Evaluation of the utility of prognostic models for patients with resected hilar cholangiocarcinoma

Abstract

BackgroundSeveral prognostic systems have been proposed to guide management strategies post-resection for patients with hilar cholangiocarcinoma. The objective of this study was to evaluate the efficacy of these conventional prognostic models, with respect to Overall Survival (OS), on patients in a modern single-centre resectional cohort. MethodPatients diagnosed with hilar cholangiocarcinoma, referred to a supra-regional tertiary referral centre between February 2009 and February 2016, were retrospectively analysed from a prospectively held database linked to Hospital Episode Statistics and Somerset Cancer Registry data. ResultsTwo-hundred and one patients were assessed for suitability for surgery. Eighty-three (41%) patients considered to have potentially resectable disease underwent surgical assessment of resectability. Fifty-six (68%) patients proceeded to resection. Multivariate analysis demonstrated that pre-operative Serum CA 19-9 (p = 0.007), Radiological Arterial Involvement (p = 0.005) and Amsterdam Medical Centre (AMC) prognostic model score (p = 0.032) retained significance in association with OS. Multivariate models developed from this cohort out-performed the conventional prognostic systems for OS. ConclusionThe cohort-derived multivariate models demonstrated significantly improved prognostic capability compared to conventional systems in explaining OS.