP levels were elevated in only 3 (7%) patients. During a comparable time period, 412 patients with hepatocellular carcinoma were evaluated at our institution.20 In contrast to FL-HCC, common HCC patients were predominantly male, had a 47% and a 70% rate of hepatitis and cirrhosis, respectively, and AFP levels were elevated in most patients.
|FL-HCC (n= 41)|
|Median age, yrs (range)||27 (14-72)|
|Hepatitis or cirrhosis||0|
|Elevated AFP||3 (7%)|
Patients typically presented with nonspecific abdominal pain or discomfort, or they were asymptomatic and had their tumors discovered incidentally during workup of an unrelated medical condition. These tumors are usually large, well circumscribed, heterogeneous lesions.21 They often have areas of hypervascularity and a central scar.7 Figure 1 demonstrates the computed tomography (CT) of 1 representative 21 year-old patient who presented with a 9 cm mass in the left liver and bulky celiac lymphadenopathy. This patient underwent a left hepatectomy along with portal and celiac lymphadenectomy and has been without evidence of recurrence for more than 2 years.
Twenty-eight patients underwent resection of their FL-HCC, and 13 patients were unresectable either because of extensive liver involvement or because of the presence of extrahepatic disease. Of the 28 patients resected, 25 had just 1 tumor and only 3 patients had more than 1 tumor (Table 2). The median size of the largest tumor was 9 cm, and the tumors were relatively equally distributed between the left and right lobes. Five tumors were located centrally in the liver such that they involved both the right and left lobes.
|Resected FL-HCC (n = 28)|
|Solitary liver tumor||25 (89%)|
|Largest tumor||9 (3-17) cm|
|Location||10 right, 13 left, 5 bilateral|
|Sublobar resection||7 (25%)|
|Extended hepatectomy||5 (18%)|
|Nodal metastases||14 (50%)|
|Vascular invasiona||9 (36%)|
|Positive microscopic marginb||5 (18%)|
|Repeat hepatectomy||17 (61%)|
Of the 28 resected patients, 9 had a preoperative CT-guided biopsy. Of these 9 biopsies, 5 showed FL-HCC, and 4 were thought to be alternative diagnoses (2 showed hepatoma, and 2 showed adenoma). Four patients had a frozen section at the time of operation: 3 of these were diagnostic for FL-HCC, and 1 was read as common HCC.
The extent of surgery for the 28 resected patients is listed in Table 2. Three-quarters of patients required either a hepatectomy or extended hepatectomy to remove their tumors. One-half of resected patients underwent a formal portal lymphadenectomy, and 3 additional patients had resection of grossly involved lymph nodes without formal portal lymphadenectomy. There were no perioperative deaths.
After pathologic analysis, 3 patients were diagnosed with an unusual mixed morphology of pure fibrolamellar features and conventional well differentiated HCC. The remainder of resected tumors showed pure fibrolamellar features. Vascular invasion was seen in 36% of cases. Of these 9 tumors, 8 showed microscopic vascular invasion, and 1 showed macroscopic invasion of the inferior vena cava. One-half of patients had lymph node metastases at the time of resection. The margins of the resected specimens were not invaded by tumor cells in 20 patients, were microscopically involved in 5 patients, and were not determined in 3 patients. The margin was microscopically positive because of proximity of the tumor to unresectable vital structures.
Patients were staged according to the American Joint Commission on Cancer (AJCC) staging criteria.18 In this staging system, Stage I disease is a single tumor of any size with no lymph node involvement and no vascular invasion, and Stage II disease is a single tumor less than 5 cm in size with vascular invasion but no lymph node involvement. Nine (36%) tumors had either microscopic or gross vascular invasion (Table 2). Most patients in this series without lymph node metastases had no vascular invasion and were, therefore, Stage I, and only 3 patients were Stage II. Stage III is comprised of patients with multiple tumors, tumors > 5 cm and invading a major vessel, or any T stage and positive lymph nodes; one-half of patients were in this category. One patient was Stage IV because of a single peritoneal implant found at the time of surgery, and this implant was resected along with the liver tumor.
Seventeen patients underwent a second operation for recurrence after a median interval of 37 months. Operations included repeat liver resection (with or without resection of other nodal or intraabdominal disease) in 6 patients, resection of nodal or intraabdominal disease in 8 patients, and resection of lung or mediastinal disease in 3 patients. The median survival of these patients after the second resection was 26 months.
Thirteen patients were found to be unresectable. Nine of these patients were initially evaluated at outside institutions, and 4 presented initially to our institution. Six of the 13 patients underwent exploratory laparotomy and were deemed unresectable because of extensive liver involvement or peritoneal metastases. Seven patients were found to be unresectable on the basis of preoperative imaging, and they underwent only biopsy.
After a median follow-up of 34 months, the median OS of the 28 patients who underwent resection was 112 months, and 5-year survival was 76% (Fig. 2). In a recent review of our overall experience with hepatocellular carcinoma (n = 193), resection achieved a median survival time of 37 months, with 5-year survival of 37%.20 For the 13 patients with unresectable disease, median survival was 12 months, and no patient survived beyond 5 years.
Univariate analysis was performed for OS on resected patients for the following variables: size of tumor, number of tumors, lymph node metastases, vascular invasion, stage of disease, and resection margin (Table 3). Only the presence of lymph node metastases was significant (Fig. 3). When stratified by nodal status, lymph node-negative patients had a 100% 5-year survival and 20% 10-year survival compared with lymph node-positive patients, who had only a 45% 5-year survival. Despite a relatively high 5-year OS, the majority of resected patients recurred. Median recurrence-free survival was 33 months, and 5-year recurrence-free survival was only 18% (Fig. 4). Univariate analysis was performed for recurrence-free survival using the same variables described above (Table 3). The only significant negative prognostic factor for recurrence was the presence of lymph node metastases (Fig. 5 -fluorouracil). Three of the 5 patients with a positive microscopic margin recurred. Positive or negative margin was included in the univariate analysis and was not significant for either recurrence-free survival or OS.
|Factor||No. (%)||Median RFS (mos)||P||Median OS (mos)||P|
|< 10 cm||15 (54%)||27||112|
|≥ 10 cm||13 (46%)||46||NR|
|> 1||3 (11%)||33||112|
|< lobe||7 (25%)||33||NR|
|≥ lobe||21 (75%)||27||112|
Recurrence and death even after 5 years is common. There were 9 patients in this study with greater than 5-years of follow-up. Of these patients, 3 died of disease at 65, 69, and 112 months after resection. Four patients were alive with disease at 61–136 months after resection. Only 1 patient had no evidence of disease, and this patient was 115 months postresection.
Although HCC is one of the most common malignancies worldwide,6 the fibrolamellar variant is rare and occurs in a distinctly different group of patients. This series confirms findings of other studies that FL-HCC occurs in young patients and usually not in the setting of hepatitis, cirrhosis, or elevations in AFP level (Table 4). FL-HCC also appears to have a less aggressive tumor biology than common HCC. Despite a rather indolent course, FL-HCC frequently recurs after completely resection, often after 5 years or longer. For unresectable patients, median survival was only 12 months.
|Author, Yr||N||Agea||Male:female||Cirrhosis/hep B/hep C||AFP elevated||Median size||> 1 tumor||Positive nodes||Vascular. invasion||Initial operation||Repeat operation||Median f/u||5-yr survival||Prognostic factors|
|Ringe, 199214||20||23||55:45||0/10%/10%||0%||12||45%||45%||30%||Phx 70% OLT 30%||NR||NR||37%||> 1 tumor positive nodes|
|Hemming, 199715||10||31||50:50||NR||10%||8||20%||20%||NR||Phx 100%||50%||101||70%||NR|
|Pinna, 19975||41||30||56:44||7%/0%/6%||10%||13||27%||34%||76%||Phx 68% OLT 32%||22%||58||66%||Vascular invasion|
|El-Gazzaz, 200025||20||27||65:35||0% hep B||0%||14||20%||30%||55%||Phx 55% OLT 45%||NR||25||50%||None|
|Current series||28||28||43:57||0%/0%/0%||7%||9||11%||50%||36%||Phx 100%||61%||34||56%||positive nodes|
FL-HCCs, as with other liver tumors, are best delineated preoperatively by abdominal CT and magnetic resonance imaging (MRI) scans. These tumors are usually heterogeneous on CT imaging with areas of hypervascularity. On MRI, tumors are usually T1 hypointense and T2 hyperintense, and the use of a gadolinium contrast agent during MRI results in heterogeneous enhancement. In 1 series of 31 FL-HCC cases evaluated by CT and/or MRI, 77% of tumors had well defined margins, 68% had calcifications, 65% had abdominal lymphadenopathy, and 71% had a central scar.22 In terms of differentiating liver tumors with central scars, Blachar et al. reported that, in a group of 64 liver tumors including 20 FL-HCC, CT scan was highly accurate in differentiating FL-HCC from focal nodular hyperplasia and hemangioma.23 On MRI, the central scar of FL-HCC tumors has low attenuation on T2 images, whereas the central scar of focal nodular hyperplasia (FNH) tumors has high attenuation.
The diagnosis of FL-HCC can often be made by characteristic CT and MRI imaging findings. For indeterminate cases, CT-guided core needle biopsy or fine needle aspiration (FNA) can be used. FNA can be useful in differentiating FL-HCC from common HCC.24 Biopsy may be necessary for patients who are unresectable or who have underlying medical conditions that preclude resection. Nine patients in our study received a preoperative biopsy. Patients were often referred from outside institutions after a biopsy had already been performed. Preoperative imaging with MRI often made the clinical diagnosis of FL-HCC, and preoperative biopsy was avoided. Patients with significant symptoms were resected without preoperative biopsy. Thus, biopsy of resectable lesions was reserved for those with asymptomatic or minimally symptomatic tumors with an indefinite radiologic diagnosis.
The best treatment for FL-HCC is surgical resection whenever possible. Resection in this series resulted in a 5-year survival of 76%, which compares favorably to the 37–70% 5-year survival rates seen in other surgical series.4, 14, 15, 23 Given that FL-HCC tumors are usually well defined and occur in young patients with healthy livers, the resectability rate may be higher than that for common HCC, despite the finding that FL-HCC tumors are usually quite large. At our institution, 68% of patients underwent resection, and 32% of patients were unresectable. In our series of common HCC, only 37% of patients underwent resection.19
Orthotopic liver transplantation has been used in patients with FL-HCC (Table 4).3, 14, 25, 26 Patients usually have better survival after partial hepatectomy than liver transplantation,14, 25 which is likely a result of more aggressive and larger tumors being treated by transplantation. However, the immunosuppression required after transplantation may increase the recurrence rate and decrease survival.27 Two patients in this study with unresectable disease were referred for liver transplantation.
Several surgical series of FL-HCC have identified factors that predict worse survival after resection. Multiple tumors, vascular invasion, and positive lymph nodes have been demonstrated to be negative prognostic factors.4, 14 Our series found only positive lymph nodes to be a significant negative prognostic factor. Patients with Stage I or II disease tended to fare better than patients with Stage III or IV disease, but this difference did not attain statistical significance (P = 0.071). The AJCC staging system that is applied to FL-HCC was developed for common HCC. Thus, although AJCC staging for HCC reliably predicts survival for common HCC, it is not surprising that this staging system is not the best predictor of survival in FL-HCC, which has a distinctly different tumor biology.
The reason for the high rate of lymph node metastases with FL-HCC compared with common HCC is unknown. Lymph node metastasis may be related to tumor size, and FL-HCCs usually are larger when discovered than common HCCs. Alternatively, common HCC occurs predominantly in cirrhotic livers, whereas FL-HCC occurs generally in normal livers, and the cirrhotic process may inhibit lymphatic outflow and subsequent lymphatic metastases.
Given the lack of effective alternative therapies for FL-HCC and its relatively indolent course, some authors advocate resection of patients with limited metastatic disease and repeat resections for recurrent disease. In a series of 41 FL-HCC patients who underwent resection from the University of Pittsburgh, 13 patients had metastatic disease, and the overall 5-year survival was still 66%.4 Only 1 patient in our series who underwent resection had metastatic disease at the time of surgical resection. However, 17 patients underwent a second operation aimed at either recurrent or metastatic disease. Surgeons involved in this study often took a relatively aggressive operative approach. This aggressive approach is illustrated by the patient described in Figure 1, who had bulky celiac lymphadenopathy resected along with his primary tumor. Thus, the spread of tumor beyond regional lymph nodes to even celiac nodes was not considered unresectable disease. Aggressive surgical resection in the presence of peritoneal disease or disease in more distant locations requires sound surgical judgment. One must consider that this is a slow-growing tumor, that nonsurgical treatment for this disease is relatively ineffective, and that patients are usually young and healthy. Because the efficacy of chemotherapy is questionable for this disease, no patient received adjuvant chemotherapy after complete resection of their tumor, and more effective adjuvant therapies are clearly needed.
Despite a relatively indolent tumor biology, FL-HCC commonly recurs after complete surgical resection, and sites of recurrence include the liver, regional lymph nodes, peritoneum, and lung.28 Over 60% of resected patients in our series underwent a second operation for recurrent disease. This rate is somewhat higher than the 20–50% rate reported in other series.4, 15 However, the median survival of these patients after the second resection was 26 months. Late recurrences beyond 5 years were common in our series. This is in contrast to 1 study where the 5-year and 10-year survival rates were equivalent at 70%,15 but similar to another study in which survival declined from 66% at 5 years to 47% at 10 years.4
Patients who underwent surgical resection fared much better than patients who were unresectable, and the median survival for unresected patients was only 12 months. By comparison, the median survival of patients with unresectable common HCC is slightly shorter, at about 9 months.20 In another series of 17 patients with unresectable FL-HCC, median survival was 14 months compared with 8 months for patients with unresectable common HCC.4 All patients in this series who were not resectable received chemotherapy. Drug regimens including isofosfamide, carboplatinum, etoposide, doxorubicin, flavopiridol, irinotecan, gemcitabine, and 5-fluorouracil (5-FU) were used. The treatment benefit for chemotherapy could not be assessed in our patients because of the variety of agents used and the small number of patients. Patt and colleagues evaluated the efficacy of 5-FU and interferon alfa-2b in FL-HCC patients,29 and 5 of 8 patients had a partial or complete response. This regimen warrants further investigation.
In conclusion, FL-HCC is a rare histologic variant of common HCC and should be suspected in patients with HCC who are young and who do not have underlying hepatitis or cirrhosis. These tumors have characteristic CT and MRI appearances that often allow their identification before resection. Aggressive initial surgical resection along with regional lymphadenectomy should be performed for primary disease. Repeat resection or orthotopic liver transplantation should be considered for these tumors given the lack of effective alternative therapies and relatively indolent tumor biology. Patients should be followed for at least 10 years given that late recurrences often occur.