Patterns of Failure and Need for Biliary Intervention among Patients with Biliary Tract Cancers Treated with Definitive Surgery and Adjuvant Chemoradiation

Abstract

Adjuvant chemoradiation therapy (CRT) after surgical resection of biliary tract cancers (BTC) has been shown to improve patient outcomes. However, there are limited data regarding patterns of failure after adjuvant chemotherapy and CRT with modern treatment regimens. In this study, we sought to characterize patterns of failure and need for biliary intervention after adjuvant CRT for gallbladder carcinoma (GBC) and cholangiocarcinoma (CCA). We present a retrospective analysis of 80 patients with GBC (n=29), extrahepatic CCA (n=30), perihilar CCA (n=13), or intrahepatic CCA (n=8). All patients underwent curative-intent resection with adjuvant CRT at a single institution from 2007–2017. Median CRT dose was 50.4 Gy (range: 36–65 Gy), which was administered with concurrent 5-FU-based chemotherapy. All but two patients received adjuvant chemotherapy. Tumor genomic mutation profiling was available for 30 patients (38%). Kaplan-Meier method was used to calculate rates of overall survival (OS) and locoregional failure (LRF), and log rank method was used to compare groups. Biliary interventions included ERCP with and without biliary stenting and percutaneous biliary tube placement. Study variables included age, performance status, site of primary disease, margin status, stage, receipt of oncologic resection, lymph node (LN) positivity, presence of lymphovascular invasion (LVI), tumor differentiation, and KRAS and TP53 mutational status. The median follow-up time was 30.5 months, and the median OS was 35.0 months. Among the entire cohort, 34% had T1/T2 disease, 40% were LN positive, and 53% underwent margin-negative (R0) resection. The overall rates of LRF and OS at 2 years were 21% and 62%, respectively. LRF was the initial site of failure in 16/80 patients (20%), and 19 patients (24%) had any LRF. On univariate analysis, only LN positivity (p=0.02) was positively associated with LRF. Site of primary disease (GBC vs. CCA), stage, LVI, and tumor mutation profile were not associated with LRF. Margin status was also not associated with LRF (2-year LRF: 18% vs. 29% in R0 vs. margin-positive (R1) resections, respectively, p=0.76). Among all patients, 23 (29%) required subsequent biliary intervention. Biliary intervention was required in 63% (12/19) of patients with LRF, compared to 18% (11/61) of patients without LRF. Of the 12 patients with LRF who required biliary intervention, 4 (33%) went on to die from biliary complications. Despite adjuvant CRT, there are high rates of LRF among patients with BTC, including GBC. Lymph node positivity may increase the risk of LRF. R1 resection was not associated with higher rates of LRF, suggesting that CRT may partially mitigate the risk of LRF in this high-risk group. LRF results in high rates of subsequent biliary interventions and can yield significant morbidity. Novel strategies should be considered to further improve upon rates of LRF.