The incidence of hepatocellular carcinoma varies worldwide and correlates with the frequency of chronic hepatitis B virus (HBV) and hepatitis C virus (HCV) infection [3]. The highest risk for hepatocellular carcinoma among cirrhotic patients, 30 cases per 100 000 persons per year, has been reported in some areas of the Far East and sub-Saharan Africa. In Europe and North America, the risk is only 2 cases per 100 000 persons per year [4]. In the United States in the 1980s, 6000 new cases were expected annually [5].

Awareness of the potential for cirrhosis-associated hepatocellular carcinoma has led to the establishment of screening programs, using ultrasonography and the measurement of -fetoprotein levels, in high-risk populations in eastern countries and in transplantation candidates in western countries [6, 7]. Such screening programs have been associated with early detection of small tumors, lower tumor stage at diagnosis, and less severe liver disease. These effects help explain the improved prognosis seen in recent years.

With asymptomatic lesions identified in increasing numbers of patients and with the development of new medical and surgical techniques, approaches to the treatment of cirrhotic patients with hepatocellular carcinoma have changed. Low surgical mortality rates are reported in cirrhotic patients undergoing resection; however, because of tumor extent, tumor location, and hepatic dysfunction, only a minority of patients are candidates for resection. In addition, tumor recurrence due to multifocality and progression of underlying liver disease has limited the efficacy of resection [8]. As orthotopic liver transplantation became available, it seemed to be a more rational approach because it addressed the multifocal nature of the tumor as well as the underlying liver disease.

In early experiences with unselected patient populations, however, liver transplantation for hepatocellular carcinoma was associated with high rates of death from recurrent disease [9-11]. Given our limited organ resources, these results called into question the justification of transplantation as a treatment for cancer. More recently, improved transplantation outcomes have been shown in selected patients with hepatocellular carcinoma.

In addition to surgical methods, transcatheter arterial chemoembolization and percutaneous ethanol injection have emerged as techniques for local disease control in patients with inoperable hepatocellular carcinoma [12, 13]. Preliminary experience with transcatheter arterial chemoembolization before transplantation has shown promising results in patients with advanced-stage tumors [14, 15].

In this review, we discuss the treatment options for cirrhotic patients with hepatocellular carcinoma and suggest an updated approach to therapy based on the severity of liver disease and tumor stage.

Methods

Top Methods Author & Article Info References

We conducted a MEDLINE search of English-language articles published between 1981 and 1997. Original articles reporting the results of resection, liver transplantation, and invasive radiologic approaches (chemoembolization and ethanol injection) for the treatment of hepatocellular carcinoma in cirrhotic patients were selected. Priority was given to large series and to comparative studies of the various therapies. We also report the initial results of a multimethod approach combining liver transplantation with chemoembolization at the Mount Sinai Medical Center and of similar approaches at other large transplantation centers. The study design of each original report was assessed, and careful attention was paid to methods and aims.

Relevant data on patient population, tumor stage distribution, treatment, survival, follow-up interval, and rate of recurrent disease were extracted and analyzed. The material was classified into treatment categories and discussed accordingly. Survival rates given in the text are cumulative survival rates as calculated by life-Table estimates or the Kaplan-Meier method, unless they are noted to be actual survival rates (or the proportion of patients surviving at a specific time interval).

Cirrhosis and Carcinogenesis

Carcinogenesis associated with cirrhosis correlates with disease duration and the cause of primary liver disease [16, 17]. Ikeda and colleagues [16] noted that the cumulative rate of hepatocellular carcinoma at 15 years was 75.2% in patients with hepatitis C and only 27.2% in patients with hepatitis B [16]. Indeed, in Japan, where the incidences of chronic HBV infection and chronic HCV infection are the same, the reported incidence of hepatocellular carcinoma is higher in patients with chronic HCV infection than in patients with chronic HBV infection (10.4% compared with 3.9%) [18]. Although the association between viral infection and tumor development is a worldwide phenomenon, the risk for tumor development varies according to geographic location. The reported annual incidence of new cases of hepatocellular carcinoma among patients with HCV-related cirrhosis is 3% in western countries and 6% in Japan [19]; this suggests that nonviral etiologic factors are also involved in the pathogenesis of this tumor.

Certain metabolic disorders that lead to cirrhosis, including hemochromatosis, hereditary thyrosinemia, ₁-antitrypsin deficiency, and type I glycogen storage disease, are also associated with an increased risk for hepatocellular carcinoma [20]. Aflatoxin B₁ contamination has been identified as a major risk factor that probably acts synergistically with HBV infection [20]. Other chemical carcinogens, inorganic arsenic, androgenic steroids, and oral contraceptive steroids may also increase the risk for hepatocellular carcinoma.

Some differences exist in the activity and modes of cancer promotion of HBV and HCV infection. In patients with HBV infection and hepatocellular carcinoma, HBV integrates into host genomic DNA in most cases, both in tumor cells and in nontransformed hepatocytes [21]. This suggests that viral integration at one or a few sites in host DNA may result in the deregulated expression of oncogenes or tumor-suppressor genes [22]. Integration of HBV also promotes genomic instability [23], which is a common feature of carcinogenesis.

The mechanism by which HCV infection promotes the development of hepatocellular carcinoma is unclear. Hepatitis C virus was found to infect and replicate in hepatocellular carcinoma cells in vivo [24]; integration of HCV viral sequences into the host genome was not seen. Other possible mechanisms of hepatocarcinogenesis after infection with HCV include the direct induction of tumorigenesis through a viral gene product, the activation of cellular proto-oncogenes, or the inactivation of cellular tumor suppressors. Recent studies [25, 26] have suggested that the third nonstructural gene of HCV (NS3) is involved in cell transformation. The NS3 protein, a serine protease with a helicase activity, can induce cell death and oncogenic transformation [26]. Another difference between HBV and HCV is that hepatocellular carcinoma is always found in the setting of chronic liver disease in patients with hepatitis C but may develop in asymptomatic carriers of hepatitis B surface antigen (HBsAg) [27]. Of note, patients infected with both HCV and HBV have an increased relative risk for hepatocellular carcinoma; this suggests that the two viruses have a synergistic effect in hepatocarcinogenesis [28]. In a prospective study of 400 cirrhotic patients [29], the annual incidence of hepatocellular carcinoma was 6.6% in patients with HBsAg alone, 7.0% in patients with anti-HCV alone, and 13.3% in patients with both HBV and HCV infection.

Histologic sections of resected specimens have shown a pathway from the regenerative nodules typically found in cirrhotic livers to adenomatous hyperplasia and hepatocellular carcinoma [30, 31]. A report of multiple foci of hepatocellular carcinoma in a macroregenerative nodule further suggests that these are precancerous lesions [32]. Although the mechanism by which cirrhosis potentiates malignant transformation is not well established, it is important to recognize this relation when designing treatment approaches for cirrhotic patients with hepatocellular carcinoma.

Resection for Hepatocellular Carcinoma

Resection remains the mainstay of treatment for hepatocellular carcinoma. Resectability of a tumor in cirrhotic patients, however, is limited by the diminished functional reserve of the cirrhotic liver and the attendant risk for intraoperative bleeding and postoperative liver failure [33]. The reported resectability rate is about 10% in western countries but 28% in eastern countries, which have earlier detection as a result of screening programs and more widespread expertise in techniques for limited resection [34].

Attempts have been made to quantify hepatic reserve and thus to estimate the extent of resection that may safely be performed. Kawasaki and coworkers [35] used volume of ascites, bilirubin level, and indocyanine green clearance in a formula to estimate liver reserve. In this series, perioperative mortality among 112 patients undergoing liver resection was only 1.8%. Tumor size is an important determinant of resectability: In one series [36], the resectability rate was only 41% in patients with large tumors (>5 cm in diameter) compared with 89% in patients with small tumors. Measurement of the hepatic-vein wedge pressure gradient was recently reported to be valuable in patients with Child A cirrhosis who were undergoing resection [37]. The presence of significant portal hypertension (hepatic-vein wedge pressure gradient >10 mm Hg) was the best predictor of postoperative hepatic decompensation associated with increased mortality.

Despite the low perioperative mortality rate seen when careful selection criteria are implemented, long-term survival after resection is limited by recurrent tumor and progression of liver disease, with 5-year survival rates in most series decreasing to between 17% and 40% (Table 1). In a report by Nagasue and colleagues [46] on 229 patients who had "curative" resection for hepatocellular carcinoma, 5- and 10-year survival rates were 26.4% and 19.4%, respectively.

Tumor size is the most important predictor of recurrence after resection. Nagao and associates [51] reported 2-year survival rates of 80% in patients with tumors less than 5 cm in diameter and 40% in patients with larger tumors. Other factors that predict decreased survival include multiplicity, vascular invasion, low histologic grade, and absence of capsule [51, 52]. Survival of patients who have small solitary lesions without vascular invasion (postsurgical tumor, node, metastasis [pTNM] stage I) depends mainly on the underlying liver disease [8, 9, 53]. In a report by Takenaka and colleagues [49] on 280 liver resections for hepatocellular carcinoma, 5-year disease-free survival rates in patients with stage I, stage II, and stage III tumors were 38%, 34%, and 17%, respectively.

The multifocal nature of hepatocellular carcinoma in cirrhotic patients explains the high rate of intrahepatic recurrence after resection. Belghiti and coworkers [38] studied intrahepatic recurrence after resection of solitary hepatocellular carcinoma in 47 patients who had undergone both preoperative computed tomography after intra-arterial lipiodol injection and intraoperative ultrasonography. Within 3 years, 60% of patients had intrahepatic recurrence; disease was adjacent to the resection margin in only 9% of cases. In another study [48], 76% of deaths long after resection of a small hepatocellular carcinoma (<2 cm in diameter) were attributed to cancer recurrence. Because most recurrent tumors arise during the first 2 years after resection, this finding might be explained by the multicentric nature of hepatomas in cirrhotic livers rather than by intrahepatic metastasis. The multifocal nature of hepatocellular carcinoma is also demonstrated by the examination of livers of patients who undergo liver transplantation. At the Mount Sinai Medical Center, Schwartz and associates [54] noted multifocality in 38% of transplant recipients with hepatocellular carcinomas less than 5 cm in diameter and 82% of those with tumors greater than 5 cm in diameter [54]. Bhattacharya and colleagues [55] examined the explants of 50 liver recipients who had undergone pretransplantation injection of lipiodol through the hepatic artery. In 5 patients with hepatocellular carcinoma, the explanted liver contained 7 to 15 foci.

It has been suggested that the likelihood of recurrence after resection varies with the cause of liver disease. A higher incidence of recurrent hepatocellular carcinoma in patients with hepatitis C than in patients with hepatitis B was reported [56]; this finding was attributed to enhanced inflammatory changes in the liver in patients with hepatitis C.

Nonsurgical Methods for Local Disease Control

Percutaneous ethanol injection and transcatheter arterial chemotherapy with or without embolization have evolved as techniques for local control of hepatocellular carcinoma. Percutaneous ethanol injection is useful only for lesions less than 4 cm in diameter [57].

Percutaneous ethanol injection was first described in 1986 by Livraghi and coworkers [13]. Under ultrasonographic guidance, 1 to 4 mL of a 95% ethanol solution is injected percutaneously through a 22-gauge needle into the tumor. Six to twelve injection sessions are done at weekly intervals according to pretreatment tumor size and response (as assessed by changes in appearance on ultrasonography or dynamic computed tomography). Livraghi and coworkers [58] reported that in 23 patients with unresectable hepatocellular carcinoma, all tumors were smaller after percutaneous ethanol injection. In 8 of 12 patients with hepatocellular carcinomas less than 2 cm in diameter, the tumor was not detectable on ultrasonography.

Ebara and associates [59] reported on their experience with percutaneous ethanol injection in 95 patients with unresectable hepatocellular carcinomas less than 3 cm in diameter. Although all index lesions shrank and 42% were undetectable on ultrasonography, 48% of patients had new tumors on follow-up. The 5-year actuarial survival rate in this series was 28%, which was better than the rate in an untreated control group. At 3 years after treatment, 66% of patients had tumor recurrence, often at sites in the liver distant from the original tumor. Castells and coworkers [60] compared the efficacy of percutaneous ethanol injection with that of resection for hepatocellular carcinomas less than 4 cm in diameter. Although the 30 patients treated with percutaneous ethanol injection had worse liver function, their survival did not differ from that of the 33 patients who had resection. Percutaneous ethanol injection thus seems to be advantageous for small hepatocellular carcinomas in patients with poor hepatic reserve in whom surgical resection is risky.

Acetic acid and not saline injections have also been successfully used to ablate small hepatocellular carcinomas. The potential benefit of these agents is that they are less irritating than alcohol if injected unintentionally into blood vessels or if they leak into the peritoneal cavity [61].

The results of systemic chemotherapy with adriamycin and 5-fluorouracil for inoperable hepatocellular carcinoma have been disappointing, with response rates of less than 20% [61, 62]. A treatment that seems to be more effective is targeted regional chemotherapy done by using transcatheter arterial chemoembolization, in which chemotherapy with or without lipiodol is selectively administered to the tumor. This is followed by embolization of the tumor with gelatin sponge particles [63]. Lipiodol is selectively taken up by the highly vascularized hepatocellular carcinoma; the cytotoxic drug it carries is then slowly delivered to the tumor in high concentrations. Embolization promotes retention of the cytotoxic drug within the tumor and induces tumor necrosis. Chemoembolization is contraindicated under the following circumstances: 1) when the main portal vein is invaded by tumor, 2) in the presence of intrahepatic arteriovenous shunting, and 3) in patients with severely impaired liver function. Fever, abdominal pain, and ileus associated with elevated liver enzyme levels often follow arterial embolization. This response is usually self-limited, although progression to liver failure may occur, especially in patients with decompensated end-stage liver disease or patients who undergo serial chemoembolization sessions [63].

Yamada and colleagues [12] used chemoembolization with mitomycin C and adriamycin in 120 patients with unresectable hepatocellular carcinoma. They found reduced tumor size in 75% of cases and marked decreases in -fetoprotein levels in 90%. Bismuth and coworkers [64] evaluated the efficacy of chemoembolization with doxorubicin in 291 cirrhotic patients with hepatocellular carcinoma; 29% of these patients had a decrease in tumor size. Morbidity and mortality after treatment correlated with severity of liver disease. Two-year survival rates were 49% for patients with Child class A cirrhosis, 29% for patients with Child class B cirrhosis, and 9% for patients with Child class C cirrhosis. Complications included cholecystitis (10%), vasculitis (14%), renal failure (13%), and jaundice (12%). Histologic sections of resected specimens in 43 patients who had resection or transplantation after chemoembolization showed complete necrosis of tumor in 26% of cases and more than 50% necrosis in another 48%. Similar results were reported by others using mitomycin C or doxorubicin [65-67].

In a recent multicenter, prospective study, chemoembolization in patients with unresectable hepatocellular carcinoma reduced tumor growth but showed no survival benefits [63]. The limited efficacy of transcatheter arterial chemoembolization in the treatment of large unresectable tumors is explained by the fact that viable tumor cells may remain after treatment, mainly when vascular invasion, small daughter nodules, and tumor thrombi are present. Transcatheter arterial chemoembolization is primarily effective in patients with hypervascular single tumors [68]. Recently, a few studies [69, 70] have reported higher response rates and improved survival in patients with unresectable hepatocellular carcinoma who receive transcatheter arterial chemoembolization followed by ethanol injection compared with patients who receive chemoembolization alone.

In summary, nonsurgical measures are effective for the temporary control of disease spread in patients with unresectable small hepatocellular carcinoma and in a selected group of patients with advanced tumors. In patients with unresectable large tumors, techniques that achieve complete tumor necrosis through a combination of transcatheter arterial chemoembolization and percutaneous ethanol injection may provide better long-term disease-free survival.

Liver Transplantation for Hepatocellular Carcinoma

Liver transplantation has become widely accepted as therapy for hepatocellular carcinoma when anatomic or functional limitations preclude resection. In Europe, hepatocellular carcinoma associated with end-stage liver disease was, until recently, the indication for orthotopic liver transplantation in 12% of liver recipients (about 200 patients per year) [53].

The initial experience with liver transplantation in unselected patients with hepatocellular carcinoma who did not receive adjuvant chemotherapy was discouraging (Table 2). Penn [77] reported on the worldwide experience with liver transplantation for hepatocellular carcinoma in 365 patients. The 5-year survival rate was 18%, and only 9% of patients survived disease-free for more than 2 years after transplantation. Despite these poor results, this experience seemed to suggest that hepatectomy with transplantation may be advantageous in a selected group of patients with end-stage liver disease.

Several large centers that perform both hepatic resection and liver transplantation have reviewed their experience with these two surgical approaches. Iwatsuki and Starzl [74] reported their results in 105 patients undergoing transplantation for hepatocellular carcinoma and 76 patients undergoing hepatic resection for this tumor; survival was significantly higher after transplantation than after resection at each stage of pTNM classification. Multivariate analysis identified large tumor size, multifocality, bilobar involvement, and microscopic vascular invasion as factors associated with poor disease-free survival after transplantation [74]. Bismuth and coworkers [42] compared results after resection with results after liver transplantation in 120 cirrhotic patients with hepatocellular carcinoma (60 patients in each group). Overall survival at 3 years was similar (50% after resection compared with 47% after transplantation), but the disease-free survival rate was better after transplantation (46% compared with 27%). In this study, transplantation conferred the greatest benefit in patients with one or two small tumors (<3 cm in diameter), achieving a much better disease-free survival rate at 3 years than that seen with resection (83% compared with 18%). When Otto and associates [78] compared outcomes in 53 patients who had resection for hepatocellular carcinoma with outcomes in 50 patients who had liver transplantation for hepatocellular carcinoma, they found significantly better 3-year survival rates after transplantation than after resection in patients with pT1 and pT2 tumors and patients with small, oligocentric tumors less than 5 cm in diameter (1 to 2 or 1 to 5 nodules). In their series, the difference in survival was explained by higher perioperative mortality among patients who had resection.

Several pretransplantation variables have been found to influence long-term, disease-free survival. Of these, tumor size and number of tumor nodules are the most important. Romani and colleagues [79] reported on 27 transplant recipients with small hepatocellular carcinomas (<5 cm in diameter); actuarial survival rates were 82% at 1 year and 71% at 3 years. Similarly, Figueras and coworkers [76] found that 1- and 3-year survival rates among 84 patients with small hepatocellular carcinomas (<5 cm in diameter) were 83% and 77%, respectively. Bismuth and coworkers [80] found that among 109 cirrhotic patients undergoing liver transplantation, the 5-year survival rate was 66% for patients with tumors less than 3 cm in diameter compared with 42% for those with tumors greater than 5 cm in diameter. In this study, 5-year survival rates were 58% in patients with fewer than three lesions and 39% in patients with more than three. Mazzaferro and associates [81] studied 48 patients with small, unresectable hepatocellular carcinomas who had liver transplantation. To be eligible for transplantation, patients could have no more than one tumor less than 5 cm in diameter and no more than three nodules with a diameter of 3 cm or less. In the 35 patients who met the selection criteria, overall and recurrence-free survival rates at 4 years were 85% and 92%, respectively.

These observations show that within certain limits, the survival rate in patients with small hepatocellular carcinomas is similar to that in transplant recipients without a tumor. This experience provided the foundation for the policy, adopted in most transplantation programs, that discovery of a small hepatocellular carcinoma during evaluation for transplantation does not rule out candidacy.

In contrast, however, certain histologic findings in the explant (such as the presence of capsular and microvascular invasion) are signs of a more aggressive tumor and are associated with a greater incidence of recurrence. In a retrospective analysis of 71 liver transplant recipients with hepatocellular carcinoma [82], no patient with encapsulated tumor had recurrence after transplantation, whereas none of the patients who had microvascular invasion were alive at 3 years.

Because most patients with hepatocellular carcinoma have an associated viral disease, the risks for recurrent viral infection must be considered in the evaluation of long-term results after transplantation. In the past, recurrence of hepatitis B after transplantation has been a serious problem. In one study [75], the 3-year post-transplantation survival rate was 18% in patients with hepatitis B and hepatocellular carcinoma compared with 69% in patients with hepatocellular carcinoma who were HBsAg-negative. More recently, passive immunization with hepatitis B immunoglobulin was shown to reduce the risk for viral reinfection. It has been suggested that HBV-associated hepatocellular carcinoma may be more virulent [83]. No data are available to date on the effect of recurrent HCV infection in transplant recipients with hepatocellular carcinoma.

The effect of long-term immunosuppression on tumor growth in patients with hepatocellular carcinoma is unknown. One report suggests that long-term steroid administration in patients with hepatocellular carcinoma who are undergoing transplantation increases the risk for tumor recurrence fourfold compared with a protocol in which steroids are withdrawn by 6 months [84]. Longer follow-up of a large patient sample is necessary to clarify this issue.

In summary, long-term results after liver transplantation in cirrhotic patients with known hepatocellular carcinoma vary according to tumor characteristics. A high recurrence rate after transplantation has been seen in patients with large or more aggressive tumors. A favorable outcome can be expected in cirrhotic patients with a small tumor who are not candidates for resection. Nevertheless, the severe shortage of donor organs and the relatively high early mortality rate after liver transplantation, averaging 13% for the first year [85], restricts the use of transplantation as a first-line treatment in cirrhotic patients with small tumors.

Multimethod Therapies

In attempts to decrease the high recurrence rate and improve long-term survival after resection or transplantation for large hepatocellular carcinomas (>3 cm in diameter), approaches that combine surgical and medical measures have been used. The hypothesis that micrometastases are present in the liver at surgery and that tumor spread may occur during manipulation of the liver has been the basis for the preoperative use of chemoembolization and percutaneous ethanol injection. Nagasue and coworkers [86] used preoperative chemoembolization in cirrhotic patients undergoing liver resection and noted tumor necrosis and a reduction in tumor size in resected specimens; survival and recurrence rates, however, did not improve. Furthermore, intra-abdominal complications resulting from treatment made surgery more difficult. Bismuth and coworkers [64] used chemoembolization followed by resection (in 23 patients) or transplantation (in 20 patients) and achieved histologic response rates with tumor necrosis of more than 50% in almost half of resected specimens.

Chemoembolization is also a rational approach in patients awaiting transplantation, particularly because transplantation candidates with tumors are not given priority on the waiting list. Chemoembolization may inhibit tumor growth. In addition, progression of tumor despite treatment may predict which patients are likely to do poorly and may permit selection of patients who are less likely to have recurrence after transplantation. Venook and associates [14] selected HBsAg-negative patients with fewer than four hepatocellular carcinomas less than 5 cm in diameter. Patients with adequate hepatic reserve had chemoembolization when they were placed on the waiting list for an organ; patients with poor hepatic reserve had transcatheter arterial chemoembolization only when a donor became available. Ten of the 11 patients who underwent transplantation remained tumor-free at 40 months; 1 patient who died at 6 months had no cancer on autopsy. Three other patients developed metastases before transplantation. Although no benefit of pretransplantation chemoembolization has yet been documented, many transplantation centers now use it (Table 3) [81, 82, 90].

Adjuvant chemotherapy after transplantation has also produced good initial results. Olthoff and associates [89] reported on 25 liver recipients, 20 of whom had stage III and IV tumors. Treatment with fluorouracil, doxorubicin, and cisplatin was initiated after surgery (once the patient was stable) and was continued for 6 months. The 3-year survival rate was 46% compared with 5.8% in a historical control. Although most of the side effects of such protocols are self-limited, treatment had to be discontinued in some cases because of doxorubicin-induced cardiomyopathy or azothemia related to cisplatin. Neutropenia developed in almost all patients, but only a few required treatment with granulocyte colony-stimulating factor.

Another approach that seems promising is one that combines pre-, intra-, and postoperative chemotherapy in transplant recipients with large tumors. Stone and associates [88] used systemic chemotherapy with doxorubicin administered before, during, and after transplantation in 20 patients, 17 of whom had large tumors. The actuarial 3-year survival rate was 59%; the tumor-free survival rate was 54%. Carr and coworkers [87] reported on 12 patients with stage III or stage IV hepatocellular carcinoma who received intrahepatic transarterial chemotherapy with doxorubicin and cisplatin combined with interferon- before transplantation. Systemic chemotherapy was continued for 12 months after transplantation. Only 2 of the 11 patients who underwent liver transplantation had recurrent tumor at 7 and 14 months; these 2 patients died 12 and 15 months after transplantation. A third patient who had unrecognized lung metastases at the time of transplantation also died. Of 14 liver recipients with similarly staged disease who did not enter the protocol, 9 had recurrence within the first year after transplantation. Cherqui and colleagues [91] used chemoembolization combined with radiation therapy before transplantation followed by systemic chemotherapy after transplantation in 9 patients; the 1-year survival rate was 64%.

In a recent experience at the Mount Sinai Medical Center [92], 66 patients with hepatocellular carcinomas greater than 5 cm in diameter were enrolled in a multimethod protocol that used chemoembolization with mitomycin C, doxorubicin, and cisplatin before transplantation. Patients with decompensated cirrhosis, renal failure (creatinine concentration > 2 mg/dL [176.8 µmol/L]), and an abnormal resting multiple gated acquisition scan, as well as those with extrahepatic disease in regional lymph nodes or the main portal vein, were excluded. Twenty-nine of the 66 patients who entered the protocol either died or had tumor progression while waiting. The 32 patients who had transplantation were given intraoperative doxorubicin (10 mg/m² body surface area); systemic treatment with the same agent was initiated 2 to 4 weeks after transplantation. Tumor recurrence after transplantation was noted in 12 patients and was the cause of death in 6. Actuarial survival rates were 85% at 1 year and 63% at 4 years. Disease-free survival rates were 72% at 1 year and 35% at 4 years. Although tumors recurred later in these patients than in a historical control, the relatively low disease-free survival rate at 4 years suggests that this adjuvant protocol ultimately did not prevent tumor recurrence. However, by identifying patients with large tumors that respond to adjuvant chemotherapy, this approach may more effectively select patients who will benefit from transplantation. This preliminary experience seems promising, but the effectiveness of multimethod approaches using targeted and systemic adjuvant chemotherapy before or after liver transplantation in patients with cirrhosis and hepatocellular carcinoma has yet to be confirmed in any large, randomized studies.

The use of molecular markers (such as messenger RNA albumin and messenger RNA -fetoprotein) for the early detection of hematogenous tumor spread may assist in the selection of appropriate candidates with large tumors for liver transplantation [93, 94].

Designing a Treatment Strategy

The treatment approach to a cirrhotic patient with hepatocellular carcinoma should be designed to achieve the lowest immediate mortality and the longest possible tumor-free survival. Several factors must be considered, including the patient's general status, the severity of liver disease, and the influence of tumor characteristics (such as size, location, and multifocality) on outcome after resection and transplantation.

In a cirrhotic patient with a tumor less than 5 cm in diameter, treatment strategy should be based on tumor location, hepatic reserve, and presence of portal hypertension [37] (Figure 1). Resection should be reserved for the patient who has a small peripheral lesion and preserved hepatic function (Child class A) and no significant portal hypertension. In patients in whom the histologic findings in the resected specimen are associated with a high probability of recurrence (for example, capsular or microvascular invasion or additional tumor nodules), liver transplantation should be considered because it produces a better long-term disease-free survival rate [42].

View larger version (28K): [in this window] [in a new window]   Figure 1. Approach to management of the cirrhotic patient with a small hepatocellular carcinoma. PEI = percutaneous ethanol injection; TACE = transarterial chemoembolization.

Patients with more advanced cirrhosis (Child class B or C) who have a small tumor or several small lesions (fewer than three nodules, each with a diameter <3 cm) will probably achieve the greatest long-term benefit from liver transplantation. While the patient is awaiting transplantation, ethanol injection and chemoembolization can be used with a strict follow-up protocol and repeated treatment sessions as needed.

Unfortunately, in patients with advanced tumor (stage VI) and minimal hepatic reserve who are unable to undergo resection, transplantation, even with adjuvant therapy, has had poor long-term results. Some of these patients may benefit from chemoembolization and percutaneous ethanol injection, although even these minimally invasive procedures may be hazardous in such patients.

Given that long-term survival rates after liver transplantation for patients with small tumors are similar to those in liver transplant recipients without tumor, and given the relatively low incidence of recurrent tumor, adjuvant systemic chemotherapy after transplantation seems to have no advantage.

Conclusions

In the past, results after surgical management of hepatocellular carcinoma in cirrhotic patients were poor, and surgery was advised only for patients with relatively well-preserved liver function. Long-term survival was limited by a high recurrent tumor rate and progression of liver disease. Today, several advances have allowed better outcomes in cirrhotic patients with hepatocellular carcinoma. Modern screening methods permit earlier detection of tumors. Invasive radiologic procedures permit targeted treatment of tumors, and liver transplantation enables surgical removal of the tumor in patients with advanced cirrhosis.

Small tumors are best managed by resection or transplantation; the choice is based on the patient's general condition, hepatic reserve, and extent of tumor. Despite evidence for long-term survival after transplantation in some patients with large tumors, most of these patients die of recurrent disease. The value of multimethod approaches that combine chemoembolization with liver transplantation to prevent tumor recurrence in patients with large tumors is still uncertain, and further prospective evaluation of such protocols in a large cohort is required. Identification of prognostic factors in this patient population may facilitate selection of suitable candidates for transplantation.

Finally, in addition to developing better methods for managing cirrhotic patients with hepatocellular carcinoma, we should study the prevention and arrest of viral infection and genetic therapies for their potential value in preventing and treating hepatocellular carcinoma.

Dr. Kaspa: The Liver Institute, Rabin Medical Center, Beilinson Campus, Petah-Tikva 49100, Israel.

Drs. Sheiner and Schwartz: Recanati/Miller Transplantation Institute, Mount Sinai Medical Center, Box 1104, One Gustave L. Levy Place, New York, NY 10029.