Malignant biliary stenosis: conventional cytology versus DNA image cytometry

Surg Endosc. 2011 Jun;25(6):1808-13. doi: 10.1007/s00464-010-1469-0. Epub 2010 Dec 18.


Background: The aim of this study is to evaluate the utility of image cytometry (ICM)-DNA analysis on cytological brush specimens in improving the sensitivity and diagnostic accuracy for biliary neoplasias.

Methods: A total of 71 patients with 89 samples of biliary tree brushing from a stenosis were included in this prospective study. Conventional cytology (CC) and DNA ploidy using ICM of the brushing were performed. Benign or malignant findings were confirmed by surgical exploration or a clinical follow-up of at least 12 months.

Results: Diagnosis was confirmed by clinical follow-up in 44 cases and surgical investigation or histology in 41 cases. A definitive diagnosis of the smears resulted in 40 malignant and 49 benign diagnoses. The sensitivity was 0.666 for CC and 0.658 for ICM, and the specificity was 0.920 and 0.937, respectively. The positive predictive value (PPV) was 0.866 for CC and 0.900 for ICM. McNemar's test did not reveal a significant difference between CC and ICM (P=0.803). Agreement of the two methods was found in 73 samples, raising specificity to 0.998 but not sensitivity (0.725).

Conclusions: ICM-DNA seems not to improve significantly the PPV and NPV for detecting neoplasias of the biliary tract compared to CC. Nevertheless a clinical advantage can be seen in the agreement of the two methods in diagnosing dysplasia or cancer, since it did not show false positive results.

Publication types

  • Comparative Study

MeSH terms

  • Aged
  • Ampulla of Vater
  • Bile Ducts / pathology*
  • Biliary Tract Neoplasms / complications*
  • Cholangiopancreatography, Endoscopic Retrograde
  • Cholestasis / etiology
  • Common Bile Duct Neoplasms / complications
  • Constriction, Pathologic
  • Female
  • Humans
  • Image Cytometry
  • Image Processing, Computer-Assisted
  • Male
  • Pancreatic Neoplasms / complications
  • Ploidies
  • Sensitivity and Specificity