Relevance of suspicious lymph nodes in preoperative imaging for resectability, recurrence and survival of intrahepatic cholangiocarcinoma

Background Intrahepatic cholangiocarcinoma (ICC) is often diagnosed at an advanced stage resulting in a low resectability rate. Even after potentially curative resection the risk for tumor recurrence is high. Although the extent and value of lymphadenectomy is part of ongoing discussion, the role of preoperative imaging for assessment of suspicious lymph nodes (suspLN) has only been studied modestly. Aim of this study is to demonstrate the influence of suspicious lymph nodes in preoperative imaging on resectability, recurrence, and long-term outcome. Methods All patients who underwent exploration for ICC between January 2008 and June 2018 were included. Preoperative imaging (CT or MRI) was analysed with focus on suspLN at the hepatoduodenal ligament, lesser curvature, interaortocaval, and superior to the diaphragm; suspLN were classified according to the universally accepted RECIST 1.1 criteria; histopathology served as gold standard. Results Out of 187 patients resection was performed in 137 (73.3%), in 50 patients the procedure was terminated after exploration. Overall, suspLN were found preoperatively in 73/187 patients (39%). Comparing patients who underwent resection and exploration only, suspLN were significantly more common in the exploration group (p = 0.011). Regarding lymph node stations, significant differences could be shown regarding resectability: All tumors with suspLN superior to the diaphragm were irresectable. Preoperative imaging assessment showed a strong correlation with final histopathology, especially of suspLN of the hepatoduodenal ligament and the lesser curvature. Sensitivity of suspLN was 71.1%, specificity 90.8%. Appearance of tumor recurrence was not affected by suspLN (p = 0.289). Using a short-axis cut-off of <> 1 cm, suspLN had significant influence on recurrence-free survival (RFS, p = 0.009) with consecutive 1-, 3-, and 5-year RFS of 41, 21, and 15% versus 29, 0, and 0%, respectively. Similarly, 1-, 3- and 5-year overall survival (OS) was 75, 30, and 18% versus 59, 18, and 6%, respectively (p = 0.040). Conclusion Suspicious lymph nodes in preoperative imaging are predictor for unresectability and worse survival. Explorative laparoscopy should be considered, if distant suspicious lymph nodes are detected in preoperative imaging. Nevertheless, given a sensitivity of only 71.1%, detection of suspicious lymph nodes in the preoperative imaging alone is not sufficient to allow for a clear-cut decision against a surgical approach.