Abstract

Background: Endoscopic retrograde cholangiopancreatography (ERCP) with biliary brush cytology is commonly used to diagnose malignant pancreatobiliary strictures. This trial compared the sensitivity of two intraductal brush cytology devices. Methods: A randomized controlled trial in which consecutive patients with suspected malignant, extrahepatic biliary strictures were randomized (1:1) to a dense or conventional brush cytology device. Primary endpoint was sensitivity. Interim analysis was conducted after 50% of the patients completed follow-up. Results were interpreted by a data safety monitoring board. Results: Between June 2016 and June 2021, 64 patients were randomized to the dense (27 patients, 42%) or conventional brush (37 patients, 58%). Malignancy was diagnosed in 60 patients (94%) and benign disease in 4 patients (6%). Diagnoses were confirmed by histopathology in 34 patients (53%), cytopathology in 24 patients (38%), and clinical or radiological follow up in 6 patients (9%). Sensitivity of the dense brush was 50%, compared to 44% for the conventional brush (p = 0·785). Discussion: The results of this randomized controlled trial showed that the sensitivity of a dense brush is not superior to a conventional brush for diagnosing malignant extrahepatic pancreatobiliary strictures. This trial was prematurely ended for reasons of futility. Trial registration: Netherlands Trial Register number; NTR5458.
Original languageEnglish
Pages (from-to)4566-4573
Number of pages8
JournalSurgical endoscopy
Volume37
Issue number6
Early online date2023
DOIs
Publication statusPublished - Jun 2023

Keywords

  • Bile Duct Neoplasms/complications
  • Biliary Tract
  • Biliary tract neoplasms
  • Cholangiopancreatography, Endoscopic Retrograde
  • Cholestasis/etiology
  • Clinical pathology
  • Constriction, Pathologic/etiology
  • Endoscopic retrograde cholangiopancreatography
  • Humans
  • Pancreatic Neoplasms/complications
  • Pancreatic neoplasms
  • Sensitivity
  • Sensitivity and Specificity
  • Specificity

Cite this