Surgery for large primary liver cancer more than 10 cm in diameter

Abstract

Purpose

Large primary liver cancer (PLC) more than 10 cm in diameter is not infrequently encountered in clinical practice. This study evaluated the clinicopathological features and long-term results after surgery for large PLC.

Methods

Comparison of clinicopathological data between patients with PLC ≥10 cm (n=1,227) and PLC <10 cm (n=2,349) during the same period.

Results

In comparison with patients with PLC <10 cm, patients with PLC ≥10 cm were significantly younger (P<0.01), had a lower incidence of asymptomatic tumors (9.1% vs 39.5%, P<0.001), higher α-fetoprotein levels ( >400 ng/ml, 78.3% vs 49.2%, P<0.001), higher γ-glutamyl transpeptidase levels ( >6U, 87.7% vs 70.5%, P<0.001), a lower incidence of a history of hepatitis (45.0% vs 61.4%, P<0.001) and associated macronodular cirrhosis (cirrhotic nodules ≥0.3 cm, 59.8% vs 66.6%, P<0.001), poor differentiation of tumor cells (Edmondson grade 3–4, 24.3% vs 19.7%, P<0.01), a lower percentage of single nodule tumors (59.9% vs 75.4%, P<0.001) and well-encapsulated tumors (28.5% vs 62.1%, P<0.001), a higher proportion of tumor emboli in the portal vein (20.5% vs 9.0%, P<0.001), a lower resection rate (50.6% vs 86.8%, P<0.001), a lower curative resection rate (54.8% vs 78.3%, P<0.001), a higher operative mortality rate (4.5% vs 2.3%, P<0.001), and less local resection (52.5% vs 80.2%, P<0.001). The 5- and 10-year survival rates after resection were 26.2% and 17.5%, respectively, for patients with PLC ≥10 cm (n=621), and 54.3% and 39.5%, respectively, for patients with PLC <10 cm (n=2039) (P<0.01).

Conclusions

Large PLC had specific clinicopathological features. Surgery is the first choice of treatment. In selected patients, resection is safe and offers the chance of long-term survival. Large PLC does not exclude the possibility of cure.

Introduction

With the use of α-fetoprotein (AFP) serosurvey and ultrasonography monitoring in high-risk populations, small hepatocellular carcinoma (HCC) has been diagnosed and resected, which has resulted in a marked improvement in 5-year survival (Tang et al. 1989; Zhou et al. 2001). In spite of this, large HCC, with a diameter of more than 10 cm, is not infrequently encountered in clinical practice. In most of these patients with large HCC, the lesion is very advanced at the time of detection, and their long-term prognosis is generally poor. The incidence of large HCC is especially high in patients aged under 35 years (Zhou et al. 1995). Delayed diagnosis may be one explanation but the question remains whether there are specific biological features related to the development of a large HCC (Lee et al. 1998). At present, hepatic resection may be regarded as the only reasonable therapeutic option for a large HCC. However, only a few reports have described a small number of patients with HCC larger than 10 cm who underwent resection and survived for long periods following this intervention (Furuta et al. 1992; Noguchi et al. 1997; Lee et al. 1998). In this paper, we present a retrospective analysis of surgical treatment of 1,227 patients with primary liver cancer (PLC) 10 cm in diameter or more. The clinicopathological features and outcome were compared with those in 2,349 patients with PLC less than 10 cm in diameter within the same period. Some factors that influence or improve long-term survival for these patients are discussed herein.

Materials and methods

Patients

Between June 1964 and December 1999, 3,576 patients with PLC—1,227 patients (34.3%) with PLC measuring≥10 cm in the greatest dimension (group A) and 2,349 patients (65.7%) with PLC measuring <10 cm in the greatest dimension (group B)—were admitted to the Liver Cancer Institute of Shanghai Medical University. The 1,227 patients who had PLC≥10 cm were the basis for this study.

Of the 1,227 patients, 1,061 were male, and 166 were female, with a male/female ratio of 6.4/1.0. The patients' ages ranged from 9 years to 82 years, with a median age of 47 years (Table 1). Tumors were asymptomatic in 9.1% of patients and were discovered by AFP serosurvey and ultrasonography in the population with a history of hepatitis, cirrhosis, and/or serum hepatitis B surface antigen (HBsAg) background or during regular health checkups. The remaining tumors were discovered clinically with obvious symptoms and signs.

Table 1.

Comparison of clinicopathological data between patients with PLC ≥10 cm and PLC <10 cm. Because not all of the following clinical parameters were available, the numbers for some of the parameters do not equal the total numbers of patients. (AFP α-fetoprotein, γ-GTP γ-glutamyl transpeptidase, HBsAg hepatitis B surface antigen, NS no statistical significance)

Items	Group A (%) (≥10 cm)		Group B (%) (<10 cm)	P
No. of patients	1,227		2,349	–
Male	86.5(1,061 of 1,227)		88.1(2,070 of 2,349)	NS
Agea (years)	47(9–82)		50(10–78)	<0.01
Asymptomatic	9.1(112 of 1,227)		39.5(929 of 2,349)	<0.001
Serum AFP(ng/ml)
≤20	26.3(313 of 1,190)		33.0(764 of 2,318)	<0.001
>20	73.7(877 of 1,190)		67.0(1554 of 2,318)
>400	78.3(687 of 877)		49.2(764 of 1,554)
Serum γ-GTP (U)
≤6	12.3(146 of 1,184)		29.5(674 of 2,284)	<0.001
>6	87.7(1,038 of 1,184)		70.5(1,610 of 2,284)
History of hepatitis
Absence	55.0(671 of 1,221)		38.6(905 of 2,344)	<0.001
Presence	45.0(550 of 1,221)		61.4(1,439 of 2,344)
>5 years	83.6(460 of 550)		85.1 (1,224 of 1,439)
Cirrohosis
Absence	18.4(215 of 1,168)		12.7(294 of 2,318)	<0.001
Presence	81.6(953 of 1,168)		87.3(2,024 of 2,318)
Macronodular	59.8(570 of 953)		66.6(1,347 of 2,024)
Serum HBsAg
Absence	24.5(266 of 1,087)		23.0(504 of 2,195)	NS
Presence	75.5(821 of 1,087)		77.0(1,691 of 2,195)
Tumor number
Single	59.9(735 of 1,227)		75.4(1,772 of 2,349)	<0.001
More than one	40.1(492 of 1,227)		24.6(577 of 2,349)
Tumor capsule
Good	28.5(273 of 958)		62.1(1,301 of 2,096)	<0.001
Poor	71.5(685 of 958)		37.9(795 of 2,096)
Tumor emboli in portal vein
Absence	79.5(967 of 1,217)		91.0(2,136 of 2,346)	<0.001
Presence	20.5(250 of 1,217)		9.0(201 of 2,346)
Edmondson grade
1	3.2(26 of 806)	4.5(81 of 1,799)		<0.01
2	72.5(584 of 806)	75.8(1364 of 1,799)
3–4	24.3(196 of 806)	19.7(354 of 1,799)

^aValue expressed in median with range in parentheses

The serum concentration of AFP was abnormal (>20 ng/ml, immunoassay) in 73.7% of patients (877 of 1,190 patients). The serum AFP concentration was >400 ng/ml in 78.3% of patients (687 of 877 patients). The serum γ-glutamyl transpeptidase (γ-GTP) levels, for which the normal value is <6 U/ml (Orlowski method), was elevated in 87.7% of patients (1,038 of 1,184 patients).

The background of liver disease among patients was as follows: 45.0% (550 of 1,221 patients) had a history of hepatitis; of these, 83.6% (460 of 550 patients) had a history lasting over 5 years. Coexisting cirrhosis occurred in 81.6% (953 of 1,168 patients), of which 59.8% (570 of 953 patients) had macronodular cirrhosis (cirrhotic nodules measuring at least 0.3 cm in the greatest dimension), and 40.2%(383 of 953 patients) had micronodular cirrhosis (cirrhotic nodules measuring <0.3 cm in the greatest dimension). Serum HBsAg, which was assayed by reverse passive hemagglutination, was observed in 75.5% (821 of 1,087 patients). Because not all of these clinical background items were available, the numbers for some of the items reported here do not equal the total numbers of patients. Indications for surgery included patients with compensated liver function, including serum bilirubin <20 ng/ml, prothrombin time >50% of the normal value, a preserved albumin/globulin ratio, and an alanine aminotransferase level less than twice the normal level.

Therapeutic modalities employed in 1,227 patients were as follows: resection in 621 patients (50.6%); hepatic artery ligation (HAL) and hepatic artery cannulation (HAC) with infusion chemoembolization with or without radiotherapy (Zhou et al. 1991a) in 495 patients (40.3%); cryosurgery (Zhou et al. 1998), microwave coagulation (Zhou et al. 1993a) or Nd:YAG (yttrium-aluminum- garnet) laser vaporization(Yu et al. 1986) in 35 patients (2.9%); conservative treatment (radiotherapy, chemotherapy, and/or Chinese herbs) in 44 patients (3.6%); and no treatment in 32 patients (2.6%). Chemotherapeutic agents commonly used in this series included cisplatin, adriamycin, epirubicin, mitomycin C, 5-fluorouracil or fluorodeoxyuridine, followed by injection of Lipiodol via hepatic artery cannulation in some patients with unresectable PLC ≥10 cm.

The types of resection employed in the 621 patients were as follows: local resection (any kind of nonsegmental resection) in 326 patients (52.5%); left lateral segmentectomy in 99 patients (15.9%); left hemihepatectomy in 129 patients (20.8%); extended left hemihepatectomy in 12 patients (1.9%); right hemihepatectomy in 52 patients (8.4%); and extended right hemihepatectomy in 3 patients (0.5%). Curative resection was done in 340 patients (54.8%) (Table 2). Curative resection refers to the complete removal of the tumor, with no macroscopically identified tumor emboli in the portal vein and no tumor residue in the remaining liver tissue or in the cut surface.

Table 2.

Comparison of type of hepatic resection between patients with PLC ≥10 cm and PLC <10 cm

Resection	Group A (%) (≥10 cm)	Group B (%) (<10 cm)	P
Local resectiona	52.5(326 of 621)	80.2(1,636 of 2,039)	<0.001
Left lateral segmentectomy	15.9(99 of 621)	10.0(203 of 2,039)
Left hemihepatectomy	20.8(129 of 621)	8.0(163 of 2,039)
Extended left hemihepatectomy	1.9(12 of 621)	0.3(7 of 2,039)
Right hemihepatectomy	8.4(52 of 621)	1.5(30 of 2,039)
Extended right hemihepatectomy	0.5(3 of 621)	0
Resection rate	50.6(621 of 1227)	86.8(2,039 of 2,349)	<0.001
Curative resection rate	54.8(340 of 621)	78.3(1,596 of 2,039)	<0.001
Operative mortality	4.5(28 of 621)	2.3(46 of 2,039)	<0.001

^aLocal resection: any kind of nonsegment resection

All were pathologically proved PLCs, HCC amounting to 96.8%, cholangiocarcinoma 2.0%, and mixed type 1.2%. Tumor size ranged from 10 cm to 29 cm, with a median size of 13 cm. Single-nodule tumors were found in 59.9% (735 of 1,227 patients), macroscopically encapsulated tumors were found in 28.5% (273 of 958 patients) of those examined, and tumor emboli presented in the portal vein in 20.5% (250 of 1,217 patients) of those examined(Table 1).

Surgical technique

The operative procedure has been described elsewhere (Zhou et al. 1994a; Zhou et al. 2001). In recent years, intraoperative ultrasound has been used to exclude previously undetected lesions and/or to localize satellite tumors that are situated under the liver surface and are not palpable. For control of bleeding during transection of the hepatic parenchyma, temporary occlusion of the portal pedicle (the Pringle maneuver) was used frequently in this series. Generally, the occlusion time should not exceed 20 min for patients with mild-to-moderate cirrhosis and 10 min for patients with severe cirrhosis. Repeated occlusion was needed sometimes for a complicated resection with 3–5 min of perfusion. Hepatic parenchyma transection was performed using finger-fracture, hemostatic clamp tip, or surgical knife handle. In patients who had large PLC in the right lobe of the liver and underwent major hepatic resection, the technique of "anterior" approach (or "retrograde" approach) was sometimes used (Lai et al. 1996; Wu et al. 1999; Liu et al. 2000). Without prior mobilization of the right lobe of the liver and the tumor, parenchyma transection was performed from the anterior surface of the liver until the anterior surface of the inferior vena cava was exposed. All venous tributaries, including the right hepatic vein, were controlled before the right lobe of the liver was mobilized. Resection was done usually 1 cm from the tumor capsule except in some patients who had a tumor near the major portal vein, hepatic vein, or inferior vena cava. The raw surface of the liver is either sutured side to side to reduce the raw area and to provide a tamponade effect on the smaller vessels, or it is covered with a piece of omentum or faciform ligament. Adequate drainage is important.

Routine preoperative medication was given for 2 days. Aggressive treatment with albumin, fresh plasma, glucose, vitamin K, antibiotics, and sufficient oxygen inhalation were needed after patients with cirrhosis underwent major hepatectomy (Zhou et al. 1994b).

Follow-up

Long-term follow-up after surgery included serum AFP and ultrasonography every 2–3 months and chest radiography every 6 months during the first 2 postoperative years and at 3–6-month intervals thereafter. Computed tomography, magnetic resonance imaging or a hepatic angiogram was done when intrahepatic recurrence was suspected. Reresection was done in 72 patients, including resection for subclinical recurrence after an initial curative resection for PLC ≥10 cm in 71 patients and for solitary pulmonary metastasis in one patient.

"Cytoreduction and sequential resection" was arbitrarily defined as surgically verified, unresectable large PLC ≥10 cm; this was treated with HAL and/or HAC with or without radiotherapy as the first step; resection was done as the second step after marked tumor regression had occurred (Tang et al. 1995). In this series, 57 patients received sequential resection.

Operative mortality indicates death within 30 days of the operation. A microcomputer was used for the storage, analysis, and statistical treatment of clinical data. Survival rates, excluding 30-day postoperative deaths, were calculated according to the life-table method. Statistical differences were tested by the log-rank method. P<0.05 was considered statistically significant.

Results

Of the 621 patients who underwent hepatectomy for PLC ≥10 cm, 28 patients died within 30 days of the resection, with a mortality rate of 4.5% (Table 2). The causes of death were hepatic failure in 18 patients (64.3%), gastrointestinal hemorrhage in five patients (17.9%), sepsis in three patients (10.7%), heart failure in one patient (3.6%), and intracranial disease in one patient (3.6%).

Comparison between patients in group A (PLC ≥10 cm, n=1,227) and group B (PLC <10 cm, n=2,349)

Compared with patients who had PLC <10 cm, patients with PLC ≥10 cm had a younger age (P<0.01), a lower incidence of asymptomatic tumors (P<0.001), higher AFP levels ( >400 ng/ml, 78.3% vs 49.2%, P<0.001), higher γ-GTP levels (>6U, 87.7% vs 70.5%, P<0.001), and a lower incidence of a history of hepatitis (45.0% vs 61.4%, P<0.001) and associated macronodular cirrhosis (59.8% vs 66.6%, P<0.001). The sex ratio and serum HBsAg data were comparable in the two groups (Table 1).

Compared with patients who had PLC <10 cm, patients with PLC ≥10 cm had a poor differentiation of tumor cells (Edmondson grade 3–4; 24.3% vs 19.7%, P<0.01), a lower incidence of single-nodule tumors (59.9% vs 75.4%, P<0.001), a lower proportion of well-encapsulated tumors (28.5% vs 62.1%, P<0.001), and a higher percentage of tumor emboli in the portal vein (20.5% vs 9.0%, P<0.001) (Table 1).

Compared with patients who had PLC <10 cm, patients with PLC ≥10 cm had a lower resection rate (50.6% vs 86.8%, P<0.001), a lower curative resection rate (54.8% vs 78.3%, P<0.001), and a higher operative mortality rate (4.5% vs 2.3%, P<0.001). A local resection was performed less frequently for patients with PLC ≥10 cm (52.5% vs 82.2%, P<0.001) (Table 2).

The 1-year, 3-year, 5-year, and 10-year survival rates were 59.7%, 30.8%, 21.6%, and 15.1%, respectively, for patients with PLC≥10 cm (n=1227), and 82.8, 61.8%, 51.3%, and 36.3%, respectively, for patients with PLC <10 cm (n=2349) (P<0.01).

Of the 621 patients with PLC ≥10 cm who underwent resection, the 1-year, 3-year, 5-year, and 10-year survival rates were 68.0%, 37.3%, 26.2%, and 17.5%, respectively. These rates were 85.0%, 65.1%, 54.3%, and 39.5%, respectively, for patients with PLC<10 cm who underwent resection (n=2,039) (P<0.01).

Factors influencing survival after resection for PLC ≥10 cm

Of the 621 patients who underwent resection for PLC≥10 cm, a significant difference was observed in survival in relation to curability, the number of tumors, tumor capsule formation, and tumor emboli in the portal vein. Patients who underwent curative resection, with a single tumor nodule, a good tumor capsule, without tumor emboli in the portal vein, had a fair prognosis (all P<0.01) (Table 3).

Table 3.

Factors influencing survival after resection for PLC ≥10 cm (n=621)

	No.	Survival rates (%)				P
		1 year	3 year	5 year	10 year
Resection type
Curative	340	78.1	47.1	31.0	21.4	<0.01
Palliative	281	57.4	20.3	20.3	11.7
Tumor number
Single	379	69.9	38.7	25.4	16.6	<0.01
More than one	242	65.1	36.2	18.9	9.2
Tumor capsule
Good	213	80.3	56.5	37.8	24.3	<0.01
Poor	408	52.6	22.1	15.2	13.1
Tumor emboli in portal vein
Absence	370	74.7	45.6	29.4	21.0	<0.01
Presence	251	50.5	23.9	16.9	11.3

Reresection for recurrence after resection for PLC ≥10 cm was performed in 71 patients, the 1-year, 3-year, 5-year, and 10-year survival rates after reresection of recurrence being 84.6, 57.6%, 47.1%, and 41.3%, respectively. One patient underwent reresection for a solitary lung metastasis 3 years and 1 month following the first resection, and has survived for 24 years and 8 months (alive, free of disease).

In this series, 57 patients with unresectable PLC ≥10 cm received cytoreduction therapy and sequential resection. The median tumor size was 12.5 cm in diameter at the first operation and 6 cm in diameter at the second operation. The median interval between the first and second operation was 6 months. The 1-year, 3-year, 5-year, and 10-year survival rates after sequential resection being 84.9%, 70.4%, 55.4%, and 50.1%, respectively.

Long-term survival after surgery for PLC ≥10 cm

Of the 1,227 patients who underwent surgery for PLC ≥10 cm, by the end of December 2000, 55 patients had survived for more than 5 years, and 17 patients survived for more than 10 years. Of the 17 patients who survived more than 10 years, 11 patients (64.7%) underwent resection of tumor. The longest-survival patient had a tumor measuring 17×13×9 cm in size on the left lobe and underwent left hemihepatectomy in August 1971. Pathologic study confirmed HCC (well differentiated) with micronodular cirrhosis. Upon follow-up after 29 years and 4 months, the patient was found to be still living and well. No adjuvant therapy was given postoperatively. One female patient underwent right hemihepatectomy with removal of a tumor measuring 15×15×12 cm in March 1973. The patient has survived for 27 years and 9 months after the operation and was married in 1986 and has had a baby.

Four patients underwent sequential resection after marked shrinkage of tumors. Initially, these patients had large tumors in the right lobe of the liver (one patient had bilateral tumors) and could not tolerate a right hemihepatectomy owing to coexisting cirrhosis. For the first-step operation, HAL plus HAC was performed, with postoperative perfusion of chemoembolization agents in the lesion site. Several months later, all four patients showed marked regression of tumor to around 50% of the original diameter, and sequential resection was carried out. Thus, these patients, formerly considered to have unresectable HCC, finally had resectable HCC. The longest-survival patient underwent reresection for recurrence 5 years and 11 months following sequential resection. The patient has survived for 20 years and 6 months after the first operation, free of disease. Survival of another three patients was 11 years and 11 months (dead with disease), 14 years and 2 months (alive with disease), and 14 years and 2 months (alive, free of disease), respectively.

In this series, two patients with unresectable HCC larger than 10 cm in diameter underwent HAL and HAC, and survived over 10 years. The longest-survival patient was admitted to our hospital in November 1982 because of minor discomfort in the right quadrant of the abdomen for 1 month. The patient had a history of hepatitis for 20 years. His AFP was 5,000 ng/ml, and HBsAg was positive. Radionuclide scan and ultrosonography showed a huge occupying lesion on the right lobe of the liver. An operation in November 1982 revealed a 10×8×8 cm tumor in the right lobe. Aspiration biopsy confirmed HCC with moderate differentiation of cancer cells. Because of coexisting cirrhosis with left-lobe atrophy, the patient could not tolerate a right hemihepatectomy. Right hepatic artery ligation plus cannulation was performed, and postoperative perfusion of chemotherapeutic agents, with 5-fluorouracil, thiotepa, had been carried out for 7 weeks. AFP declined to normal 3 months after the operation. During the ensuing years, the patient received long-term conservative treatment consisting of chemotherapy and Chinese medicinal herbs. Recently, ultrosonography showed that the tumor was in complete remission. The patient has been well for 18 years and 1 month, with negative AFP.

Discussion

Resection of a large PLC is a surgical challenge. Although the operative mortality was higher than those associated with resection for smaller PLC, the death rate of 4.5% within the postoperative 30 days for patients with PLC ≥10 cm in this study was acceptable and comparable to other recent series of liver resections reported in the literature, the operative mortality being around 6% (Franco and Usatoff 2001). Our results demonstrated that in selected patients surgical resection was well tolerated for large PLC because these patients had a low incidence of cirrhosis compared with those with smaller PLC. Furthermore, when a tumor-bearing lobe is occupied mainly by tumor tissue, deterioration of liver function after resection is mild (Lee et al. 1998).

At present, the surgical principles of hepatic resection are well established. For control of bleeding, temporary occlusion of the portal pedicle (the Pringle maneuver) was frequently used in this series. Excellent exposure of the liver can be achieved using a bilateral subcostal incision and a self-retaining retractor. The right liver can be mobilized and rotated to the wound surface if necessary. This procedure is easy when the liver is soft and normal and the tumor is relatively small. However, when a large tumor adheres tightly or invades the diaphragm or adjacent structures, mobilization is difficult because the space available for manipulation is limited. For these patients the anterior approach might be used, which involves initial completion of the parenchyma transection before the right hepatic lobe is mobilized. However, the anterior approach is potentially dangerous and should not be undertaken lightly by beginners because bleeding at the deeper plane of parenchyma transection can be difficult to control if the posterior surface of the liver has not been previously mobilized (Liu et al. 2000).

Some previously reported clinicopathological features of large PLC (Furula et al. 1992; Noguchi et al. 1997; Lee et al. 1998) observed in this study included: 1) increased frequency in young patients; 2) lower incidence of asymptomatic tumors; 3) higher AFP level; 4) higher γ-GTP level; 5) lower percentage of patients with a history of hepatitis and associated liver cirrhosis; 6) higher proportion of Edmondson grade 3 and 4 disease; 7) lower incidence of single-nodule tumor; 8) lower percentage of well-encapsulated tumor; 9) higher proportion of tumor emboli in the portal vein; 10) lower resection rate; 11) lower curative resection rate; 12) higher operative mortality; and 13) less frequency of local resection. These clinicopathological features reflect the relatively advanced stage of the disease in patients with PLC ≥10 cm. In this series, the incidence of patients with serum AFP levels exceeding 400 ng/ml was 78.3% in group A and 62.4% in group B (P<0.001), and γ-GTP levels exceeding 6 U was 87.7% in group A and 70.5% in group B (P<0.001), consistent with our previous observation that AFP and γ-GTP levels correlated with tumor size, and with disease stage to a certain extent (Zhou et al. 1991b; 1995). Recently, some authors observed that huge PLC is usually accompanied by mild cirrhosis or non-cirrhosis, grows expansively outward, and rarely invades the vascular system to cause cancer embolus, as compared with small PLC (Yang et al. 2002; Chen et al. 2002).

The less favorable survival rates after resection of large PLC also reflect the relative advanced stage of the disease. In this series, the 5-year survival rate after resection for PLC was 26.2% in group A and 54.3% in group B (P<0.01). However, there is a subgroup of patients with potentially less malignant large PLC, whose prognosis is not necessarily dismal. Large PLC ≥10 cm does not exclude the possibility of cure. In this series, 17 patients have survived for more than 10 years after surgery. The longest-survival patient has been alive for 29 years and 4 months following left hemihepatectomy with removal of tumor measuring 17×13×9 cm and is still free of disease.

It has been demonstrated that reresection for subclinical recurrence and solitary lung metastasis remains an important approach to prolonging survival after curative resection of PLC (Tang et al. 1984; Zhou et al. 1993b). In this series, 71 patients with PLC ≥10 cm underwent reresection for disease recurrence, the 5-year survival rate being 47.1% after reresection of recurrence. One patient underwent lobectomy for solitary lung metastasis and has survived for 24 years and 8 months, free of disease.

In patients with a large PLC for which it is technically difficult to do a curative resection, the best way might be to decrease the tumor size as the first step and resect the tumor as the second step. Effective approaches for successful cytoreduction might include hepatic artery ligation and infusion chemotherapy, transhepatic arterial chemoembolization (TACE) with or without radiotherapy, and radioimmunotherapy. However, if the huge tumor is well-encapsulated, without daughter nodules and cancer embolus in the vascular system, and cirrhosis is not severe, resection is still the first choice of treatment.

Regional chemotherapy via the distal hepatic artery is based on the concept that a high concentration of drug could be achieved selectively to the tumor-bearing area with decreased systemic toxicity if the agent is injected into the arterial supply of a tumor. HAL+HAC has some advantages. First, since the periphery of liver tumor remains unaffected after HAL, regional chemotherapy can treat the residual tumor cells with ischemic injuries. Second, after HAL the arterial circulation of the liver is slowed down, and the concentration of the drug achieved in hepatic circulation is more than that of intraarterial administration. Our previous results demonstrated that HAL+HAC yielded better survival rates than a single procedure (Zhou et al. 1991b). In this series, two patients with unresectable HCC underwent HAL and HAC, and survived 10 years and 3 months (dead with disease) and 18 years and 1 month (alive, free of disease), respectively. These results indicate that HAL+HAC+combination treatment might provide a hope for prolongation of survival or even resection in some patients with localized initial unresectable PLC.

In conclusion, large PLC more than 10 cm in diameter is not infrequently encountered in clinical practice. Large PLC has specific clinicopathological features. Surgery is the first choice of treatment. In selected patients, resection is safe and offers the chance of long-term survival. Large PLC does not exclude the possibility of cure.