Pancreatic adenocarcinoma: regression analysis to identify improved cytologic criteria

Diagn Cytopathol. 1991;7(4):341-5. doi: 10.1002/dc.2840070404.

Abstract

The incidence of pancreatic adenocarcinoma is increasing and it is usually unresectable at the time of diagnosis. Consequently, fine-needle aspiration biopsy (FNAB) is being used more frequently for diagnosis. The reported sensitivity of diagnosing pancreatic adenocarcinoma by FNAB has varied between 50% and 100%. In an attempt to increase the diagnostic sensitivity, we retrospectively reviewed a series of pancreatic FNABs. Fifteen cytologic criteria were evaluated in 78 patients who had pancreatic FNABs. Of these patients, 49 had primary adenocarcinomas and 29 had benign, non-neoplastic lesions. Using a stepwise logistic regression analysis we identified three key cytologic criteria for this diagnosis. Our study identified anisonucleosis (P = 0.001), large nuclei (P = .007), and nuclear molding (P = .03) as the significant cytologic features for diagnosing pancreatic adenocarcinoma. In combination, these three criteria had a sensitivity of 98% and a specificity of 100%.

MeSH terms

  • Adenocarcinoma / pathology*
  • Adenocarcinoma / ultrastructure
  • Biopsy, Needle
  • Cell Nucleus / pathology
  • Humans
  • Pancreatic Neoplasms / pathology*
  • Pancreatic Neoplasms / ultrastructure
  • Predictive Value of Tests
  • Regression Analysis
  • Retrospective Studies