Long-Term Survival and Pattern of Recurrence After Resection of Small Hepatocellular Carcinoma in Patients With Preserved Liver Function

Abstract

Objective

To evaluate the survival results and pattern of recurrence after resection of potentially transplantable small hepatocellular carcinomas (HCC) in patients with preserved liver function, with special reference to the implications for a strategy of salvage transplantation.

Summary Background Data

Primary resection followed by transplantation for recurrence or deterioration of liver function has been recently suggested as a rational strategy for patients with HCC 5 cm or smaller and preserved liver function. However, there are no published data on transplantability after HCC recurrence or long-term deterioration of liver function after resection of small HCC in Child-Pugh class A patients. Such data are critical in determining the feasibility of salvage transplantation.

Methods

From a prospective database of 473 patients with resection of HCC between 1989 and 1999, 135 patients age 65 years or younger had Child-Pugh class A chronic liver disease (chronic hepatitis or cirrhosis) and transplantable small HCC (solitary ≤5 cm or two or three tumors ≤3 cm). Survival results were analyzed and the pattern of recurrence was examined for eligibility for salvage transplantation based on the same criteria as those of primary transplantation for HCC.

Results

Overall survival rates at 1, 3, 5, and 10 years were 90%, 76%, 70%, and 35%, respectively, and the corresponding disease-free survival rates were 74%, 50%, 36%, and 22%. Cirrhosis and oligonodular tumors were predictive of worse disease-free survival. Patients with concomitant oligonodular tumors and cirrhosis had a 5-year overall survival rate of 48% and a disease-free survival rate of 0%, which were significantly worse compared with other subgroups. At a median follow-up of 48 months, 67 patients had recurrence and 53 (79%) of them were considered eligible for salvage transplantation. Decompensation from Child-Pugh class A to B or C without recurrence occurred in only six patients.

Conclusions

For Child-Pugh class A patients with small HCC, hepatic resection is a reasonable first-line treatment associated with a favorable 5-year overall survival rate. A considerable proportion of patients may survive without recurrence for 5 or even 10 years; among those with recurrence, the majority may be eligible for salvage transplantation. These data suggest that primary resection and salvage transplantation may be a feasible and rational strategy for patients with small HCC and preserved liver function. Primary transplantation may be a preferable option for the subset of patients with oligonodular tumors in cirrhotic liver in view of the poor survival results after resection.

Hepatocellular carcinoma (HCC) is a leading cause of cancer death in Asia, and its incidence is also on the rise in Western countries. ^1,2 Hepatic resection has been the mainstay of surgical treatment for HCC. It is now well established that liver transplantation is the treatment of choice for unresectable small HCC (≤5 cm) in patients with decompensated Child-Pugh class B or C cirrhosis. ^3–5 The best treatment strategy for small HCC in patients with preserved liver function is controversial. Studies have shown superior survival results after transplantation compared with resection, especially in terms of disease-free survival rates. ^6–10 On that basis, some authors have recommended transplantation as the treatment of choice even for resectable small HCC in Child-Pugh class A patients. ^7–10 However, other studies have shown similar overall survival rates after resection and transplantation for small HCC, and hence concluded that resection rather than transplantation should be the first-line treatment for small HCC in patients with preserved liver function. ^11–13 In one study, the disease-free survival rates after resection and transplantation for HCC 5 cm or smaller were found to be comparable. ¹¹

Whether a patient with an HCC 5 cm or smaller and preserved liver function should be treated by resection or transplantation is not only a debatable issue, ¹⁴ but also a therapeutic dilemma that many hepatic surgeons are facing with increasing frequency because of the growing number of small HCCs detected by screening of patients with chronic hepatitis or cirrhosis. ¹⁵ The donor organ shortage has been a major deterrent to the use of liver transplantation for small resectable HCC. ^13,16,17 In a recent intent-to-treat analysis comparing resection and transplantation for early HCC, a high dropout rate from tumor progression as a result of a long waiting time was identified as a major factor hampering survival in patients with early HCC listed for transplantation. ¹⁸

Recently an alternative strategy of primary resection followed by transplantation for recurrence or deterioration of liver function (salvage transplantation) has been proposed for patients with small HCC and preserved liver function. 13,16,19,20 Although no study has formally evaluated the efficacy of such a strategy, it is already being used in some transplantation units under the pressure of organ shortage. ¹⁹ Such a strategy depends on the premise that the majority of recurrences after resection of small HCC are still transplantable. However, so far there have been no published data on transplantability after HCC recurrence or long-term deterioration of liver function after resection of small HCC in Child-Pugh class A patients. Such data are fundamental in determining whether a strategy of primary resection and salvage transplantation is practicable. Hence, we conducted a study of the long-term survival outcome and pattern of recurrence after resection of potentially transplantable small HCC in a cohort of Child-Pugh class A patients. Based on the results, we examined the implications on the strategy of salvage transplantation.

PATIENTS AND METHODS

Between January 1989 and December 1999, 473 consecutive patients underwent hepatic resection for HCC in the Department of Surgery, University of Hong Kong Medical Center, Queen Mary Hospital, Hong Kong, China. Clinicopathologic data including liver function status, tumor size, tumor number, any radiologic evidence of macroscopic venous invasion, and histopathologic findings of resected tumors in these patients were prospectively collected in a computerized database. From this database, we identified 135 patients age 65 years or younger with Child-Pugh class A chronic liver disease and transplantable small HCC according to the following widely accepted criteria: solitary tumor 5 cm or smaller or two or three tumor nodules (oligonodular) each 3 cm or less; absence of extrahepatic metastasis; and absence of radiologic evidence of macroscopic portal vein or hepatic vein invasion. 3–5,10,16,18,19 These 135 patients constituted the subjects of this study. Six other patients with solitary small HCC 5 cm or smaller who presented with tumor rupture were excluded. The clinicopathologic characteristics and type of resection of the 135 patients are shown in Table 1. Among these patients, 87% had HCC related to hepatitis B virus (HBV) infection, which can lead to chronic hepatitis or cirrhosis, depending on the degree of chronic liver damage. Ninety patients had histologically confirmed cirrhosis; the other 45 patients had evidence of chronic hepatitis. In our center, hepatic resection was considered the treatment of choice for anatomically resectable HCC in Child-Pugh class A patients. All patients were managed by the same team of hepatic surgeons. The details of our preoperative assessment, surgical techniques, and perioperative care have been given in a previous report. ²² These patients did not receive any neoadjuvant or adjuvant therapy, except for 12 patients who had postoperative adjuvant transarterial chemotherapy as part of a previously conducted randomized controlled trial. ²³

Table 1. CLINICOPATHOLOGIC AND SURGICAL DATA

HBsAg, hepatitis B surface antigen; ICG-R15, indocyanine green retention at 15 minutes; AFP, alpha-fetoprotein. PTNM, pathologic tumor-node-metastasis. ²¹

* Figure indicates size of the largest tumor for patients with two or three tumor nodules.

All the patients were followed up by the same surgical team in an HCC clinic with a standard protocol of surveillance that included a contrast-enhanced computed tomography (CT) scan at 1 month after resection, followed by liver function test, serum alpha-fetoprotein (AFP) level, ultrasonography or CT scan and chest radiograph every 3 months. The median follow-up period was 48 months (range 14–135; interquartile range 30–80). Diagnosis of recurrence was based on typical imaging findings in a contrast-enhanced CT scan and an increased serum AFP level. In cases of uncertainty, hepatic arteriography and a post-Lipiodol CT scan were performed, and if necessary, fine-needle aspiration cytology was used for confirmation. All recurrent diseases were prospectively recorded in the database immediately on confirmation of diagnosis, and details including the site, number, and size of recurrent tumors were documented. The strict follow-up for recurrence and prospective collection of data on recurrence in our unit afforded the opportunity for a reliable analysis of transplantability after HCC recurrence or long-term deterioration of liver function. As suggested in a previous study, the recurrences were classified as transplantable and nontransplantable using the same criteria as those of primary transplantation for HCC (i.e., solitary tumor ≤5 cm or two or three tumor nodules each ≤3 cm; absence of extrahepatic metastasis or macroscopic venous invasion). ¹⁹

Patients with recurrence were managed with various therapeutic modalities, including re-resection for resectable recurrence, transarterial chemoembolization (TACE) or percutaneous ethanol injection for unresectable intrahepatic recurrence, and systemic chemotherapy using intravenous epirubicin for selected patients with extrahepatic metastatic disease. ²⁴ The treatment decision was based on the pattern of recurrence and liver function reserve. Since 2000, we have offered transplantation for selected patients with transplantable recurrence.

Statistical Analysis

Data were retrieved from the prospective database for analysis. Statistical comparison between groups was performed using the chi-square test with Yates correction (or the Fisher exact test where appropriate) for nominal data and the Mann-Whitney test for numerical data. Cumulative overall survival and disease-free survival rates were computed according to the Kaplan-Meier method and compared between groups by the log-rank test. The Cox proportional hazards model was used for multivariate analysis. Overall survival analysis included all hospital deaths, defined as deaths within the same hospital admission after surgery. Hospital deaths were excluded from the disease-free survival analysis. All analyses were performed using statistical software (SPSS, Inc., Chicago, IL). P < .05 was considered statistically significant.

RESULTS

Survival Results

Of the 135 patients, 5 (3.7%) died in the hospital after surgery, 2 died of cardiac complications, and 3 died of intraabdominal sepsis. The hospital death rate was similar for the 45 patients with chronic hepatitis (4.4%) and the 90 patients with cirrhosis (3.3%, P = .542). The hospital death rate for 83 patients operated on after 1995 was 1.2% (1/83) compared with 7.7% (4/52) of those operated on before 1995 (P = .050).

Taking hospital deaths into account, the overall 1-, 3-, 5-, and 10-year survival rates of the 135 patients were 90%, 76%, 70%, and 35%, respectively (Fig. 1). Excluding the five hospital deaths, the 1-, 3-, 5-, and 10-year disease-free survival rates of 130 patients were 74%, 50%, 36%, and 22%, respectively (Fig. 2). At the time of analysis, 54 of the 130 patients were alive without recurrence and 36 were alive with recurrent disease. Forty patients had died, including 31 of tumor recurrence, six of cirrhotic complications without recurrence (variceal bleeding, n = 3; liver failure, n = 2; spontaneous bacterial peritonitis, n = 1), and three of unrelated medical conditions.

Figure 1. Cumulative overall survival curve of 135 Child-Pugh class A patients after resection of small hepatocellular carcinoma (≤5 cm).

Figure 2. Cumulative disease-free survival curve of 130 Child-Pugh class A patients after resection of small hepatocellular carcinoma (≤5 cm; 5 patients with hospital death were excluded from analysis).

To examine whether there were certain clinical subsets of patients who had a less favorable outcome after resection, the patients were stratified in turn by the status of nontumorous liver, tumor number (solitary or oligonodular), and tumor size (≤3 cm or 3.1–5 cm). There was a trend toward better overall survival rates in patients with chronic hepatitis (n = 45; 1-, 3-, 5-, and 10-year rates 91%, 83%, 78%, and 39%, respectively) compared with patients with cirrhosis (n = 90; 1-, 3-, 5-, and 10-year rates 90%, 72%, 66%, and 27%, respectively), but the difference was not significant (P = .078). The disease-free survival rates of patients with cirrhosis (1-, 3-, 5-, and 10-year rates 70%, 41%, 28%, and 21%, respectively) were significantly lower than those of patients with chronic hepatitis (n = 43; 1-, 3-, 5-, and 10-year rates 81%, 67%, 54%, and 28%, respectively) (P = .022;Fig. 3).

Figure 3. Cumulative disease-free survival curves of patients with chronic hepatitis (n = 43) and those with cirrhosis (n = 87) (P = .022).

Similarly, there was a trend toward better overall survival rates in patients with solitary tumors (n = 115; 1-, 3-, 5-, and 10-year rates 91%, 76%, 72%, and 40%, respectively) compared with those with oligonodular tumors (n = 20; 1-, 3-, 5-, and 10-year rates 86%, 71%, 60%, and 0%, respectively) (P = .082). The disease-free survival rates of patients with solitary tumors (1-, 3-, 5-, and 10-year rates of 76%, 52%, 40%, and 24%, respectively) were significantly better than those with oligonodular tumors (1-, 3-, 5-, and 10-year rates 57%, 33%, 0%, and 0%, respectively) (P = .029, Fig. 4).

Figure 4. Cumulative disease-free survival curves of patients with solitary tumor (n = 110) and those with oligonodular (two or three) tumors (n = 20) (P = .029).

In contrast, there were no significant differences in the overall survival rates between 70 patients with tumors 3 cm or smaller (1-, 3-, 5-, and 10-year rates 91%, 78%, 74%, and 37%, respectively) and 65 patients with tumors 3.1 to 5 cm (1-, 3-, 5-, and 10-year rates 89%, 73%, 65%, and 33%, respectively, P = .473), nor were there significant differences in the disease-free survival rates (1-, 3-, 5-, and 10-year rates 78%, 55%, 38%, and 21% vs. 71%, 44%, 34%, and 34%, respectively, P = .318).

Table 2 summarizes the survival results when patients were categorized by both cirrhotic status and number of tumors. Patients with concomitant oligonodular HCC and cirrhosis had significantly worse overall and disease-free survival rates than the other subgroups. The cumulative 5-year disease-free survival of patients with oligonodular tumors in cirrhotic liver was 0%.

Table 2. OVERALL AND DISEASE-FREE SURVIVAL RATES ACCORDING TO CIRRHOTIC STATUS AND NUMBER OF TUMORS

* Overall survival of group 4 significantly worse than group 1 or 2; no significant differences between other groups.

† Disease-free survival of group 4 significantly worse than group 1, 2, or 3, and disease-free survival of group 3 significantly worse than group 1 or 2; no significant difference between groups 1 and 2.

Apart from cirrhosis and oligonodular tumors, microscopic venous invasion and advanced pTNM stage were also adverse prognostic factors for disease-free survival. The 1-, 3-, 5-, and 10-year disease-free survival rates were 53%, 30%, 12%, and 0%, respectively, compared with 78%, 55%, 41%, and 25%, respectively, among patients with (n = 26) or without (n = 104) venous invasion (P < .001). The corresponding disease-free survival rates were 58%, 28%, 14%, and 0%, respectively, in patients with stage IIIA/IVA disease (n = 36), compared with 80%, 51%, 45%, and 27%, respectively, in those with stage I/II disease (n = 94) (P < .001). The other clinical or pathologic factors listed in Table 1 (sex, age, hepatitis B status, preoperative indocyanine green retention at 15 minutes, serum AFP, and any microsatellite lesions) had no significant influence on the disease-free survival rate. When all the clinicopathologic variables were entered into a multivariate analysis, cirrhosis (risk ratio [RR] 1.542, 95% confidence interval [CI] 1.023–2.325, P = .046), oligonodular tumors (RR 2.012, 95% CI 1.200–4.395, P = .021), and presence of microscopic venous invasion (RR 1.824, 95% CI 1.046–3.178, P = .034) were independent predictors of unfavorable disease-free survival.

The only pathologic factor significantly predictive of overall survival was pTNM stage. The 1-, 3-, 5-, and 10-year overall survival rates were 84%, 64%, 58%, and 0%, respectively, in 38 patients with stage IIIA/IVA disease, compared with 93%, 80%, 75%, and 44%, respectively, in 97 patients with stage I/II disease (P = .012).

Pattern of Recurrence

At a median follow-up of 48 months, 67 (51%) of 130 patients had recurrence. The cumulative recurrence rates at 1, 3, 5, and 10 years were 24%, 48%, 60%, and 76%, respectively. Figure 5 shows the pattern of recurrence according to site and number of recurrences. Thirty-nine patients had solitary, 14 had oligonodular, and six had multinodular (four or more) intrahepatic recurrences. At the time of diagnosis, the size of the intrahepatic recurrences ranged from 0.5 to 4.0 cm (mean 2.2 ± 1.2), and none were associated with macroscopic venous invasion. Five patients had extrahepatic recurrence alone (diaphragmatic recurrence, n = 1; lung metastasis, n = 2; bone metastasis, n = 2), and another three had concurrent intrahepatic and extrahepatic recurrence (defined as diagnosis of extrahepatic recurrence simultaneously or within 3 months of detection of intrahepatic recurrence). The median time between resection and tumor recurrence among the 67 patients was 16 months (range 1–84).

Figure 5. Pattern of recurrence according to the site (intrahepatic or extrahepatic) and number of recurrent tumors. Oligonodular means two or three tumor nodules; multinodular means four or more tumor nodules. IHR, intrahepatic recurrence; EHR, extrahepatic recurrence.

Using the same criteria as for primary transplantation for HCC, 53 patients with solitary or oligonodular tumors confined to the liver remnant were eligible for salvage transplantation, constituting 79% of all patients with recurrence. The median time to recurrence of these 53 patients with transplantable recurrences was 20 months (range 3–84). Twenty of these recurrences were observed in year 1 after resection, 10 in year 2, 8 in year 3, 7 in year 4, 4 in year 5, and 4 after 5 years. The other 14 patients with extrahepatic or multinodular intrahepatic recurrences were classified as nontransplantable. The median time to recurrence in these 14 patients was 6 months (range 1–55), significantly shorter than that of transplantable recurrences (P = .024). Six (11%) of the 53 patients with transplantable intrahepatic recurrence subsequently had extrahepatic metastasis 12 to 36 months (median 20) after the first diagnosis of intrahepatic recurrence, whereas in the other 47 patients there was no evidence of extrahepatic metastasis for a period of 6 to 97 months (median 18.5) from the time of diagnosis of recurrence until the date of last follow-up or death.

Among the 67 patients with recurrence, the transplantability of recurrence was analyzed in relation to the nontumorous liver status, number of primary tumors, presence of venous invasion in the primary tumor, and pTNM stage of the primary tumor, which were the factors predictive of recurrence (Table 3). There was no significant difference in the transplantability of recurrence between patients with cirrhosis and chronic hepatitis, nor was there a significant difference between patients with solitary and oligonodular primary tumors. In contrast, venous invasion and advanced pTNM stage (IIIA or IVA) in the primary tumors were associated with a lower transplantability rate of recurrence. Among 14 nontransplantable patients, 10 had primary tumors of stage IIIA (n = 8, all with solitary tumor >2 cm with microscopic venous invasion) or stage IVA (n = 2, both with three tumor nodules located in both lobes).

Table 3. TRANSPLANTABILITY OF RECURRENCE

Main treatment modalities for the 53 patients with transplantable recurrence included re-resection of intrahepatic recurrence (n = 7), TACE (n = 40), ethanol injection (n = 4), and salvage transplantation (n = 2). The first patient with salvage transplantation for recurrence was a 51-year-old man with HBV-related cirrhosis in whom a 2-cm recurrent tumor developed in the right lobe 10 months after resection of a 3-cm HCC in the left lobe. The tumor was first controlled with TACE, and the patient subsequently underwent living-related liver transplantation with right lobe graft from his wife. The second patient was a 39-year-old man with HBV-related chronic hepatitis who was found to have a 4-cm recurrent tumor in the liver remnant 2.5 years after resection of a 4.5-cm HCC. He also underwent living-related liver transplantation with right lobe graft from his wife after initial tumor control with TACE. Both pairs of recipients and donors recovered without complications from surgery. At the time of analysis, the two patients were disease-free for 5 and 4 months, respectively. Treatments for the 14 patients with nontransplantable recurrence included TACE for multinodular intrahepatic recurrence (n = 6), excision of extrahepatic recurrence (diaphragmatic recurrence, n = 1; lung metastasis, n = 1), systemic chemotherapy (n = 3), and symptomatic palliation alone (n = 3).

The overall 1-, 3-, 5-, and 10-year survival rates from the time of recurrence of all 67 patients were 77%, 49%, 26%, and 16%, respectively. The 1-, 3-, 5-, and 10-year survival rates after recurrence of the 53 patients with transplantable recurrence were 85%, 53%, 30%, and 22%, respectively, compared with 43%, 32%, 16%, and 0%, respectively, in the 14 patients with nontransplantable recurrence (P = .033).

Long-Term Deterioration of Liver Function

Of the 130 patients who survived the initial resection, 6 patients (4.6%) had deterioration of liver function from Child-Pugh class A to class B or C at 3 to 37 months (median 10.5) after surgery. All of these six patients had cirrhotic liver (6.9% of all patients with cirrhosis), and all underwent segmentectomy. The preoperative indocyanine green retention at 15 minutes of these six patients (median 12.5%, range 10.6–21.5%) was not significantly different from that of the other patients with cirrhosis who remained in Child-Pugh class A (median 10.3%, P = .092). All six patients died of cirrhotic complications 2 to 5 months after decompensation. These six patients could have been considered candidates for salvage transplantation.

DISCUSSION

This study focused on the long-term results after resection of small HCC in a cohort of patients with chronic liver disease but preserved liver function who otherwise could have been candidates for transplantation, in an attempt to provide survival and recurrence data suitable for comparison with those after primary transplantation. Although previous studies have reported survival and recurrence rates after resection of small HCCs less than 3 to 5 cm, ^25–27 there is a paucity of specific data on the results of resection in patients with preserved liver function and transplantable HCC. Our results show that resection of small HCC in Child-Pugh class A patients can be performed with a low death rate, and the 5-year overall survival rate of 70% is comparable to that of 60% to 74% reported in recent series of transplantation for HCCs 5 cm or smaller. 16,18,20,28,29 Even among patients with cirrhosis, the 5-year survival rate was 66%. Therefore, in terms of overall survival result, hepatic resection can be justified as the first-line treatment for small HCC in patients with preserved liver function. However, the 5-year disease-free survival rate in this cohort of patients was only 36%, whereas the corresponding rate after transplantation for small HCC was about 60%. ^7,8,16 The lower tumor recurrence rate has been the main argument for advocating transplantation for small HCC even in Child-Pugh class A patients.

Provided that a strategy of salvage transplantation is feasible for recurrent disease, it can be argued that resection may be the primary treatment of choice for patients with small HCC and preserved liver function. Although tumor recurrence is much less a problem after transplantation, there are other complications specific to transplantation that can compromise long-term survival, such as graft rejection, opportunistic infections, and development of other malignancies as a result of immunosuppression. In Western countries where the majority of HCCs are related to hepatitis C infection, hepatitis C recurrence is a major long-term complication. These long-term complications explain an overall survival outcome similar to that after hepatic resection, despite a lower tumor recurrence rate after transplantation. ¹³ The long-term immunosuppression may also lead to chronic complications and compromised quality of life. ^30,31 Further, transplantation is a costly procedure and requires a graft. The worldwide graft shortage is the most important restriction for transplantation. In a series of 87 patients with early HCC listed for transplantation in a Spanish center, 8 patients died of tumor progression before a graft was available. ¹⁸ In another analysis, the survival benefit of primary transplantation over resection begins to disappear once the waiting time for a graft exceeds 6 to 10 months. ¹⁰ A strategy of resection and salvage transplantation saves scarce organs only for those with recurrence or deterioration of liver function. Our data show that 36% and 22% of the patients may be expected to be disease-free at 5 and 10 years, respectively. Hence, transplantation may not be necessary in a considerable proportion of patients.

The strategy of salvage transplantation should work best for patients with a low risk of recurrence after resection. ¹⁹ Hence, we performed subgroup analyses to identify any subset of patients with a high risk of recurrence after resection, who may be candidates for primary transplantation from the outset. Cirrhosis and oligonodular tumors were found to adversely affect the disease-free survival rate. Both factors may be associated with a high risk of multicentric recurrence. ³² The survival results of patients with concomitant oligocentric tumors and cirrhosis were particularly poor compared with other subgroups, with a 5-year disease-free survival rate of 0%. It makes intuitive sense to suggest that primary transplantation may be a more appropriate option for this subset of patients. Microscopic venous invasion was also predictive of poor disease-free survival, presumably because of a higher risk of intrahepatic or extrahepatic metastasis. Further, if recurrence does occur in patients with microscopic venous invasion, a higher proportion of recurrences will not be transplantable. It is tempting to suggest that patients with microscopic venous invasion may benefit from primary transplantation, but this is a histologic finding that cannot be used as a criterion for preoperative selection. Further, some studies have shown that microscopic venous invasion also predicted a higher risk of recurrence even after transplantation for HCC, ^28,33 although others have not found such a prognostic influence. ^7,20 It is not surprising that pTNM stage IIIA or IVA disease portended a poor 5-year disease-free survival rate, because the classification incorporates both venous invasion and number of tumors. In contrast, transplantation for small HCC of stage IIIA or IVA has been shown to result in disease-free survival rates similar to those of earlier-stage tumors. ^3,20 Thus, primary transplantation may be the preferred option for patients with HCC 5 cm or smaller but of stage IIIA or IVA. Unfortunately, accurate preoperative staging is impossible because of the lack of information regarding microscopic vascular invasion, which frequently results in understaging. ^3,20

This study shows that 79% of the recurrences after resection of small HCC were considered transplantable using the same criteria of primary transplantation for HCC, as adopted by some authors. ^18,19 However, recurrence in the liver remnant may be due to intrahepatic metastasis rather than metachronous de novo tumor in some patients. Theoretically, the former may be associated with less favorable results after transplantation because there is a possibility of concurrent hematogenous metastasis. Currently it is impossible to differentiate between the two except by DNA clonal analysis of the resected tumors. ³⁴ In our study, most of the intrahepatic recurrences were confined to the liver for a long period; even in six patients in whom extrahepatic metastasis subsequently developed, the distant metastasis was detected more than 1 year after diagnosis of the intrahepatic recurrence. These findings may suggest that most intrahepatic recurrences after resection of small HCC were metachronous de novo tumors, as conjectured in a previous study. ²⁷ However, it is also possible that intrahepatic metastases may be confined to the liver remnant without concurrent distant metastasis in many instances, because they are believed to arise from tumor cell spread via the portal venous circulation rather than the systemic circulation. ³² Our previous experience has shown that re-resection of intrahepatic recurrence can result in a survival rate comparable to that after primary resection for HCC. ²⁴ It is not unreasonable to envisage that salvage transplantation for recurrence in appropriately selected patients may result in survival comparable to that after primary transplantation. ¹⁹ Most intrahepatic recurrences are not amenable to re-resection because of unfavorable tumor location, presence of more than one lesion, or inadequate liver function reserve after previous hepatic resection, but salvage transplantation may be suitable for such patients. A recent report from European centers has provided encouraging results, with recurrent disease found in only 1 of 12 patients transplanted for postresection recurrent HCC. ³⁵

Apart from tumor recurrence, deterioration of liver function is another potential indication for salvage transplantation. In our cohort of patients, only six patients had hepatic decompensation to Child-Pugh class B or C without recurrence. This remarkably low incidence of long-term hepatic decompensation may be attributable to the fact that most of our patients had chronic liver disease caused by HBV infection. It has been shown that Child-Pugh class A patients with chronic hepatitis or cirrhosis related to HBV infection have a slow course of progression and a low incidence of hepatic decompensation after years of follow-up. ³⁶

For salvage transplantation to be a viable strategy, donor organs must be available when recurrence or deterioration of liver function occurs. In Western countries, where cadaveric grafts are more available, salvage transplantation may be practicable if the growth of intrahepatic recurrence can be controlled while waiting for a graft. TACE appears to be an effective bridging therapy; a good response of recurrent tumors can be expected because of their small size. ³⁷ In Eastern countries, where the cadaveric donor rate is extremely low, living-related liver transplantation is considered a justified option for the treatment of unresectable small HCC. ³⁸ With the experience accumulated since our first adult-to-adult living-donor liver transplantation in 1996, we have shown that this procedure is safe for both the recipients and the donors, provided that patients are carefully selected. ^39,40 In places such as Hong Kong and Japan, where living donor is the main source of graft for liver transplantation, primary resection backed up by salvage transplantation is a particularly appealing strategy for patients with small HCC and preserved liver function. When there is an option of resection with a low surgical risk and a comparable overall survival result, it seems unethical to incur the risk of donor hepatectomy on a living donor for a Child-Pugh class A patient with small HCC. Even if this is considered an option, understandably it will be difficult for the relatives to accept the need for living-related transplantation knowing that there is a simpler and safer choice. However, in the face of recurrence or deterioration of liver function after resection, the option of living-related transplantation becomes more justified, provided that the recipient has a reasonable chance of survival. Our initial experience with the two patients transplanted for recurrent HCC using right lobe grafts from living donors was encouraging, but long-term follow-up results in a larger number of patients are required to establish the role of living-related transplantation for recurrent HCC. The combined strategy of resection and transplantation may improve the overall survival results of surgical treatment compared with what has been achieved with either treatment alone. In this study, the 5-year survival rate after recurrence was 30% for the 53 patients with transplantable recurrences, most being treated with TACE. Assuming that the 5-year survival rate after transplantation for recurrence was 60%, with an allowance for potentially worse survival in patients with intrahepatic metastasis, ¹⁹ the overall 5-year survival rate could have been improved had transplantation been available as a salvage treatment for recurrence. Compared with a policy of primary transplantation for all patients, avoidance of unnecessary transplantation in some patients who may survive disease-free for long periods after resection can reduce the death rate from complications after transplantation.

A recent biometric analysis comparing salvage transplantation with primary transplantation based on hypothetical assumptions found increased life expectancy and graft saving with the former strategy in a scenario assuming an average 5% to 10% yearly hepatic decompensation rate, a 15% yearly recurrence rate, and an 80% transplantability rate after recurrence. ¹⁹ The recurrence and transplantability rates of recurrence in our study approximated those assumed in the above analysis, and the rate of hepatic decompensation was lower. Our data suggest that salvage transplantation is a plausible approach. However, there are certain caveats. First, most patients in this study had HBV-related HCC. Some Japanese studies have shown worse liver function and an increased incidence of multicentric tumors after resection of hepatitis C-related HCC compared with HBV-related HCC. ^41,42 Data on the recurrence rate, transplantability of recurrence, and deterioration of liver function after resection of small HCC in patients with hepatitis C are needed to clarify whether salvage transplantation is a rational strategy for such patients. The second caveat is that the favorable results in this cohort of patients were obtained in a center with extensive experience in resection of HCC. Finally, it cannot be overemphasized that the feasibility of salvage transplantation depends on strict surveillance after resection for early detection of intrahepatic recurrence.

In conclusion, this study shows that resection can be performed safely with an overall 5-year survival rate of 70% in Child-Pugh class A patients with small HCC. A considerable proportion of patients may survive disease-free for 5 or even 10 years; among those with recurrences, most appear to be transplantable. Our data suggest that primary resection followed by salvage transplantation for recurrence or deterioration of liver function may be a feasible and rational approach for such patients. The role of liver transplantation for HCC is continuously evolving. At present, the worldwide experience with salvage transplantation for recurrent HCC is limited, but data on the long-term results after transplantation for recurrent HCC will probably be available in the coming years to allow a better assessment of the strategy. Ultimately, randomized trials conducted on an intent-to-treat basis will be needed to clarify the role of this strategy compared with primary transplantation or resection alone for small HCC in patients with preserved liver function.