The role of perineural invasion in long-term survival following curative resection of pancreatic carcinoma

Abstract

BACKGROUND & AIMS: Most pancreatic carcinomas demonstrate perineural invasion, however its prognostic significance is unclear. The aim of this study is to grade the severity of perineural invasion and correlate this with survival. METHODS: A retrospective analysis of all resected pancreatic carcinomas (n=44) in a UK teaching hospital from 1997-2002. Perineural invasion was graded 0 (absent) to 3 (severe). Factors predictive of hospital mortality and 5-year survical were determined. RESULTS: Hospital mortality (18%), was significantly higher for patients with albumin <35g/L, operating time> 5.25 hours, blood loss > 5 litres, lowest quartile Townsend score and reoperation. Multivariate analysis identified albumin < 35 g/ dL, (odds ratio (OR) 22.5, p=0.043), reoperation (OR 30.3, p=0.029) and lowest quartile Townsend score (OR 39.8, p=0.0l8) as independent predictors of hospital mortality. Actuarial survival at 1,3 and 5-years was 56%, 29% and 24%. Perineural invasion grade 0-1 than grade 2-3: median survival 37 versus 17 months respectively, p=0.007. Five-year survival was significantly less with vascular invasion and tumours > 2cm.Cox proportional hazard analysis identified grade 2-3 perineural as the only significant independent indicator of poor prognosis, hazard ratio 2.S (95% Cl 1.1-7.2) p=0.031, CONCLUSIONS: Grade 2-3 perineural invasion is a strong prognostic indicator. Grading should become standard practice in reporting of pancreatic cancer. Previous studies may have concluded perineural invasion was not a prognostic indicator because invasion was graded present/ absent.