Does resection improve overall survival for intrahepatic cholangiocarcinoma with multiple nodal metastases?

Abstract

Background: Intrahepatic cholangiocarcinoma (ICC) often presents with advanced locoregional or metastatic disease. The potential benefit of surgical resection in patients with pretreatment evidence of multifocal regional node involvement in the absence of distant metastatic disease has not been definitively determined. We sought to evaluate the benefit of resection in this population and the relationship of the lymph node ratio (LNR): positive nodes to total nodes examined, with survival. Methods: The National Cancer Database was queried to identify patients with histologically confirmed primary ICC and clinical evidence of N1 lymph node involvement. Patients with metastatic disease, those missing survival data and those who underwent procedures with non-curative intent were excluded. Patients were stratified based on treatment received, with non-operative management including systemic chemotherapy acting as reference. Resected patients were further stratified into LNR quartiles (1: <0.10; 2: 0.10–0.39; 3: 0.40–0.79; 4: ≥0.80). The relationship between treatment type and overall survival (OS) was determined using Cox proportional hazards models adjusting for variables determined a priori including: age, sex, race, Charlson comorbidity score (CCI), tumor size and treating cancer facility characteristics. Unadjusted survival data were estimated using the Kaplan-Meier (KM) method and compared using the log-rank test. Results: 1,292 patients with T1-T3 N1 M0 ICC were identified. 226 (17.5%) underwent surgical resection. Those who underwent resection were more likely to be younger (age 60 vs 64, p 0.1). Among patients undergoing resection, the median number of nodes sampled was 4 (IQR: 2,7) and median number of nodes positive was 2 (IQR: 1,4). The mean LNR was 0.4. KM analysis showed a median survival of 3.0 months with no therapy, 11.6 months with systemic chemotherapy, 16.7 months with resection when LNR ≥ 0.4, and 22.2 months with resection when LNR < 0.04 (log rank p < 0.05). On adjusted Cox modeling, patients undergoing no therapy had a 2-fold increased risk of death relative to those undergoing systemic chemotherapy (HR 2.01, p < 0.01). On Cox modeling stratified by margin status and LNR quartile, R0 resection afforded an OS benefit regardless of LNR quartile (HRs 0.39–0.52, p < 0.03) whereas only patients with the lowest LNR by quartile demonstrated a survival benefit with an R1 resection (HR 0.24, p < 0.01) (Figure 1). Conclusion: For patients with pretreatment evidence of locoregionally advanced ICC, R0 resection offers a benefit in risk-adjusted overall survival regardless of the degree of regional lymph node involvement. The survival benefit associated with resection is not evident in patients who undergo a margin positive resection and have high LNRs. These findings suggest that surgical resection is warranted in ICC with multifocal node involvement when pretreatment studies support feasibility of a margin negative resection.