Infiltrated margins of resection (R1) and lymph node invasion are dominant negative predictors of survival in patients with a resectable perihilar cholangiocarcinoma. Less clear is whether survival predictors stratify differently between R0 (tumor-free margins) and R1 patients and whether the prognosis of the latter patients is influenced by the pattern of neoplastic infiltration (ie, radial versus longitudinal infiltration). We retrospectively evaluated a series of reported resected perihilar cholangiocarcinoma to obtain insights on the predictive power of these histologic features.
The study includes 264 patients with perihilar cholangiocarcinoma treated between 2004 and 2019 in our center and followed up for >18 months. There were 176 patients with R0 (66.6%) and 88 patients with R1 (33.3%), 31 with radial infiltration only, 30 with longitudinal infiltration only, and 27 with both infiltration patterns. In all patients, the criteria for resection was the absence of metastatic involvement (ie, distant organ metastases, liver metastases, and lymph node metastases beyond the hepatoduodenal ligament). Histopathologic specimens of the resected tumors were centrally reviewed by a pathologist unaware of the clinical outcomes.
Three- and 5-year long-term survival were significantly better in R0 (respectively) compared to R1 patients (55% and 42% vs 42% and 18%, respectively, P < .05). In R1 patients with radial infiltration only and those with radial + longitudinal infiltration, both disease-free and overall survival were worse than those with longitudinal infiltration only (median disease-free survival of 18 and 23 months, respectively, P < .05, median overall survival of 33 and 39 months, respectively, P < .05). At multivariable analysis, nodal status, side of hepatectomy, grading, and presence of radial margin infiltration were associated with long-term outcome.
Radial infiltration of resection margins enhances the negative prognostic value of R1 margins in perihilar cholangiocarcinoma patients and should specifically be accounted for in the prediction of the outcome of adjuvant therapy.

Copyright © 2023 Elsevier Inc. All rights reserved.