Underestimation of the influence of satellite nodules as a risk factor for post-transplantation recurrence in patients with small hepatocellular carcinoma



Liver transplantation offers good results in patients with small hepatocellular carcinoma. However, 3 to 15% of patients still have recurrence, suggesting that factors other than the size and number of nodules are implicated. The aim of our study was to identify predictive factors of recurrence in patients with small hepatocellular carcinoma. Seventy consecutive patients fulfilling Milano criteria and who were transplanted for hepatocellular carcinoma were studied. Forty-six patients had pretransplantation adjuvant local therapy. The size and number of tumors, the clinical and biological characteristics of the patients were recorded before liver transplantation, and histological analysis was performed on the explanted liver. Overall survival rates at 1 and 3 years were 81% and 66%, respectively. Recurrence-free survival rates at 1 and 3 years were 80% and 65%, respectively. Seven patients had tumor recurrence with 1- and 3-year recurrence rates of 5% and 10%, respectively. Satellite nodules on the explanted liver were the only statistically significant predictor of recurrence (P = .0003). None of the patients who did not have satellite nodules had recurrence. There was a significant correlation between satellite nodules and microvascular invasion. Patients with pretransplantation adjuvant therapy had significantly more tumor necrosis, but did not have less satellite nodules. In conclusion, microscopic satellite nodules are a significant predictive factor of tumor recurrence in patients transplanted for small hepatocellular carcinoma. (Liver Transpl 2004;10:S86–S90.)

There is increasing evidence that, in the long term, liver transplantation is the best therapeutic option for patients with small cirrhosis-associated hepatocellular carcinoma (HCC).1, 2 Patients who meet the widely accepted Milano criteria (i.e., a single nodule < 5 cm or 2 or 3 nodules each < 3 cm and no macroscopic vascular invasion) were shown to have 4-year survival rates as high as 70%.2 Apart from the number and size of the nodules, macroscopic vascular invasion is a major risk factor for recurrence. However, vascular invasion is tightly correlated to the size of the nodules.3 Therefore, it is rarely present in selected patients with small HCC and has a limited impact in this population.

Overall, the results of transplantation in patients with small HCC are encouraging. However, it is important to note that even if strictly selected on the basis of the Milano criteria, patients with small HCC still carry a 10–15% risk of recurrence at 5 years and that, in some of them, recurrence appears shortly after transplantation. Therefore, it can be suspected that tumor factors other than size, number of nodules, and macroscopic vascular invasion may have an influence on the risk of recurrence and that these factors may be helpful for refining selection criteria. The aim of this study was to search for additional tumor-related predictive factors for recurrence in a selected group of patients who all met the Milano criteria.


HCC, hepatocellular carcinoma; US, Doppler-ultrasound; CI, confidence interval.

Patients and Methods

From January 1991 to May 2002, 82 consecutive patients underwent liver transplantation with a diagnosis of cirrhosis-associated HCC. At evaluation, tumor staging included Doppler-ultrasound (US) examination computed tomography scan of the thorax and abdomen and bone scintigraphy in all patients. When needed, MRI was used in addition to computed tomography scan for characterization of small nodules. Thirteen out of these 82 patients who were outside the Milano criteria at evaluation were excluded from the study. The remaining 69 patients were considered for analysis.

There were 60 males and 9 females. Mean age was 52 ± 8 years. The main characteristics of these patients are summarized in Table 1.

Table 1. Characteristics of the 69 Patients in the Study Population
  1. Data are presented as number of patients (with percentages in parentheses) or mean ± SD (with ranges in parentheses).

Child score 
 A35 (51%)
 B29 (42%)
 C5 (7%)
Cause of the cirrhosis 
 HCV infection34 (50%)
 Alcohol16 (23%)
 HBV infection16 (23%)
 Others3 (4%)
Waiting time (months)4.8 ± 4.4 (0.1–29)
Number of nodules (1/2/3)56 (81%) / 11 (16%) / 2 (3%)
Size of the largest nodule (mm)21.5 ± 10.6 (0.5–4.5)
Alpha-fetoprotein level (ng/mL)142 ± 419 (2–9,043)

Pretransplantation adjuvant therapy was decided according to the following policies. From 1991 to 1994, patients with Child's grade A or B cirrhosis received arterial chemoembolization provided they did not have portal vein thrombosis. After 1994, percutaneous ablation was the first-line treatment in patients with less than 3 nodules, unilobar tumors, platelets over 50 × 109/L, prothrombin time over 50% of normal, and no ascites. Percutanous ablation consisted in ethanol injection from 1995 to 1999 and radiofrequency thereafter. Patients who did not meet these criteria were oriented to arterial chemoembolization. No adjuvant therapy was performed when patients met neither the criteria for percutaneous ablation nor those for arterial chemoembolization or when the interval from listing to transplantation was expected to be short. In keeping with these policies, 12 patients received chemoembolization, 28 received percutaneous ablation (ethanol injection in 15, radiofrequency in 11, and both in 2), 5 received sequentially percutaneous ablation and chemoembolization and the remaining 24 patients did not have any adjuvant therapy before transplantation.

Pathological analysis of the explanted liver was performed by the same pathologist (VP) to evaluate macroscopic and microscopic characteristics of the tumors: size, number, capsule formation with or without effraction, macrovascular invasion defined as gross invasion of the main branches of the portal or hepatic veins, tumor cell differentiation according to Edmondson and Steiner,4 microvascular invasion defined as microvascular tumor emboli in the central veins or portal vessels in adjacent liver tissue, satellite micronodules defined as microscopic nodules of hepatocellular carcinoma separated from the tumor by an interval of non-tumoral liver parenchyma, and lymph node metastasis. The characteristics of the tumor according to pretransplantation adjuvant therapy groups are summarized in Table 2.

Table 2. Tumor Characteristics According to Pre-transplantation Therapy at Evaluation and at Transplanted, on the Explanted Liver
 Pre transplantation TherapyTotal (n = 69)P*
None (n = 23)CE (n = 12)PA (n = 29)CE and PA (n = 5)
  • *

    Between therapy groups.

  • Median value, CE, chemoembolization; PA, percutaneous ablation.

Number of nodules      
 At evaluation1 (1–3)1 (1–3)1 (1–3)1 (1–2)1 (1–3)ns
 Explanted liver2 (0–7)1 (1–8)2 (0–5)1 (1–3)2 (1–8)ns
Size of the largest nodules (cm)      
 At evaluation2 (0.8–4.3)2.4 (0.8–4)2 (0.5–4.5)2 (2–4)2.1 (0.7–6)ns
 Explanted liver2 (0–5)2.5 (0.5–6)2.6 (0–5)3.5 (3–4.5)2.6 (0.5–6)ns
Tumor necrosis      
 <10%17 (74%)3 (35%)5 (17%)025 (37%) 
 10-60%4 (17%)25 (17%)3 (10%)1 (20%)33 (47%)<0.001
 >60%2 (9%)7 (58%)21 (73%)4 (80%)34 (49%) 
Satellite nodules9 (39%)6 (50%)11 (38%)2 (40%)28 (40%)ns
Microvascular invasion12 (52%)5 (42%)12 (41%)5 (100%)24 (34%)ns

Quantitative variables are expressed as mean ± standard deviation or median values with range in parentheses, and qualitative variables as absolute numbers with percentages in parentheses. Differences between patient groups were tested by means of Fisher's exact test for qualitative variables and Wilcoxon rank sum test or Kruskal-Wallis test for quantitative variables. Probabilities of overall and relapse-free survival were estimated by the Kaplan-Meier method, with ranges of 95% confidence intervals (CI) given in parentheses. Univariate analysis of survival, recurrence-free survival, and time-to-recurrence was performed using log-rank test to assess predictive factors for recurrence. A P value < .05 was considered statistically significant. Analysis was performed using SPSS 11.0 software (SPSS, Chicago, IL).


Influence of Pretransplantation Adjuvant Therapy on Tumor Status

The number or size of the nodules at evaluation were not significantly different between pretransplantation adjuvant therapy groups (Table 2). On the explanted liver, tumor necrosis more than 60% was significantly more frequent in patients who had pretransplantation therapy as compared with those who did not (P = .0005). However, it must be noted that in the group without pretransplantation therapy, tumor necrosis was found to be between 10 and 60% in 17% of the patients and more than 60% in 8%. Seventy-one percent of the patients who had percutaneous ablation had complete tumor necrosis and none of them had recurrence. The proportion of patients with satellite nodules and micro-vascular invasion was not significantly different between treatment groups.

Survival and Recurrence

The median follow-up for the 69 patients was 30 months (range: 0.1–143). The overall survival rates in the 69 patients at 1 and 3 years were 81% (95% CI: 72–90%) and 66% (95% CI: 54–78%), respectively. The mean recurrence-free survival rate at 1 and 3 years were 80% (95% CI: 70–89%) and 65% (95% CI: 53–77%), respectively. Death was related to the recurrence of HCC in 5 patients. Other causes of death were septic shock in 5, primary liver dysfunction in 4, cerebral stroke in 3, recurrence of initial liver disease (other than HCC) in 2, de novo malignancy in 2, and miscellaneous causes in 4.

HCC-recurrence occurred in 7 of 69 patients (10.1%). The median intervals from transplantation to recurrence and from recurrence to death were 12 months (range: 3–47) and 7 months (range: 0.4–10), respectively. Four patients had liver metastases, and 1 patient had lung and brain metastases. One patient with a previous biopsy had a parietal recurrence probably related to needle tract seeding. The remaining patient had bilateral adrenal metastases, which were surgically resected. This patient was alive and tumor free 59 months after transplantation.

Predictive Factors for Recurrence and Survival

The 1- and 3-year overall recurrence rates were 5 (95% CI: 0–11%) and 10 (95% CI: 3–18%), respectively. Univariate analysis showed that the presence of satellite nodules was the only significant predictive factor for recurrence (Table 3). No recurrence was observed in the group of 41 patients who did not have satellite nodules on the explanted liver. Conversely, among the 28 patients who had satellite nodules, 7 had post-transplantation recurrence (Figure 1). In this series in which all the patients were within the Milano criteria, the size and number of nodules were not significantly associated with recurrence, neither at the time of evaluation nor on the explanted liver. There was no significant correlation between the size or the number of nodules and the presence of satellite nodules. Satellite nodules were present in 9/24 (37%) patients without pretransplantation adjuvant therapy, 11/28 (39%) patients with percutaneous ablation, 6/12 (50%) patients with chemoembolization and 2/5 (40%) with a combination of chemoembolization and percutaneous ablation. These proportions were not significantly different.

Table 3. Predictive Factors for Recurrence of Hepatocellular Carcinoma After Liver Transplantation (Univariate Analysis)
FactorNumber of Patients (%)Recurrence (%)P*
  • *

    Log-rank test.

Satellite nodules   
 Absent41 (59)0/410.001
 Present28 (41)6/28 
Microscopic capsule   
 Absent17 (25)2/170.89
 Present52 (75)5/52 
Microscopic vascular  invasion   
 Absent35 (51)2/350.39
 Present34 (49)4/34 
Pre-LT treatment   
 None23 (33)3/230.18
 Percutaneous29 (42)0/29 
 Chemoembolisation12 (17)2/12 
 Both5 (7)1/5 
Figure 1.

Effect of presence or absence of satellite nodules on the explanted liver on the recurrence rate in 69 patients transplanted for cirrhosis with hepatocellular carcinoma. SN = satellite nodules.

Focal macrovascular invasion was found on the explanted liver in only 3 patients. One had a recurrence 7.5 months after transplantation. Another patient died 24 months after transplantation due to recurrence of hepatitis C. The remaining patient is alive without recurrence 42 months after transplantation. Microvascular invasion was significantly correlated with the presence of satellite nodules (P = .002), but was not a significant predictive factor for recurrence. None of the tumor-related variables which were recorded were significant predictive factors for survival and recurrence-free survival.


In the present series, the tumor-related factors which have been shown in the past to be the more strikingly related to post-transplantation survival in patients with HCC (i.e., the size and number of nodules were not significantly predictive for recurrence).5, 6 This apparent paradox might be explained by the fact that in this series, patients were highly selected and that all met the Milano criteria.1 On the basis of previous reports, there is no doubt that for patients who are outside Milano criteria, the risk for recurrence rapidly increases with the number and size of nodules.5, 7, 8 However, our findings suggest that for those who are within the range of the Milano criteria, these basic variables do not significantly influence the risk of recurrence. Macroscopic vascular invasion, whatever the size and number of nodules, is a determinant risk factor for recurrence. Since in our series all patients had small HCCs, it is not surprising that only 3 of 69 had focal macroscopic vascular invasion on the explanted liver and that this factor did not play a significant role in the whole study population. Otherwise, since 10% of the patients had tumor recurrence, these results support our initial hypothesis that other tumor-related factors are implicated in post-transplantation recurrence in patients who are within Milano criteria.

The most important finding in this study is that, whatever the size and number of nodules, the presence of satellites nodules on the explanted liver was a strong predictor for recurrence. Indeed, none of the 42 patients without satellite nodules had recurrence while, in contrast, 6 of the 28 patients (21%) who had satellite nodules experienced recurrence. There was a strong correlation between microscopic vascular invasion and the presence of satellite nodules. Furthermore, patients who had microscopic vascular invasion had a risk of recurrence almost twice as high as those who did not. Therefore, the absence of statistically significant influence of microscopic vascular invasion in our study population is likely to be due to the small size of the samples, and it can be anticipated that microscopic vascular invasion along with satellite nodules has important prognostic implications. However, this apparent discrepancy between satellite nodules and microscopic vascular invasion suggests that this pathological change, if localized, could be missed even by a detailed examination of the resected liver. Conversely, satellite nodules due to their large size are more unlikely to be unrecognized by pathological examination.

As well as in transplanted patients, satellite nodules have been identified as a prognostic factor after surgical resection for HCC.9 It can be speculated that patients with satellite nodules have more prominent tumor-related vascular changes, develop more frequently some degree of vascular invasion leading to circulating tumor cells which in turn, give rise to distant metastases. Again, this hypothesis is supported by the strong statistical relation that we found between satellite nodules and microscopic vascular invasion. Overall, a clue for identifying more precisely those patients who are at risk for recurrence might be to identify those who have vascular invasion and circulating tumor cells at the time of transplantation. Unfortunately, the results of techniques aimed at detecting circulating tumor cells, some of them based on albumin messenger RNA, have not been validated.9–11 Our results suggest that improvements in imaging techniques aimed at detecting more accurately satellite nodules less than 5 mm in diameter, an indirect but strong marker of vascular invasion, could also help improve the selection of the optimal candidates.

Another important finding in this study is that the incidence of satellite nodules was not significantly reduced by pretransplantation adjuvant therapy. Indeed, although pathological examination of the resected liver showed that tumor necrosis was more frequent in those who underwent arterial chemoembolization and/or percutaneous ablation than in those who did not, these pretransplantation therapies failed to prevent the occurrence of satellite nodules. This finding could explain at least in part that arterial chemoembolization and/or percutaneous ablation did not reduce the incidence of recurrence although they were significantly effective in terms of tumor necrosis. In the future, preventing the occurrence of satellite tumoral proliferation in parallel with tumor destruction should be an determinant goal in the development of novel adjuvant local therapies.

In conclusion, this study confirms that about 10% of the patients who meet Milano criteria still have post-transplantation recurrence and that, in those patients, recurrence does not seem to be related to the size and number of nodules. In these selected patients, microscopic vascular invasion might be a determinant factor for recurrence. Our results suggest that the development of satellite nodules favor vascular invasion, and in turn, recurrence. Therefore, future directions in the development of pretransplantation adjuvant therapies should focus on the prevention of peripheral tumor progression.