Resection Prior to Liver Transplantation for Hepatocellular Carcinoma

Abstract

Objective:

To evaluate the feasibility and postoperative course of liver transplantation (LT) in cirrhotic patients who underwent liver resection prior to LT for HCC.

Summary Background Data:

Although LT provides longer survival than liver resection for treatment of small HCCs, donor shortage and long LT wait time may argue against LT. The feasibility and survival following LT after hepatic resection have not been previously examined.

Methods:

Between 1991 and 2001, among 107 patients who underwent LT for HCC, 88 met Mazzafero’s criteria upon pathologic analysis of the explant. Of these, 70 underwent primary liver transplantation (PLT) and 18 liver resection prior to secondary liver transplantation (SLT) for recurrence (n = 11), deterioration of liver function (n = 4), or high risk for recurrence (n = 3). Perioperative and postoperative factors and long-term survival were compared.

Results:

Comparison of PLT and SLT groups at the time of LT revealed similar median age (53 vs. 55 years), sex, and etiology of liver disease (alcohol/viral B/C/other). In the SLT group, the mean time between liver resection and listing for LT was 20 months (range 1–84 months). Overall time on LT waiting list of the two groups was similar (3 vs. 5 months). Pathologic analysis after LT revealed similar tumor size (2.2 vs. 2.3 cm) and number (1.6 vs. 1.7). Perioperative and postoperative courses were not different in terms of operative time (551 vs. 530 minutes), blood loss (1191 vs. 1282 mL), transfusion (3 vs. 2 units), ICU (9 vs. 10 days) or hospital stay (32 vs. 31 days), morbidity (51% vs. 56%) or 30-day mortality (5.7% vs. 5.6%). During a median follow-up of 32 months (3 to 158 months), 3 patients recurred after PLT and one after SLT. After transplantation, 3- and 5-year overall survivals were not different between groups (82 vs. 82% and 59 vs. 61%).

Conclusions:

In selected patients, liver resection prior to transplantation does not increase the morbidity or impair long-term survival following LT. Therefore, liver resection prior to transplantation can be integrated in the treatment strategy for HCC.

Transplantation following hepatic resection for hepatocellular carcinoma (HCC) is feasible. Resection may function as a bridge treatment, allows selection of transplant candidates based on pathologic analysis of the entire specimen, and preserves the possibility of salvage transplantation. Resection in selected patients could reduce the number of patients on the transplant waiting list.

Liver transplantation (LT) is the most effective treatment of patients with small hepatocellular carcinomas (HCC).^1–4 The international transplantation community has largely adopted an approach to LT for HCC based upon the criteria proposed by Bismuth et al,⁵ which were amended and popularized by Mazzafero et al⁶ and adopted by UNOS.⁷ LT based on these Milano criteria (solitary liver nodule not exceeding 5 cm in maximum diameter, or 2 or 3 tumors not exceeding 3 cm in diameter) has been shown to provide very good disease-free survival, so that it is considered to be the optimal treatment of small HCC, especially in patients with underlying chronic liver disease.1–4,6,8,9 The evidence that LT should be therefore the preferred treatment choice for these patients has increased the demand resulting in longer waiting lists in the face of a relative shortage of available donors.¹⁰ During long waiting times, some patients suffer progression of disease such that they can never benefit from LT. Bridge treatments to halt or delay tumor progression during the waiting period for LT include radiofrequency tumor ablation, transarterial chemoembolization, and liver resection.^11–16 Poon et al proposed to resect HCC in selected patients eligible for LT and to reserve LT for those who develop recurrence or deterioration of liver function.¹¹ This approach, which proposes resection as a bridge treatment to prevent tumor progression during the waiting period, has not been studied in detail. Further, resection might enable selection of patients with HCCs who are likely to obtain the maximum benefit from LT. However, transplant surgeons are concerned that prior liver resection could complicate the operative transplant procedure, increase the risk of postoperative complications, and even impair the survival advantage of transplantation over resection alone.

To evaluate the impact of liver resection prior to LT, we retrospectively studied a homogeneous group of patients with cirrhosis and HCC who strictly fulfilled Milano criteria based on pathologic analysis of the explanted specimen after LT. The aim of this study was to report the perioperative course, morbidity, and survival in the group of patients who underwent secondary liver transplantation (SLT) after liver resection compared with those who underwent primary liver transplantation (PLT) for HCC.

PATIENTS AND METHODS

Patients Studied and Indications for Liver Transplantation

From 1990 to 2001, among 548 patients with HCC treated surgically at our institution, 107 (20%) underwent elective LT. All of these transplanted patients were selected for LT based on the following pretransplant criteria: age younger than 60 years, solitary liver nodule not exceeding 5 cm in maximum diameter, or 2 or 3 tumors not exceeding 3 cm in diameter on preoperative spiral computed tomography (CT) or MRI. Patients with emergency transplantation, evidence of extrahepatic spread of tumor, or who were listed and then not transplanted are not included in this group. Patients with HCC were not given priority for LT over those on the waiting list with cirrhosis without HCC.

Nineteen of the 107 LT patients (18%) were excluded from analysis for one or more of the following reasons: retransplantation, simultaneous liver and renal transplantation, split LT, absence of underlying liver disease on the specimen, and/or patients in whom tumors did not meet the Milano tumor size or number criteria on pathologic analysis of the specimen. Thus, 88 patients all with cirrhosis who underwent whole-liver cadaveric LT for HCC who strictly fulfilled pathologic tumor criteria on the explanted specimen were analyzed. The clinicopathological and surgical data for all studied patients at the time of LT are presented in Table 1. Patient ages ranged from 30 to 60 years. The etiology of cirrhosis was alcohol or hepatitis C infection in 83% of patients. Tumor sizes ranged from 0.8 to 3.5 cm, and the range of tumor nodules was 1 to 3 tumors per patient without exceptions. Seventy-three percent of patients had either one or more antitumor treatments before LT, including transarterial chemoembolization, percutaneous alcohol injection, radiofrequency, or liver resection. Among 88 patients, 18 underwent resection for their first occurrence of HCC prior to LT and represented the group of SLT. The clinicopathological and surgical data at the time of resection for the SLT group are presented in Table 2. All resected patients had cirrhosis, Child-Pugh score A. Fifty-five percent had prior nonsurgical tumor therapy before liver resection (including transarterial chemoembolization and percutaneous therapies). Patients who underwent resection were listed for LT after a mean time of 20 months (median 11 months; range 1–84 months). Eleven of 18 (61%) underwent LT for morphologic evidence of recurrence. The remaining 7 did not have objective evidence of recurrence but were transplanted either for deterioration of liver function (4 patients, 22%) or “de principe” (deliberately), without waiting for evidence of recurrence, due to positive margin or satellite nodules on the specimen (3 patients, 17%). Of the latter 7 patients, 4 (57%) were found incidentally to have HCC in the explanted liver.

TABLE 1. Descriptive Characteristics of All 88 Studied Patients at the Time of Liver Transplantation

TABLE 2. Descriptive Characteristics of All 18 Secondary Liver Transplantation Patients at the Time of Resection

In the SLT group, tumor sizes ranged from 1 to 4 cm, and the maximum number of nodules found on specimen explant was 3. Transthoracic approach was used in 27% (due to tumor location in the dome of the liver near diaphragm).¹⁷ Twenty-eight percent underwent major resection defined by removal of 2 or more Couinaud segments (right hepatectomy in 2 patients, 2 or more segment-oriented resections in 3 patients). Minor resections including local resections (transabdominal or transthoracic tumorectomy) or single segment-oriented resections were performed in 72% of the patients.

Clinicopathological characteristics, perioperative surgical morbidity, and survival were compared between the 70 PLT and 18 SLT patients at the time of LT.

Morbidity and Complications

Minor complications, such as asymptomatic pleural effusions which did not require treatment, are not discussed. Major complications are defined as follows: major infection (hemoculture positivity, pulmonary infection requiring antibiotic therapy, symptomatic urinary tract infection requiring antibiotic therapy, central venous catheter infection requiring catheter removal, exchange and/or antibiotic therapy, intra-abdominal abscess requiring reoperation or percutaneous drainage), renal insufficiency (serum creatinine level ≥ 120 μmol/L), acute rejection, need for reoperation, need for retransplantation. Global morbidity represents the proportion of patients with one or more of the above complications. Postoperative death was defined as death in the first 30 postoperative days. Orthotopic LT with preservation of inferior vena caval flow utilizing a temporary portocaval shunt and the modified piggyback technique was performed in all cases as previously described.^18,19

Principle Criteria for Analysis

Clinicopathological data were analyzed in both groups, including preoperative liver function, etiology of underlying liver disease, the delay between resection and listing for the LT, and wait time between listing and LT. Pathologic analysis of explanted specimens was available in all patients. Perioperative morbidity, including perioperative blood loss, transfusions, and operative duration, was examined. Short-term postoperative outcomes, including complications, intensive care unit stay, hospital stay, and frequency of reoperation were compared. Disease-free and overall survival were determined. Statistical analyses using standard tests (χ², t test) were performed where appropriate. Cumulative survival rates were computed according to the Kaplan-Meier method and compared between groups by the log-rank test. Significance was defined as a P value <0.05. Statistical analyses were performed using StatView for Windows.²⁰

Follow-up

Follow-up after liver resection and after LT was conducted at 3-month intervals during the first 2 postoperative years and every 6 months thereafter. At each follow-up visit, physical, serologic (liver function tests and serum alpha-fetoprotein level), and radiologic examinations (chest radiography, liver US, and CT and/or MRI scan if indicated by preceding examinations) were obtained. After resection, patients were registered for the LT if they developed recurrence or deterioration in liver function (except for 3 patients who were transplanted de principe).

RESULTS

Clinicobiological Characteristics of Patients at the Time of LT

The characteristics of the two groups at time of LT, including preoperative clinical factors, biologic factors, and pathologic tumor factors are compared in Table 3.

TABLE 3. Comparison of Primary and Secondary Liver Transplantation Groups at the Time of Transplantation

There were no significant differences between patients who underwent PLT versus SLT with regard to clinical factors, measures of liver function, etiology of cirrhosis or median waiting time to performance of the LT (PLT, 3 months vs. SLT 5 months). For patients with morphologic evidence of HCC recurrence after resection, the mean delay between resection and morphologic recurrence was 20 months (median 9 months; range 3–68 months). For those who underwent SLT for either deterioration of liver function or de principe, the delay between resection and listing was 14 months (median 7.5 months; range 1–15 months). Of the 7 SLT patients without morphologic evidence of tumor recurrence at the time of LT, 4 (57%) were found incidentally to harbor HCC in the resected specimen after LT. For patients with tumor in the explanted liver (70 of 70 in the PLT group, 15 of 18 in the SLT group), there was no difference in the mean number or size of tumors between groups (Table 3).

Perioperative and Postoperative Course and Complications From LT

There were no significant differences between PLT and SLT groups in terms of perioperative factors including operative time, operative blood loss, or transfusion of blood products (Table 4). Further, the overall incidence of complications was similar in both groups (51% for PLT, 56% for SLT). Although there were more reoperations in the SLT group (39% vs. 13% for PLT), this did not translate into difference in retransplantation or prolonged duration of intensive care unit or hospital stay for the SLT group (Table 4). Hematoma or bleeding which led to reoperation was more common after SLT than PLT (22% vs. 3%). There was no statistical difference in the incidence or reoperation for biliary fistula or arterial thrombosis. Incidence of other complications such as sepsis, renal insufficiency, and acute rejection did not differ between groups. Although statistical subgroup analysis was not possible because of the number of studied patients in the SLT group, the type of resection appeared to influence the perioperative morbidity of the transplantation procedure (Fig. 1). Following major resection, LT appeared to be more difficult, reflected in a longer operating time, increased perioperative transfusion rate, and longer hospital stay. In contrast, following minor resections, LT appeared to be less difficult; after transthoracic resection, only a few adhesions to the diaphragm were encountered which had very little impact on the transplant hepatectomy.

TABLE 4. Comparison of Perioperative and Postoperative Characteristics of Primary Versus Secondary Liver Transplantation Groups

FIGURE 1. Analysis of perioperative and postoperative factors in secondary liver transplantation group (n = 18) based on surgical approach to resection prior to liver transplantation. A trend toward greater operative difficult and longer hospital course emerged in patients who underwent major (n = 5) as opposed to minor resection (n = 13) prior to SLT. Conversely, those who underwent prior transthoracic tumor resection had the lowest perioperative and postoperative morbidity measured by need for blood transfusion, duration of operation, and length of hospital stay.

Follow-up and Survival After LT

In the SLT group, the mean time from resection to listing for LT was 20 months (median 11 months, range 1–84 months). The overall survival rates excluding postoperative deaths after LT in PLT and SLT groups were not different (Fig. 2). There were 4 postoperative deaths (5.7%) in the PLT group (idiopathic peritonitis and septic shock on postoperative day 5; acute rejection, graft arterial thrombosis, and renal failure on postoperative day 11; aspergillosis on postoperative day 20; graft failure followed by multiple organ dysfunction on postoperative day 30) and a single death (5.6%) in the first 30 days postoperatively in the SLT group (hepatorenal failure on postoperative day 4). Excluding all postoperative deaths, the 3- and 5-year disease-free survivals for SLT (82% and 61%) were similar to those for PLT (82% and 59%). Of the 7 patients who underwent transplantation without morphologic evidence of HCC, 4 were found to have HCC in the resected specimen. With a median follow-up for the whole group of 32 months (range 3–135 months), three recurrences were recorded following PLT and one following SLT (not significant).

FIGURE 2. Kaplan-Meier survival plots for PLT versus SLT groups. The survival rates measured from the time of liver transplantation in the group of patients who underwent primary (—) versus secondary (—) liver transplantation for HCC. There was a single death (5.6%) in the first 30 days postoperatively in the SLT group and 4 postoperative deaths (5.7%) in the PLT group. Patients who died in the postoperative period were excluded.

DISCUSSION

Transplantation, which is the treatment of choice for small HCC in cirrhotic patients, is considerably limited by organ shortage. As a result of prolonged wait times before transplantation, tumor progression may counteract the benefit of LT.¹¹ Surgical resection of the tumor is an optional bridge treatment, which has been anecdotally proposed in some series.^21–23 Poon et al noted that nearly 80% of patients who recur after primary hepatic resection for HCC remain eligible for LT.¹¹ Thus, a new strategy has been proposed for patients with preserved liver function and HCC: hepatic resection prior to “salvage” or secondary transplantation. This approach may not only reduce the impact of a long waiting list and donor shortage but might function as a strategy to select patients with HCCs who are likely to obtain the maximum benefit from LT. The major drawbacks of this concept of primary liver resection prior to transplantation could be the increased technical difficulty during transplantation procedure and the risk for impaired posttransplant survival. Results of this study showed that in selected patients with cirrhosis and resectable HCC, liver resection prior to transplantation is feasible and does not contraindicate LT or impair the long-term survival from transplant.

Liver resection prior to LT did not increase significantly the operative technical difficulty of the transplantation procedure. Indeed, compared with PLT, reoperations after SLT were more frequent, but this did not translate into difference in retransplantation or prolonged duration of intensive unit or hospital stay for the SLT group. Several factors may explain these findings. First, the study population was highly selected and treated at a single specialized hepatobiliary surgery and transplantation center. Progress in hepatic surgery has shown that reoperation for recurrent HCC^27,28 or liver metastasis^29,30 is a feasible, efficient treatment with acceptable morbidity in selected patients. Second, tolerance to prior liver resection and fitness for subsequent LT may actually represent a sign of good candidacy for surgery. Third, minor resection, especially via the transthoracic approach, appeared to have very little impact on the conduct of the LT.¹⁷ The concept that major resection may increase the difficulty of subsequent surgery is noted but does not appear to contraindicate LT. In addition, the anticipation of potential difficulty with reoperation or postoperative course may have led to particular precaution in the care of these patients transplanted by senior surgeons. Thus, the concept of liver resection before LT is feasible and is to be considered in selected patients with HCC as several treatment strategies.

Resection as a Bridge to Transplantation

Tumor progression during the wait period for transplantation is the most important impediment to LT as a treatment of HCC.¹¹ An estimated 10% of patients with HCC on transplant waiting lists die before undergoing LT.³² Therefore, bridge treatments including transarterial chemoembolization and radiofrequency ablation are increasingly used to halt or delay tumor progression during the waiting period to transplant. The finding that LT after resection is feasible reopens the potential utility of liver resection as a bridge to transplantation for patients with preserved liver function and small HCC. Advantages of resection as a bridge treatment include the complete removal of the tumor with a longer period of tumor control than other bridge treatments. To minimize the potential morbidity of a future LT, elective transthoracic resection for tumors adjacent to the diaphragm should be considered (which are not appropriate for radiofrequency).¹⁷ When transabdominal resection is indicated, anatomic segment-oriented resections are preferred.^33,34

Resection Selection

In the era when the transplant community debates about expanding LT criteria for HCC, the strategy of prior resection provides whole-specimen pathology before LT planning. Presence of vascular invasion, precise size and number of satellite nodules, presence of capsule, and degree of differentiation are among the most powerful predictors of survival in patients with HCC. In addition to pathologic factors, natural history of tumor behavior can be incorporated into future treatment planning. Patients with poor prognostic criteria such as macroscopic vascular invasion and poor differentiation may not be candidates for transplantation and could therefore be removed from the list.^35–37 Other patients might be selected for early transplantation de principe (deliberately) considered to be at high risk for recurrence while they are within transplantation criteria. This group might include those with a positive microscopic surgical margin or the presence of satellite nodules in the resected specimen. Indeed, in our study, 2 of 3 patients transplanted de principe were found incidentally to harbor HCC in the resected specimen. Similarly, some patients will be transplanted for deterioration of liver function without morphologic evidence of tumor recurrence. In this small subgroup in our study, 2 of 4 were found to have incidentally discovered HCC after transplantation. Some patients, likely not to recur after resection (particularly those with solitary tumors without high risk pathologic findings), might be the best candidates for surveillance after resection until and unless recurrence occurs.¹¹

Salvage Transplantation

Given the known value of resection in small HCC for patients with preserved liver function,^21–23 Poon et al propose to reconsider primary resection for small HCC, with reservation of transplantation as a “salvage” treatment of recurrence or progressive liver dysfunction.¹¹ The Hong Kong group showed that when recurrence occurs after resection of small HCC, up to 79% of patients are considered transplantable using the same criteria of primary transplantation for HCC.¹¹ Thus, it may not be necessary to list patients who undergo resection for LT unless recurrence is detected on surveillance. The present series focused on the impact of prior liver resection for HCC on perioperative morbidity and survival of LT in patients who strictly met Milano criteria at the time of resection and at the time of LT. This study reveals that disease-free and overall survival, measured from the date of transplantation, are equivalent for primary LT and secondary LT. Moreover, secondary LT patients enjoyed a mean 20-month disease-free interval before recurrence or listing for LT, which argues in favor of resection prior to transplantation in these patients. Finally, patients are not on the transplant waiting list after resection, so that organs can be allocated to other patients during the interval of observation. Decision-making regarding living donation for transplantation may be facilitated in this strategy.^38–40 The delay given by prior resection and the pathologic analysis of the resected specimen may help to organize this procedure.

CONCLUSION

Transplantation following hepatic resection for HCC is feasible. In selected patients, liver resection prior to LT does not significantly increase the technical difficulty of the transplantation procedure or impair the survival after LT. Liver resection may function as a bridge treatment to transplantation, may enhance selection of candidates for LT based on analysis of the entire pathologic specimen, and preserves the possibility of salvage transplantation in the event of recurrence. Finally, implementation of this approach might reduce the number of patients on the waiting list for transplantation.

Discussion

Dr. S.T. Fan: Professor Belghiti, I congratulate you on your excellent study, I think this is the first documented evidence that salvage transplantation is good for patients with recurrent HCC.

On the whole, I agree with you; however, I have one comment and two questions.

The comment is: although you said that the two groups of patients had the same outcome, as a matter of fact, the group of patients who had salvage transplantation had much more infections, bleeding, and reoperations. In other words, we have to spend extra effort to achieve the same result. Therefore, we may have to tell our patients the possible scenario if we really want to adopt this kind of a sequential type of management.

My question is: since radiofrequency operation is coming into scene, and you can treat many small tumors effectively and successfully, do you think in the future surgical resection could be replaced by radiofrequency ablations because surgery eventually and actually brings a lot of problems to the patients? (There could be a number of technical difficulties, especially when the liver hilum has been touched, or there has been dissection in the liver hilum before the liver transplant.)

The second question is about vascular permeation of the tumor. Since you mentioned that the pathologic examination of the resected specimen from partially hepatectomy, can give you some idea about the selection of the patients, if the patient has vascular permeation you in the tumor specimen? Would you or would you not consider him a candidate for liver transplant?

Dr. J.B. Belghiti: Thank you for your kind comments. I would start with the second question. I think that as much as we look through our work we open many doors and we have more questions than in the beginning. I agree with you that resection selection brings up two possibilities. The first possibility is that when you find some factors, such as positive margin of resection, which create a risk for early postoperative recurrence, you should discuss transplantation as quickly as possible after resection. On the other hand, for some completely resected tumors, the prognosis after resection is similar as the prognosis after liver transplantation so the patient should not be transplanted immediately but undergo surveillance. Concerning vascular invasion, there is no doubt that if major vascular invasion is discovered, secondary liver transplantation is not indicated. If microvascular invasion is discovered, liver transplantation remains indicated. We consider that these questions remain open since pathologic data have never before been available for consideration before transplantation.

The other question is what to do with tumors situated on the top of the liver where the percutaneous radiofrequency is difficult or contraindicated. Our tendency now is not to hesitate to perform transthoracic resection, firstly because the tumor is completely removed, secondly because you have the specimen for pathologic analysis, and thirdly because this approach had no deleterious technical consequences during the transplantation procedure.

Dr. P. Neuhaus: Dr. Belghiti, you answered already half of my question: do you see in your concept a principal biologic difference between resection and radiofrequency ablation if this is complete, or do you consider that both treatments are comparable to primarily treat such a patient?

Also, I would like to ask you what about the difference between Child A or B or C patients and third with the possibility of using living related donors when such a family comes and you have the option to do the transplant immediately or you do a primary ablation or resection or therapy and then proceed. If you get recurrence, how would you position yourself in this instance?

Dr. J.B. Belghiti: Your comment concerning the possibility of living donation is very important. Finally, decision making regarding living donation for transplantation may be facilitated in this strategy. The delay given by prior resection and the pathologic analysis of the resected specimen may help to organize living donation. Secondly, you make a very important point concerning liver function. In this retrospective study, all of our patients resected had a good liver function (Child A). Because resection in patients with compromised liver function (Child B and C) usually results in significant morbidity and mortality, we consider liver transplantation as a primary treatment.

Concerning the efficacy of radiofrequency as a bridge treatment before transplantation, we consider resection as a better treatment because in our group of patients, exclusively treated by percutaneous ablation, we only found in the explant liver a 60% rate of complete necrosis associated in half of cases with persistent satellite nodules.

Dr. K.G. Tranberg: This is a very interesting paper. When trying to interpret your findings, the problem, as you have indicated, is the selection of patients. You have looked at patients having undergone liver resection and then being suitable for transplantation. But what happened to the other patients that had liver resection? To be clinically helpful, I think that the results in all patients primarily undergoing liver resection should be compared with the results in patients undergoing primary liver transplantation. The importance of selection is underscored by the fact that those who underwent secondary or salvage liver transplantation had a waiting time of some 20 months. Maybe primary liver transplantation is the best choice, but yet we don’t know.

Dr. A.V. Höckerstedt: Excellent study, Dr. Belghiti. I only have three short questions.

In cancer studies, we would like to know what is the disease-free survival? Was there any difference in the two groups? The second one is, that you excluded in your survival analysis death within the first 3 months, which in my opinion is not appropriate, so we would like to see the real survival, also those included who died within the first 3 months.

And finally, why did you keep your patients in the ICU for such a long period, about 10 days in both groups?

Dr. J.B. Belghiti: Thank you for your comments and pertinent questions. Firstly, the disease-free survival is similar in the two groups after inclusion of deaths within the first 3 months. Concerning the long duration of ICU, it is habitual in France for patients to go to the ICU systematically and to stay until early major postoperative complications are eliminated.

Dr. R.A. Adam: Thank you very much for your communication. I first have a comment and then a question.

My comment is that this is a very clinically relevant study. And because of this, I think we should be very careful on the conclusions.

As you know with the Milan and the Geneva groups, we reported in 1999 the first 12 cases of secondary liver transplantation with results which appeared similar to that of primary transplantation. But indeed, we have changed our mind because what we have to consider is an intention-to-treat analysis. In our series, from 358 hepatocellular carcinomas with cirrhosis operated during the same period, 163 were treated by primary resection and 195 by primary transplantation. In the transplantation group, 10% of the patients were dropped out from transplantation, so that the final transplantability was 90%. In the primary resection group, from the 163 patients, 60% of whom were considered retrospectively as transplantable, only a minority of patients come to transplantation. They accounted for 17% of patients who had tumoral recurrence and 3% of patients who had no recurrence. Therefore, on an intention-to-treat basis, in a center where the two treatments were available, the transplantability rate of this resected but initially transplantable population was only 20%.

When we see now the disease-free survival at 5 years, it was 60% for the transplanted group and only 29% for the group of patients resected but initially transplantable, including those with a secondary transplantation. In view of these results, what was apparently an attractive policy (resection and secondary transplantation), is indeed associated with a much lower long-term survival than primary transplantation on an intention-to-treat basis. This is the reason why I think we should be very careful on the message we deliver.

As you have an extensive experience of more than 100 liver resections for hepatocellular carcinoma on cirrhosis, my question is: How do you explain that in your study, so few resected patients come to transplantation thereafter?

Dr. J.B. Belghiti: Thank you for your comment. The question is why did we resect these good candidates for liver transplantation. Firstly, it is a retrospective study without intention to treat. Secondly, some patients or families do not accept the option of transplantation; and thirdly, during the early phase of this study, transplantation was not the standard for small HCC.

Dr. P. Neuhaus: Thank you. This is probably an ongoing discussion for the next few years, and maybe since a number of surgeons in this ESA group are doing these tumor transplants, we can convince ourselves to a prospective study sometime.