br Methods br Results br
Discussion ICC represents an aggressive malignancy arising from the intrahepatic biliary tree. Recent years have seen a dramatic increase in the incidence of ICC, making it the second most common primary liver cancer. Prognosis of ICC remains poor, with 5-year OS rates ranging from 14% to 40%. This study was a review of OS in ICC patients after surgery in a Taiwan medical center over a period of 10 years. Our data showing 5-year OS rates of 45.6% are similar to results from recent studies. Surgical resection of the liver is the only curative treatment for patients with ICC that achieves long-term survival. Accurate information on prognostication is important for decision making and counseling of patients. Widely used prognostication systems, such as the AJCC TNM classification, include a limited number of tumor-related variables, and lack flexibility in terms of allowing physicians to tailor prognostication for specific patients. We therefore reviewed the known prognostic factors of ICC and evaluated correlations between demographic and clinicopathological characteristics and survival (Table 1). As this study was a retrospective analysis, missing data did exist in some variables. For a more accurate analysis, we excluded CEA levels and CA19-9 levels with missing data in >10% of cases for our multivariable analysis. We used univariable analysis to show that nonreceptor tyrosine kinase node metastases, hepatolithiasis, liver abscesses, high CEA and CA19-9 levels, and large tumor diameters were adverse preoperative prognostic factors. Similarly, positive resection margins, periductal infiltration, poor differentiation in tumors, vascular invasion, and perineural invasion were adverse postoperative prognostic factors (Table 2). It is especially notable that hepatolithiasis and liver abscesses adversely influenced survival of ICC. Su et al reported that patients with hepatolithiasis-associated cholangicarcinoma including ICC had a significantly worse survival compared to patients with only cholangiocarcinoma. In a study of 66 patients with hepatolithiasis-associated ICC, radical resection was possible in only 38 patients. Liver abscesses may mask ICC leading to delayed diagnosis. Our multivariable analysis identified lymph node metastases, positive resection margins, periductal infiltration, and poor differentiation in tumors as independent prognostic factors of ICC. These results were consistent with the recent studies that evaluated prognosis of resectable ICC. Although several clinicopathologic factors have been reported to influence survival after resection for ICC, nodal status may be the most strongly predictive. In this study, the presence of lymph node metastases was an independent negative predictor of OS in multivariate analysis. Despite several national guidelines advocating the removal of clinically suspicious lymph nodes, there are wide practice variations regarding routine LND among patients without clinically suspicious lymph nodes. In our study group, there was no significant difference in survival between Nx (no LND) and N0 (no lymph node metastases). We therefore performed a further analysis of demographic and clinicopathologic characteristics of patients with or without LND (Table 3). Patients in the D1 (with LND) group tended to have more late T stages (T3–T4) compared to patients in the D0 (without LND) group. Furthermore, among the patients in the D1 group, lymph node metastases were associated with late T stages, multiple tumors, and elevated serum CEA and CA19-9 levels (Table 4). Marubashi et al found that patients with solitary lesions less than 5 cm in diameter and peripheral-type ICC showed a very low probability of lymph node metastasis. Similarly, Miwa et al also suggested that patients with tumors less than 4.5 cm in diameter located in the peripheral liver had less lymph node metastases. Moreover, there was a higher incidence of vascular and perineural invasion in the D1 group compared to the D0 group. A retrospective study in a tertiary institution revealed that a significantly greater proportion of patients with lymph node metastases had lymphovascular or perineural invasion. This study result suggested the finding that LND of liver hilum may not lead to survival benefits, but was useful for nodal staging, which is an essential prognostic factor of ICC. In addition, later T stages, multiple tumors, and high preoperative serum tumor marker levels were associated with lymph node metastases. Those might be preoperative indicators for LND.