Satellite lesions in patients with small hepatocellular carcinoma with reference to clinicopathologic features
Abstract
BACKGROUND
It is not rare to find satellite lesions in patients with small hepatocellular carcinoma (HCC). The purpose of this study was to elucidate the factors associated with satellite lesions in these patients.
METHODS
We investigated the prevalence of satellite lesions, the relationship of clinicopathologic factors to satellite lesions, and the distance from the main tumor to the satellite lesion in 149 patients. Patients, who had a solitary HCC of 3.0 cm or less in diameter but no satellite lesions on preoperative imaging procedures, underwent potentially curative resection. The main tumors were macroscopically classified into four groups: early HCC, a vaguely nodular type showing preservation of the preexisting liver structure; single nodular type; single nodular type with extranodular growth; and confluent multinodular type.
RESULTS
Of 149 resected specimens, 28 (19%) showed satellite lesions. Of the clinicopathologic factors investigated, the macroscopic type and tumor differentiation were significantly associated with the prevalence of satellite lesions. Both the single nodular type with extranodular growth and the confluent multinodular type showed satellite lesions more frequently than the early HCC and the single nodular type. A significantly higher prevalence of satellite lesions was observed in poorly differentiated HCC than in well and moderately differentiated HCC. The satellite lesions were located 0.5 cm or less from the main tumor in 8 (33%) specimens, 0.6–1.0 cm in 12 (50%), and 1.1–2.0 cm in 4 (17%). No identifiable factors were significantly related to the distance from the main tumor to the satellite lesion. However, all satellite lesions located more than 1.0 cm from the main tumor coexisted with poorly differentiated HCC, which were the single nodular type with extranodular growth or the confluent multinodular type.
CONCLUSION
In the single nodular type with extranodular growth, confluent multinodular type, and poorly differentiated HCC, extensive treatment achieving a large safety margin and/or frequent posttreatment follow-up examinations may be needed because of the high prevalence of satellite lesions. Cancer 2002;95:1931–7. © 2002 American Cancer Society.
DOI 10.1002/cncr.10892
Hepatocellular carcinoma (HCC) is one of the most common malignant tumors in Japan. The screening of high-risk populations for HCC using ultrasonography and serum α-fetoprotein (AFP) levels has recently increased the number of candidates for effective local treatments such as hepatic resection.1 However, the majority of patients with small HCC cannot undergo surgical treatment because of associated cirrhosis. For these patients, several local ablation methods, such as percutaneous ethanol injection (PEI), microwave coagulation therapy (MCT), and radiofrequency ablation therapy (RFA), have been developed as minimally invasive therapies and are accepted as alternatives to surgery in patients with small HCC.2-4
However, the prognosis of patients with small HCC is still unsatisfactory because of frequent recurrence even after complete resection or complete destruction of the tumor. The high recurrence rate may be due, at least in part, to untreated satellite lesions, which are too small to be detected in imaging methods before treatment. It is important to analyze the factors related to satellite lesions to identify patients at high risk and to detect recurrent HCC nodules early enough to reapply effective treatment. The current study was conducted using 149 resected specimens with solitary small HCC nodules to analyze satellite lesions overlooked in pretreatment evaluation with reference to clinicopathologic features in patients with small HCC.
MATERIALS AND METHODS
The subjects were 170 patients with a solitary HCC of less than or equal to 3.0 cm in diameter who underwent potentially curative resection from January 1992 to October 1999 at the National Cancer Center Hospital. Of these 170 patients, 21 patients were excluded from this study because satellite lesions and/or portal vein invasion were noted during preoperative evaluation, including ultrasonography (US), computed tomography (CT) scan, and angiography. In the remaining 149 patients, we assessed satellite lesions in resected specimens (Table 1). Of the 149 patients, 85 (57%) patients underwent preoperative transcatheter arterial embolization (TAE). The main surgical procedures applied were partial resection in 92 (62%), subsegmentectomy in 29 (19%), segmentectomy in 27 (18%), and lobectomy in 1 (1%).
| Characteristics | Total no. of patients | No. of patients with satellite lesions (%) | P value |
|---|---|---|---|
| Total | 149 | 28 (19) | |
| Age (yrs) | |||
| 21–60 | 62 | 12 (19) | |
| 61–85 | 87 | 16 (18) | 0.88 |
| Gender | |||
| Men | 121 | 20 (17) | |
| Women | 28 | 8 (29) | 0.14 |
| HBs Ag | |||
| Positive | 20 | 4 (20) | |
| Negative | 129 | 24 (19) | 0.99 |
| HCV Ab | |||
| Positive | 114 | 19 (17) | |
| Negative | 35 | 9 (26) | 0.23 |
| Alcohol abusea | |||
| Positive | 68 | 10 (15) | |
| Negative | 81 | 18 (22) | 0.24 |
| Total bilirubin (mg/dL) | |||
| 0.3–0.9 | 84 | 16 (19) | |
| 1.0–1.7 | 65 | 12 (18) | 0.93 |
| Albumin (g/dL) | |||
| 2.6–3.8 | 77 | 14 (18) | |
| 3.9–4.4 | 72 | 14 (19) | 0.84 |
| ICG-15b | |||
| 1–20 | 65 | 11 (17) | |
| 21–57 | 65 | 15 (23) | 0.38 |
| Cirrhosis | |||
| Positive | 47 | 10 (21) | |
| Negative | 102 | 18 (18) | 0.60 |
| Tumor size (cm) | |||
| 0.8–2.1 | 75 | 13 (17) | |
| 2.2–3.0 | 74 | 15 (20) | 0.65 |
| Tumor encapsulation | |||
| Positive | 129 | 27 (21) | |
| Negative | 20 | 1 (5) | 0.12 |
| Macroscopic type of tumor | |||
| Early HCC | 18 | 0 (0) | |
| Single nodular type | 59 | 3 (5) | |
| Single nodular type with extranodular growth | 38 | 13 (34) | |
| Confluent multinodular type | 34 | 12 (35) | < 0.01 |
| Status of differentiationc | |||
| Well | 39 | 3 (8) | |
| Moderate | 71 | 13 (18) | |
| Poor | 26 | 12 (46) | < 0.01 |
| α-fetoprotein (ng/mL) | |||
| 1–20 | 74 | 11 (15) | |
| 21–19,930 | 73 | 17 (23) | 0.28 |
- HBsAg: hepatitis B surface antigen; HCV Ab: hepatitis C virus antibody; HCC: hepatocellular carcinoma.
- a Ethanol intake ≥ 80 g per day for ≥ 5 years.
- b Indocyanine green retention at 15 minutes.
- c Evaluation of the histologic grade was not performed for 13 tumors because of extensive necrosis after preoperative transcatheter arterial embolization.
The resected specimens were serially sectioned at 10-mm intervals and examined macroscopically. The criteria used to identify satellite lesions were essentially those indicated by the Liver Cancer Study Group of Japan,5 i.e., tumors surrounding the main tumor with multiple other satellite nodules or small solitary tumors located near the main tumor that are histologically similar or less differentiated than the main tumor. The satellite lesions in this study did not include multicentric nodules. Portal vein invasion in the parenchyma was included as satellite lesions in this study because it is not always easy to make a distinction between them and because portal vein invasion can also be the origin of satellite lesions. The main tumors were macroscopically classified into four groups according to the following criteria (Fig. 1): early HCC, HCC of a vaguely nodular type showing preservation of the preexisting liver structure; single nodular type; single nodular type with extranodular growth; and confluent multinodular type.5, 6 For light microscopic examination, paraffin-embedded sections were stained with hematoxylin and eosin. The histologic grade of tumor differentiation was assigned according to the classification of the Liver Cancer Study Group of Japan.5 When two of more histologic patterns were shown in the same tumor, the predominant pattern was described. In 13 tumors with extensive necrosis after preoperative TAE, evaluation of the histologic grade was not performed.
Macroscopic type of main tumor. (A) Early hepatocellular carcinoma (HCC): HCC of a vaguely nodular type showing preservation of preexisting liver structure. (B) Single nodular type. (C) Single nodular type with extranodular growth. (D) Confluent multinodular type.
The relationship beyween clinicopathologic factors and satellite lesions was also investigated. The factors were classified as host or tumor related. The host-related factors were age, gender, hepatitis B surface antigen, hepatitis C virus antibody, alcohol abuse (ethanol intake ≧ 80 g per day for ≧ 5 years), serum total bilirubin level, serum albumin level, indocyanine green retention at 15 minutes, and associated cirrhosis. The tumor-related factors were tumor size, tumor encapsulation, macroscopic type, status of differentiation, and serum AFP level. In patients who underwent preoperative TAE, their values before TAE were adopted for the assessments of tumor size and serum AFP level. Each quantitative factor was divided into two groups by the median. We also assessed the distance between the main tumor and the satellite lesion. The distance was defined as the distance from the margin of the main tumor to the side of the satellite lesion facing away from the main tumor. When two or more satellite lesions were detected in the same specimen, the satellite lesion most distant from the main tumor was selected. In addition, we evaluated the differences in the distance from the main tumor to the satellite lesion with respect to the clinicopathologic factors.
Frequencies in 2 × 2 and larger contingency tables were compared with the chi-square or Fisher exact test. Distributions of continuous variables were compared with the Mann–Whitney or the Kruskal–Wallis test. All P values in this study were two tailed. Significance was defined as a P value of 0.05 or less. Statistical analyses were performed with Stat View version 5.0 (SAS Institute, Cary, NC).
RESULTS
Of the 149 resected specimens, 28 (19%) showed satellite lesions. Of the 28 specimens, the maximum diameter of the satellite lesion was 0.1 cm in 6 (20%), 0.2 cm in 8 (30%), 0.3 cm in 5 (18%), 0.4 cm in 4 (14%), and 0.5 cm in 5 (18%). The mean was 0.28 cm, the standard deviation was 0.14 cm, and the median was 0.25 cm. The number of satellite lesions per specimen was 1 in 6 (20%), 2 in 7 (25%), 3 in 9 (33%), 5 in 2 (7%), 7 in 1 (4%), and more than 10 in 3 (11%). Table 1 shows the relationship of the clinicopathologic factors to the satellite lesions. Of the factors investigated, the macroscopic type of tumor and the histologic grade of tumor differentiation were significantly associated with the prevalence of satellite lesions. Both the single nodular type with extranodular growth (34%) and the confluent multinodular type (35%) showed satellite lesions more frequently than early HCC (0%; P < 0.01 for the single nodular type with extranodular growth, P < 0.01 for the confluent multinodular type) and the single nodular type (5%; P < 0.01 for the single nodular type with extranodular growth, P < 0.01 for the confluent multinodular type). A significantly higher prevalence of satellite lesions was observed in poorly differentiated HCC (46%) than in well (8%; P < 0.01) and moderately differentiated HCC (18%; P < 0.01). The prevalence did not differ between the larger (22–30 mm: 20%) and the smaller tumors (21 mm or less: 17%; P = 0.65). Even in patients with smaller HCC, a significantly higher prevalence of satellite lesions was observed in poorly differentiated HCC (5 of 11, 45%) than in well (2 of 21, 10%; P < 0.05) and moderately differentiated HCC (6 of 36, 17%; P < 0.05). The single nodular type with extranodular growth (8 of 23, 35%) showed a significantly higher prevalence of satellite lesions and the confluent multinodular type (4 of 15, 27%) had satellite lesions more frequently, in comparison to early HCC (0 of 12, 0%; P < 0.05 for the single nodular type with extranodular growth, P = 0.10 for the confluent multinodular type) and the single nodular type (1 of 25, 4%; P < 0.01 for the single nodular type with extranodular growth, P = 0.06 for the confluent multinodular type).
In 4 of the 28 specimens, the distance from the main tumor to the satellite lesion could not be measured because the satellite lesion was sectioned separately from the main tumor. In the other 24 specimens, the satellite lesions were located 0.5 cm or less from the main tumor in 8 (33%), 0.6–1.0 cm in 12 (50%), 1.1–1.5 cm in 1 (4%), and 1.6–2.0 cm in 3 (13%). Although no identifiable clinicopathologic factors were significantly related to the distance from the main tumor to the satellite lesion, all satellite lesions located more than 1.0 cm from the main tumor coexisted with poorly differentiated HCC, which were the single nodular type with extranodular growth or the confluent multinodular type. The satellite lesions in well and moderately differentiated HCC and/or single nodular type were located 1.0 cm or less from the main tumor (Table 2).
| Characteristics | Distance from the main tumor to SL (cm) | Total | P value | |||
|---|---|---|---|---|---|---|
| 0.1–0.5 | 0.6–1.0 | 1.1–1.5 | 1.6–2.0 | |||
| Age (yrs) | ||||||
| −60 | 3 | 4 | 1 | 1 | 9 | |
| 61- | 5 | 8 | 0 | 2 | 15 | 0.62 |
| Gender | ||||||
| Men | 5 | 8 | 1 | 2 | 16 | |
| Women | 3 | 4 | 0 | 1 | 8 | 0.91 |
| HBs Ag | ||||||
| Positive | 1 | 1 | 0 | 1 | 3 | |
| Negative | 7 | 11 | 1 | 2 | 21 | 0.76 |
| HCV Ab | ||||||
| Positive | 6 | 9 | 0 | 2 | 17 | |
| Negative | 2 | 3 | 1 | 1 | 7 | 0.45 |
| Alcohol abusea | ||||||
| Positive | 3 | 4 | 0 | 2 | 9 | |
| Negative | 5 | 8 | 1 | 1 | 15 | 0.62 |
| Total bilirubin (mg/dL) | ||||||
| −0.9 | 5 | 8 | 1 | 0 | 14 | |
| 1.0- | 3 | 4 | 0 | 3 | 10 | 0.15 |
| Albumin (g/dL) | ||||||
| −3.8 | 4 | 5 | 1 | 1 | 11 | |
| 3.9- | 4 | 7 | 0 | 2 | 13 | 0.68 |
| ICG-15b | ||||||
| −20 | 4 | 4 | 1 | 0 | 9 | |
| 21- | 4 | 7 | 0 | 3 | 14 | 0.27 |
| Cirrhosis | ||||||
| Positive | 4 | 2 | 0 | 2 | 8 | |
| Negative | 4 | 10 | 1 | 1 | 16 | 0.21 |
| Tumor size (cm) | ||||||
| −2.1 | 5 | 5 | 1 | 2 | 13 | |
| 2.2- | 3 | 7 | 0 | 1 | 11 | 0.57 |
| Tumor encapsulation | ||||||
| Positive | 7 | 12 | 1 | 3 | 23 | |
| Negative | 1 | 0 | 0 | 0 | 1 | 0.55 |
| Macroscopic type of tumor | ||||||
| Early HCC | 0 | 0 | 0 | 0 | 0 | |
| Single nodular type | 1 | 2 | 0 | 0 | 3 | |
| Single nodular type with extranodular growth | 4 | 5 | 0 | 2 | 11 | 0.88 |
| Confluent multinodular type | 3 | 5 | 1 | 1 | 10 | |
| Status of differentiation | ||||||
| Well | 1 | 2 | 0 | 0 | 3 | |
| Moderate | 5 | 7 | 0 | 0 | 12 | |
| Poor | 2 | 3 | 1 | 3 | 9 | 0.23 |
| α-fetoprotein (ng/mL) | ||||||
| −20.9 | 2 | 5 | 1 | 1 | 9 | |
| 21.0- | 6 | 7 | 0 | 2 | 15 | 0.51 |
- SL: satellite lesion; HBsAg: hepatitis B surface antigen; HCV Ab: hepatitis C virus antibody; HCC: hepatocellular carcinoma.
- a Ethanol intake ≥ 80 g per day for ≥ 5 years.
- b Indocyanine green retention at 15 minutes.
DISCUSSION
Many treatment modalities such as surgical hepatic resection, PEI, MCT, and RFA, have been developed as effective radical treatments for patients with small HCC.4 PEI has been the most widely performed and is now accepted as an alternative to surgery in patients with small HCC.7, 8 However, it has several limitations, including the uncertainty of tumor ablation and its unsuitability for large tumors. To overcome these limitations, several other options for local ablation therapy, such as MCT and RFA, have been developed9-11 and introduced clinically as more effective treatment modalities to achieve tumor necrosis in fewer treatment sessions than PEI.12-14 TAE, another treatment option in use, has a marked antitumor effect especially for expanding encapsulated small HCC, but its effect is poor for patients with hypovascular or infiltrative HCC.15
However, the long-term prognosis is still disappointing because recurrence occurs frequently even after complete local control of the tumor is obtained. Satellite lesions, which derive from the main tumor via the portal system, are one of the major causes of HCC recurrence. Despite recent progress in imaging diagnosis, no currently available imaging technique is adequately sensitive to detect small satellite lesions before treatment. CT scan during arterial portography, which is one of the most sensitive techniques available for depicting small nodules, still has several limitations including the frequent appearance of nodular psuedolesions that mimic HCC.16-18
It is not rare to find satellite lesions in patients with small HCC. A previous pathologic study showed that 37.7% of small HCC of less than or equal to 3.0 cm in diameter had satellite lesions and/or portal vein invasion including nodules large enough to be identified before treatment.6 The current study revealed that 19% of HCC nodules of 3.0 cm or less in diameter had satellite lesions that were not detected during pretreatment evaluation. In the clinicopathologic factors investigated, the macroscopic type of the main tumor was one of the factors significantly associated with the prevalence of satellite lesions. Both the single nodular type with extranodular growth and the confluent multinodular type showed satellite lesions more frequently than early HCC and the single nodular type. A review of 240 autopsy series of patients with HCC revealed that both the single nodular type with extranodular growth and the confluent multinodular type showed aggressive tumor spread, i.e., a high prevalence of satellite lesions, portal vein invasion, and lymphogenous and hematogenous metastasis.19 Our study indicated that aggressive tumor spread in the single nodular type with extranodular growth and in the confluent multinodular type was observed even in patients with small HCC.
The findings in this study suggested that a detailed pretreatment imaging diagnosis of the macroscopic type may be helpful in predicting satellite lesions in patients with HCC. However, to our knowledge, the role of imaging modalities in the classification of the macroscopic type in patients with small HCC has not yet been fully evaluated. Conventional CT scans and magnetic resonance imaging (MRI) are insufficiently sensitive, whereas conventional B-mode US provided a correct diagnosis for macroscopic type in 61% of patients with small HCCs.20-22 Recent advances in technology have facilitated morphologic characterization of small tumors by imaging modalities, including US with a microbubble contrast agent, dynamic thin-sliced CT scans, and fast three-dimensional gadolinium-enhanced MRIs. These diagnostic approaches improve imaging quality and contribute positively to the characterization of the hepatic nodule.23-27
The histologic grade of tumor differentiation was another factor significantly related to the prevalence of satellite lesions. Several studies reported that a low degree of tumor cell differentiation was a risk factor for HCC recurrence after hepatic resection or PEI.28-30 These findings were supported by the findings of this study, in which a higher prevalence of satellite lesions was noted in patients with small HCC with a lower grade of tumor differentiation. Tumor biopsy before ablation therapy may be beneficial to predict satellite lesions not detected during the pretreatment evaluation. Tumor biopsy using a 21-gauge needle has been performed routinely in our hospital before ablation therapy for patients with small HCC except for subcapsular HCC.31 This procedure is safe with a low risk of tumor seeding, although direct puncture of subcapsular HCC or the use of a large needle may cause hemorrhage and tumor dissemination.31-38
Tumor size was not associated with the prevalence of satellite lesions in patients with small HCC. In HCC of 2.1 cm or less in diameter, the macroscopic type and tumor differentiation were significantly associated with the prevalence of satellite lesions. In the single nodular type with extranodular growth, the confluent multinodular type, and in poorly differentiated HCC, extensive treatment and/or frequent posttreatment examinations may be needed even though the tumor is relatively small.
In this study, the satellite lesions were located 2.0 cm or less from the main tumor, but no identifiable factors were related to the distance from the main tumor to the satellite lesion. Local ablation therapy has cell-killing effects on the main tumor and on the adjacent surrounding tissue including the satellite lesion.39, 40 Heat ablation therapy is more effective than the classic method (e.g., PEI) to achieve a wide safety margin.41-44 The multimodality treatment approach combined with ablation therapy and transcatheter arterial balloon occlusion/embolization has the potential to enlarge the volume to be treated.45-50 Extensive therapies or hepatectomy with a wide resection margin may be applied to the single nodular type with extranodular growth, the confluent multinodular type, and poorly differentiated HCC, all of which have a high prevalence of satellite lesions located far from the main tumor. However, there is a possibility that some satellite lesions will survive even after initial treatment appeared to be successful. To detect the untreated nodules early enough to reapply effective treatment, frequent follow-up examinations after treatment may be needed especially in the single nodular type with extranodular growth, the confluent multinodular type, and poorly differentiated HCC.
In conclusion, this study demonstrated that the macroscopic type of tumor and tumor differentiation were significantly associated with the prevalence of satellite lesions. Both the single nodular type with extranodular growth and the confluent multinodular type showed satellite lesions more frequently than early HCC and the single nodular type. In addition, a significantly higher prevalence was observed in poorly differentiated HCC than in well and moderately differentiated HCC. The findings in the current study may be helpful in determining treatment strategies and follow-up examination schedules after treatment of patients with small HCC.
Acknowledgements
The authors acknowledge the helpful suggestions of Professor Keigo Yasuda. They thank Ms. Yuriko Kawaguchi for help with the preparation of the manuscript.
