Living Donor Liver Transplantation for Hepatocellular Carcinoma: A Single-Center Experience in Taiwan

Abstract

Background. 

Living donor liver transplantation (LDLT) demonstrates certain survival benefits over deceased donor liver transplantation for hepatocellular carcinoma (HCC) but there is no consensus on criteria for the use of LDLT for HCC for hepatocellular carcinoma (HCC) taking into account strategies to improve survival.

Methods. 

Thirty-five patients (89% men) underwent LDLT for HCC. The mean age was 51 years (range, 22–61). The median disease severity scores were B, 11–20, and 2B for Child-Turcotte-Pugh, Model for End-stage Liver Disease, and United Network for Organ Sharing, respectively. The transplant records were retrospectively analyzed.

Results. 

All were within Milan criteria at time of transplantation. A novel approach to downstaging tumors initially beyond the Milan criteria was evaluated using transarterial embolization or percutaneous ethanol injection. Our initial results were encouraging as recipients whose tumors had been downstaged had not had recurrence to date. Seven (20%) patients underwent hepatectomy for HCC before undergoing transplant. The overall mean posttransplant follow-up in this series was 40.3 months (range, 23–75). The overall posttransplant complication rate requiring intervention was 11%. There was only one malignancy recurrence for an overall recurrence rate of 3%. Vascular invasion and small- for-size transplants did not seem to influence tumor recurrence. The nonestimated recipient 1-year, 3-year, and 5-year survivals were 98%, 96%, and 90%, respectively.

Conclusion. 

This review emphasizes the need for early disease recognition and prompt intervention when Milan criteria are met to improve survival from HCC after LDLT.

Early cancer detection and accurate pretransplant staging are the key factors to successful liver transplantation (LT) outcomes for hepatocellular carcinoma (HCC). Using Milan criteria, excellent recurrence-free survival of >92% at 3 years is now achievable (¹).

Although living donor liver transplantation (LDLT) assumed a key role in the pre–Model for End-stage Disease (MELD) era, the current MELD prioritization for HCC reduced the necessity for LDLT in countries such as the United States and parts of Europe (²). Unfortunately, there is a prolonged waiting time for candidates to receive allocated deceased organs in Taiwan due to a very low deceased donation rate (7.2 per million population in 2006), which contributes in patient dropout due to tumor progression. Advances in adult LDLT demonstrate survival benefits over deceased donor liver transplantation (DDLT) in regards to decreasing waiting time and dropout because it is not restricted by organ allocation system (³). However, other studies observe that although LDLT decreases waiting time, it is also associated with higher rates of disease recurrence (⁴).

There is no consensus on criteria for the use of LDLT for HCC. Centers offering LDLT use currently accepted criteria in DDLT or adopted their own criteria to justify its use for HCC. We review our center's experience and treatment strategies in LDLT for HCC including selection criteria, pretransplant locoregional tumor control, immunosuppression, adjuvant doxorubicin chemotherapy posttransplant with microvascular tumor invasion, and salvage transplantation.

PATIENTS AND METHODS

From January 1999 to December 2004, we performed 88 adult LDLT. Thirty-five (40%) of these transplantations were for HCC. Demographic and clinical data were collected prospectively and analyzed retrospectively.

Recipient Selection

The criteria for hepatic resection for HCC included anatomically resectable disease, absence of distant metastasis, and adequate liver reserve based on liver function and indocyanine green assay. Cirrhosis alone was not considered a contraindication to resection. Figure 1 summarized the decision-making process for when to consider resection over locoregional treatments. Patients with unresectable tumors due to advanced cirrhosis but with no signs of metastatic disease or vascular invasion may be accepted for LT. Using imaging methods, the Milan criteria were used for selecting patients for LT.

FIGURE 1.: Decision-making flow chart for primary resection vs. locoregional therapy.

With initial recipient evaluation, candidates were offered the option of DDLT or LDLT. Because there is shortage of deceased donor grafts in Taiwan and there was no prioritization based on the diagnosis of HCC, the option of LDLT was often times chosen by the patient and family as they come with the donor and avoid a long wait before LT.

Acceptance into the waiting list required that a candidate fits the Milan criteria as required by the Taiwan National Health Insurance. Because 99% of the country's population was covered by national health insurance, a cost-effective preliminary screening was made which consist of plain and contrast-enhanced computed tomography (CT). If by this screening a candidate was found to be beyond Milan criteria (i.e., tumor number and size, gross vascular invasion), then this candidate will not be listed for transplantation or a tumor downstaging therapy may be instituted before listing for transplant. If a recipient-candidate was HBeAg reactive or HBV DNA positive, lamivudine, and/or adefovir were given to decrease viral load before transplant. Meticulous search for occult metastases secondary to HCC as well as search for silent malignancies from other organ systems was performed. Findings of these metastases preclude transplantation. A meticulous search for occult HCC metastases included brain, chest, abdomen, and pelvis plain and contrast-enhanced CT scans, and bone scan. A positron emission tomography scan was only requested when there were equivocal findings on CT. When there was still question on the possibility of metastasis, particularly in the evaluation of pulmonary nodules (e.g., pulmonary tuberculosis vs. metastasis vs. other pathology) or enlarged abdominal lymph nodes, histological confirmation was made prior to transplant.

While awaiting LDLT, alpha-feto protein (AFP) and abdominal ultrasound were performed every 1–2 months. In those with adequate liver reserve, transarterial embolization (TAE) or percutaneous ethanol injection (PEI) was performed for tumor recurrence during the waiting period. Patients with tumor progression beyond the Milan criteria that could not be effectively downstaged were removed from the LT list after a multidisciplinary meeting decision.

Tumor Downstaging

Patients with tumors beyond the Milan criteria at initial presentation but with no signs of metastatic disease or vascular invasion as documented by CT of the chest, abdomen and pelvis were downstaged using TAE and/or PEI. Superselective embolization was performed as much as possible using a mixture of iodized oil and ethanol. The result of embolization was evaluated by CT 2 weeks after the procedure. The results were classified as complete if lipiodol occupied the whole tumor (100%) or incomplete embolization if lipiodol occupied below 100%. All cases with incomplete embolizations received a second embolization and/or PEI if liver function was acceptable. Radiological restaging after TAE was based on completely embolized tumor being considered as no tumor; for incompletely embolized tumor, a reassessment of the unembolized part of the tumor and number was made. Superparamagnetic-iron oxide magnetic resonance imaging was used to assess small equivocal tumors. Patients who met the Milan criteria after TAE and/or PEI were transplanted as soon as possible.

Donor Selection

Our protocol for evaluating the live liver donor had been described previously (⁵). We emphasized the importance of the donor's voluntary and altruistic intent. The psychological status of the donor was evaluated by both social worker and clinical psychologist. The donor is given the opportunity to withdraw anytime with the assurance of a medical excuse provided by the transplant team. If the donor had a compatible blood group, negative serology for HBsAg and anti-hepatitis C virus (HCV), CT angiography with volumetry was performed to estimate liver volume and fatty content. The right lobe was selected to provide a graft of at least 40% of the recipient's standard liver volume as estimated using the Urata formula (⁶) and at least 30% remnant liver volume for the donor.

Operative Management

Our techniques of donor graft hepatectomy and recipient total hepatectomy in LDLT have been described previously (^{5, 7}). The graft consisted of the right lobe with or without the middle hepatic vein (HV) (⁸). The recipient right HV opening was widened by trimming the vessel edges. The size of this opening was adjusted to measure wider than the graft HV. Assuring the correct orientation of the graft and recipient vessels, the graft HV was anastomosed to the recipient right HV opening with the IVC cross-clamped. Venovenous bypass was not used in any recipient. Multiple graft HV were reconstructed either via direct caval anastomosis, graft venoplasty (⁹) or by use of interposition grafts to the inferior vena cava (IVC). The graft was reperfused after portal vein anastomosis followed by hepatic artery reconstruction using microsurgical techniques. Biliary reconstruction was performed via duct-to-duct anatomosis without stent whenever possible. Intraoperative Doppler ultrasound was performed to check vascular flow patterns and velocities after vascular reconstruction, and before and after abdominal closure.

Histopathologic Analysis

The explanted liver was examined by an experienced LT histopathologist. All specimens were sliced at 5-mm intervals and detected tumors were carefully noted and examined. Tumor grading was performed according to the Edmondson criteria. Tumors with extensive necrosis where histologic grading can not be determined were classified as “undetermined due to necrosis.” Lymph nodes were examined for metastatic disease.

The nonneoplastic liver underwent histopathologic and immunohistochemistry studies to determine degree of cirrhosis and hepatitis virus activity, if present. Liver inflammation grading was based on Hepatitis Activity Index score.

Posttransplant Management

Immunosuppression consisted of tacrolimus (Fujisawa, Osaka, Japan) in combination with steroids. Mycophenolate (Roche, Basel, Switzerland) was used in patients whose indications included: a) more potent immunosuppression was desirable, b) renal-sparing benefits, and c) to decrease the dose of tacrolimus required. Steroids were tapered and discontinued within 6 months in the majority of patients. Tumor recurrence was assessed by monthly measurements of AFP and liver ultrasonography every 3 months. If ultrasonographic findings were inconclusive, a CT of the liver and abdomen was requested.

Patients whose explanted liver showed microvascular tumor invasion received a 10-cycle doxorubicin-based chemotherapy (10 mg/m², first cycle given at 3 weeks postLT, then every week for a total of 10 cycles). All rejections were biopsy-proven and managed with either increased immunosuppression or intravenous pulse methylprednisolone (10 mg/kg/body weight). Steroid pulse therapy was avoided whenever possible particularly in HCV patients.

PreLT HBV reactive recipients were placed on indefinite treatment with lamivudine and/or adefovir, and hepatitis B immune globulin postLT. HCV recurrence, based on serologic parameters with concordant increase in liver enzymes, was managed with pegylated interferon alpha-2a and ribavirin regimen for a minimum of 6 months.

Data Analyses

The minimum follow-up of transplanted patients was 23 months or until death. The outcome measures of interest were survival and recurrence-free survival. We reported all causes of mortality and recurrence of HCC as our primary end-points of interest.

Univariate analysis was performed on multiple clinical and pathological data. The clinical data studied included age, gender, blood type, albumin, bilirubin, AFP, prothrombin time, ascites, encephalopathy, pretransplant hepatic resection for HCC, neoadjuvant treatments for HCC, functional status, etiology of underlying hepatitis disease, and perioperative blood transfusion. The tumor-pathologic data analyzed were tumor size, location, number, presence of tumor capsule, histologic grade, tumor necrosis, vascular invasion, and bile duct invasion. Tumor data were categorized separately as to that described preoperatively and at liver explant analysis.

Baseline data were expressed as mean or median when appropriate. Comparison of relevant variables used the Wilcoxon rank sum test. Comparison between groups was done by using the Student's t test for quantitative variables and the Fisher's exact test for qualitative variables. A P value of <0.05 was considered statistically significant. Survival analysis was based on long-term actual survival (nonestimated). Data were analyzed using computer software STATA (Stata Corporation, College Station, TX).

RESULTS

There were 31 (89%) male and 4 (11%) female recipients. The mean age was 48 years (range, 22–61) for men and 54 years (range, 50–60) for women. One patient had a preoperative diagnosis of primary HCC without cirrhosis. All were either with hepatitis B, C, or B and C viruses. The median disease severity scores were B, 11–20 (before the addition of 22 points for stage II HCC), and 2B for Child-Turcotte-Pugh, MELD, and United Network for Organ Sharing, respectively. All were within the radiologic Milan criteria for LT for HCC at time of transplantation. Seven (20%) patients underwent hepatectomy for HCC before undergoing LT. Twenty-five (71%) patients underwent preoperative locoregional tumor control before LT. Seventy-one percent (71%) had normal pretransplant AFP results. The median timeline between pretransplant therapy of the recipients and the transplant operation was 3 months (range, 1–6). Tables 1 and 2 summarize the patients' demographic and clinical data, respectively.

TABLE 1: Patient demographics

TABLE 2: Patients' pretransplant clinical characteristics (n=35)

Operative Outcomes

Four recipients received right lobe grafts with the middle HV and 31 recipients received right lobe grafts without the middle HV. The mean graft-to-recipient weight ratio was 1.06 (range 0.7–1.7). The mean graft-to-liver volume ratio was 56.6 (range, 41.4–81.3). The mean recipient operative time was 688 min (range, 490–952). The mean cold ischemia time was 68 min (range, 35–119).The mean warm ischemia time was 58 min (range, 31–91). The average recipient intraoperative blood loss was 1968 mL (range, 220–11,375). The average amount of packed red blood cells transfused was 1336 g (range, 0–6716). Vascular outflow patency and velocity determinations using two-dimensional Doppler ultrasonography were done after reconstruction. There was no intraoperative HV, portal vein, or hepatic artery complication in this series.

There were four postoperative complications that required interventions. These included two biliary strictures and two portal vein stenoses. The biliary strictures were managed by revision of the initial duct-to-duct anastomoses to Roux-en-Y hepatodocho-jejunostomy, whereas the portal vein stenoses were both successfully managed by percutaneous interventional radiology balloon dilatation. The overall operative complication rate requiring corrective intervention was 11%.

The overall rejection rate was 23%, which included six mild acute cellular rejection (ACR) and two moderate ACR. All were successfully managed by either pulse methylprednisolone treatment or increase in current immunosuppression regimen. There was no steroid-resistant ACR; muromonab CD3 (OKT3) had not been used. Likewise, there was no chronic rejection cases. All rejections were biopsy proven before treatment.

HCV recurrence requiring treatment with pegylated interferon and ribavirin occurred in two of eight (25%) recipients, whereas HBV recurred in only 1 of 25 (3.7%) recipients. One (¹) HCV-reactive recipient developed de novo HBV infection 1 year posttransplant.

There were three mortalities in this series. One recipient (LDLT 180) died of hepatic failure secondary to small-for-size syndrome 2 months posttransplant as in-hospital mortality. Another recipient (LDLT 92) died of multiple lung metastases from HCC recurrence 1 year posttransplant. The third recipient (LDLT 137) died of veno-occlusive disease of the liver of unknown etiology (due to nonavailability of donor graft for retransplant) at 26 months posttransplant. The overall mortality rate was 9%. The mortality rate from HCC recurrence was 3%. There was 1 retransplantation due to secondary cirrhosis caused by biliary stricture 2 years after LDLT. There was no graft loss due to rejection or hepatitis recurrence. The mean follow-up was 40.3 months (range, 23–75). The overall 1-year recipient survival rate was 98%. The actual, nonestimated 3-year and 5-year survival rates were 96% and 90%, respectively.

Tumor Characteristics and Histopathological Analyses

In one patient who underwent pretransplant TAE, no malignancy was seen in the explanted liver examined. Tumor histologic grade could not be determined in six explants due to extensive tumor necrosis (>60% necrosis). Twenty-two (65%) of explants showed moderately differentiated HCC. Nine (26%) explants showed microvascular invasion, whereas six (18%) explants showed satellite nodules. There was no tumor bile duct invasion in all liver explants examined. Table 3 summarizes the histopathologic data.

TABLE 3: Histopathology results of explanted livers (N=34)

Using histopathologic Milan criteria for tumor number and size, there were only two (6%) recipients beyond the criteria in terms of tumor number (1–4 tumors; 1–5 tumors) and one (3%) recipient beyond the tumor size of 5 cm for a single lesion.

Eight (23%) patients beyond the Milan criteria at the time of LDLT listing underwent tumor downstaging. The tumor size, numbers, and outcomes of downstaging are shown in Figure 2.

FIGURE 2.: The use of TAE and/or PEI in pretransplant downstaging of tumor size and number to fit the Milan criteria (each asterisk represents a patient). The eight patients were outside Milan criteria prior to listing for transplantation. After tumor downstaging, all eight patients fit the Milan criteria at the time of transplantation.

Recurrence and Disease-Free Survival

HCC recurred in only one of 35 recipients. In this patient (LDLT 92), the graft weight-to-recipient weight ratio was 1.4 and the graft volume-to-standard liver volume ratio was 61.3%; further, he satisfied the pretransplant radiologic Milan criteria. However, on histopathologic examination, the liver explant showed microvascular invasion. The histologic grade was moderately differentiated and the tumor showed >60% necrosis from previous TAE. He did not receive prophylactic chemotherapy immediate posttransplant. Six months posttransplant, he developed multiple lung metastases; on the seventh month posttransplant, he had multiple liver HCC recurrence. He received four cycles of doxorubicin-cisplatin salvage chemotherapy upon detection of lung metastases. Six months later (at 1 year posttransplant), he succumbed to respiratory failure from lung metastases.

Three patients with microvascular invasion did not receive chemotherapy. The first patient (LDLT 180) died due to small-for-size syndrome before chemotherapy could be instituted. The second and third patients (LDLT 92 and 188) did not undergo chemotherapy per patient preference.

On analysis, there was no association between recurrence and tumor pathological variables as histologic grade, presence of tumor capsule, satellite nodule, or degree of tumor necrosis. There was also no correlation between presence of microvascular invasion and recurrence (1/9 vs. 0/26, P=0.257, Fisher's exact test). Table 4 summarizes the data on microvascular invasion and recurrence.

TABLE 4: Histopathologic profile of patients with microvascular tumor invasion, patient management, and outcome (n=9)

Salvage Transplantation for Post-HCC Resection

Seven patients had undergone hepatic resection for HCC before receiving LT as protocol. All seven patients were within Milan criteria at the time of hepatic resection. Excluding two recipients whose indication for LT was decompensated cirrhosis years after HCC resection, the mean interval to HCC recurrence was 24 months (range, 3–79). The median interval to transplant was 12 months (range, 3–84) after the recurrence detection. There was no microvascular invasion noted in all histopathologic reports of explanted livers in this group. There was no HCC recurrence in this group after LT. The mean follow-up was 41.9 months posttransplant (range, 16–53) in this subset of patients. Table 5 summarizes the data on patients with previous hepatectomy prior to LT.

TABLE 5: Interval to recurrence, interval to transplantation, and histopathologic profile of patients who underwent liver resection prior to transplantation (N=7)

Although two recipients showed no radiological evidence of tumor recurrence prior to LT, liver explant analysis revealed HCC tumor.

DISCUSSION

Liver transplantation has been successful in treating limited-stage HCC affecting cure of both the tumor and cirrhosis (^10–12). LT for HCC is limited by organ availability (^{3, 13, 14}). LDLT improves access to transplant. Because LDLT is not restricted by waiting time and organ allocation from a deceased donor, it offers a substantial advantage in patients with early stage HCC who would otherwise have waited for several months or years for a DDLT. This is supported by analytical studies that demonstrated the theoretical advantage of LDLT over DDLT based on the latter's long waiting time and dropout rate (^{15, 16}). Despite this theoretical advantage and direct survival advantage of LDLT over DDLT, there was a higher recurrence rate observed in an LDLT series reported by Lo (³). This observation contradicts an earlier series by Gondolesi, who reported that the incidence of recurrence as well as patient survival and freedom from recurrence in LDLT are comparable to results after DDLT (¹⁴). We re-examine factors and our strategies that may influence outcome, survival, and freedom from recurrence in our series of LDLT patients.

Staging and criteria for selection of patients who would benefit from LT the most are keys to predict the prognosis of patients with HCC. The landmark study by Mazzaffero in 1996 led to careful selection of patients who satisfied specific tumor criteria. The Milan criteria have established through time that excellent recurrence-free survival of 92% at 3 years and a 4-year survival rate equivalent to patients transplanted without HCC can be achieved (¹). We have used the Milan criteria in selecting our patients undergoing LDLT because the Taiwan National Health Insurance System would only guarantee insurance coverage for patients satisfying the Milan criteria. All our patients satisfied the pretransplant radiologic Milan criteria including tumor down-staged patients.

We have used TAE and PEI as pretransplant therapy for patients with localized HCC while awaiting donor livers and to down-stage tumor size and number to satisfy the Milan criteria. Sixty-eight percent of our patients received TAE and 28% received PEI prior to LDLT. In an earlier study involving 29 patients at our center to support this practice, we found that TAE is an effective treatment for HCC before LT. Moreover, excellent survival was achieved in patients who initially did not fit Milan criteria (¹⁷). In our experience, meticulous pretransplant TAE helps in reducing the dropout rate from the waiting list. Figure 2 summarized our experience with downstaging tumor size and number using TAE and/or PEI to meet the Milan criteria. We speculate that downstaged tumors may represent tumors with favorable biologic behavior resulting in good posttransplant outcome. In a corollary study evaluating the effectiveness of pretransplant treatments for HCC, Wong showed that pretransplant treatment with TAE, radiofrequency ablation, or cisplatin injection can generally achieve tumor necrosis (¹⁸). However, this study failed to demonstrate whether such treatments allowed longer waiting time for transplant and eventually affected long-term outcome.

We only have one recurrence in this series, for an overall recurrence rate of 3%. There are several theories that suggest why LDLT has a higher observed recurrence rate when compared to DDLT. One theory suggests that histologic differentiation may play a role in the recurrence. Since the patient comes with a donor and LT is “fast-tracked,” there may be not enough time for the tumor to exhibit its biologic behavior or aggressiveness as would have been seen if the patient is allowed the waiting time for tumor progression. At Northwestern University, they observed the similar high recurrence in LDLT patients whose evaluation and transplantation have been fast-tracked (⁴). Our series does not show association between histologic tumor grade and recurrence. Another theory for the high recurrence rate is due to tumor-enhancing effects of cytokines, growth factors, and transcription factors associated with liver regeneration (⁴). Researchers have linked this theory with the recurrence of HCV after LDLT. This theory requires the presence of extrahepatic microscopic tumor at time of transplantation since almost all patients are within the Milan criteria and surgical techniques are also almost the same. This second theory may explain the recurrence in our patient because he had microscopic portal vein tumor invasion on explant pathology.

Other authors have associated small-for-size grafts with increased risk of HCC recurrence (¹⁹). The rapid regeneration rate necessary for small-for-size grafts to compensate for the needed liver function may also be associated with tumor-enhancing effects of cytokines and growth and transcription factors. There were two patients (LDLT 180, 193) with a graft-to-recipient weight ratio of <0.8 and six patients (LDLT 119, 139, 170, 180, 188, 193) with a graft-to-liver volume ratio of <50%. In these two groups of patients, there was no HCC recurrence. The only HCC recurrence developed in LDLT 92 who had a graft-to-recipient weight ratio of 1.4 and graft-to-liver volume ratio of 61.3%. If we exclude LDLT 180 who died at 2 months posttransplant, it would seem that small-for-size per se does not influence HCC recurrence in this series. However, due to relatively small number of graft-to-recipient weight ratio <0.8 or graft-to-liver volume ratio <50% and a single recurrence, again, we can only speculate on this observation.

In 71% of our patients, the AFP was <20 ng/mL. Only 26% and 3% were that of 21–400 and >400 ng/mL, respectively. The low AFP values observed in this series may create two hypotheses: 1) these tumors are non-AFP or low AFP producing, or 2) these are well-controlled tumors from previous treatment prior to transplant. Although we do not have pretreatment AFP values (as most patients have undergone loco-regional treatment/s prior to transplant referral) or cell biology behavior indicators (i.e., cell mitotic indices, tumor doubling time, etc.), by deductive reasoning we can speculate on the following based on this series: 1) the low AFP may represent favorable tumor biology, 2) patients with favorable tumor biology respond well to locoregional treatment, and 3) tumors which respond well to locoregional treatment have low HCC recurrence outcome posttransplant. Further, if we consider that all tumors are AFP-producing, theoretically, then tumor size and number may be correlated with an elevated AFP and AFP may be used as a surrogate marker for tumor size and number. It is worthwhile to note that in 51% of the cases, there was only one tumor; in 74% of the cases, the tumor was <3 cm size. Hence, this may explain the use of AFP as surrogate marker for HCC recurrence in some studies.

Tumor-related factors as size, number, and histologic grade were not related to recurrence, which is explained by the fact that patients are highly selected based on criteria and that all met the criteria. The presence of a satellite nodule is not a factor for HCC recurrence even in small HCC in this series where 76% of the tumors are <3 cm size. This finding is in contrast to Plessier's finding in 70 patients with small HCC which showed that satellite nodules were the only statistically significant predictor of recurrence (²⁰). We also find no correlation between presence of satellite nodules and microvascular invasion. Tumor necrosis is also not related to tumor recurrence after LT in this series. Although partial necrosis on HCC nodules were related to tumor recurrence after LT is reported (²¹), 59% of patients in our series showed partial necrosis (between <10% and 10–60% necrosis) and yet none of these patients developed recurrence. The recurrence occurred in a patient with >60% necrosis.

Surgical resection is the first option offered to HCC patients with resectable tumors without significant portal hypertension and acceptable liver function reserve. This strategy of offering resection as first treatment option reserving LT for disease recurrence avoids the unnecessary risks of early use of immunosuppression. In this series where we have seven patients (Table 5) who initially underwent resection before LT, the mean interval to recurrence was 24 months and the median interval to transplant was 12 months from recurrence. There was no recurrence in any patient previously resected after LT. We believe this strategy is cost-effective when considering cost of immunosuppression medication. The reoperation for transplant did not increase overall recipient morbidity. The waiting time is justified as no patient has recurrence to date.

Salvage transplantation has been advocated for patients with high pathological risk of recurrence after surgical resection as manifested by microvascular invasion and/or presence of satellite nodule (²²). Although the series is small in a reported analysis, these two parameters identify patients at higher risk of recurrence, thus allowing them to be listed for liver transplantation at time without proven recurrent malignant disease. This conclusion should be taken with caution. As discussed, tumor biology may play a role in recurrence postLT and that vascular invasion makes a locoregional tumor a systemic disease. Since this subgroup of patients already showed tumor aggressiveness preLT, the chances of recurrence postLT is high. We have had only one patient (LDLT 84) where we used “fast-track” LDLT for high pathologic risk of recurrence as described above. This patient received doxorubicin chemotherapy preLT and is recurrence-free 44 months postLT.

High immunosuppression and prolonged use of steroids predispose patients to their side effects. Theoretically, the more immunosuppression is provided, the higher the risk of recurrence due to the body's failure to produce natural killer cells to autoregulate malignant transformation in cells. Our strategy is to keep immunosuppression at a minimum as long as clinically possible; thus we are able to achieve near normal liver enzyme results. Whenever possible, steroids are withdrawn within 6 months. In cases of ACR, and particularly in HCV recipients with ACR, we prefer to increase immunosuppression rather than give boluses of intravenous methylprednisolone whenever possible. This low immunosuppression strategy, early discontinuation of steroids, and avoidance of intravenous bolus methylprednisolone whenever possible may play roles in the low HCC and hepatitis recurrences seen in our series.

All HCC patients in this series were hepatitis virus- related. Two patients have both HBV and HCV. We find a trend that HCC with HBV is different from HCC with HCV. HCC with HCV tends to be more associated with microvascular invasion (four of nine) and lesser chance of undergoing resection prior to transplant (one of seven). Although we have only one HCC recurrence, it seems that the type of co-existing hepatitis virus does not influence tumor recurrence.

In this review, the long-term actual survival (nonestimated) and freedom from recurrence were taken as the outcome parameters and risk factors were identified on the basis of these rates since actuarial survival rates can differ significantly from actual rates indicating that actuarial rates do not always reflect reality. A drawback, however, is that analysis using actual survival rates used for identification of risk factors is based on a smaller number. The observation that the low HCC recurrence and high graft and recipient survival rates in this series are attributed to adherence to the Milan criteria and prompt interventions when criteria are met. But due to small number and nonavailability of a cohort to compare our LDLT outcome with that of DDLT, we can only speculate on survival issues. However, in the absence of a direct comparing group, a review of the results of LT for HCC published in literature (^23–27) indicates that our LDLT within Milan criteria and reported DDLT within Milan criteria may be similar in outcome. Our treatment strategies have yielded beneficial and excellent outcome to our patients within the Milan criteria demonstrating its durability to justify the use of LDLT for patients with HCC.