Prospective randomized controlled study comparing cell block method and conventional smear method for bile cytology

Abstract

There is a paucity of data on the cell block (CB) method for bile cytology. We compared the diagnostic efficacy of the CB method with that of conventional smear cytology for bile obtained by endoscopic retrograde cholangiopancreatography (ERCP) in a randomized controlled trial manner. A total of 137 patients with biliary tract lesions suspicious of malignancy who had undergone bile collection under ERCP were recruited to this study. After sampling, the bile was randomized to the CB method (n = 69) or to smear cytology (n = 68). CB sections were prepared using the sodium alginate method and subjected to hematoxylin-eosin, Alcian blue-periodic acid-Schiff stain, and immunohistochemical stains. Both Papanicolaou and Giemsa stains were used for smear cytology. The final diagnosis was malignancy in 94 patients: bile duct cancer, 42; pancreatic head cancer, 34; gallbladder cancer, 16; and ampullary cancer, two. The diagnostic accuracy of the CB method and that of smear cytology were 64% and 53%, respectively (P = 0.20). The sensitivity of the CB method (53%) was significantly better than that of smear cytology (28%; P = 0.014). Their respective sensitivities were 80% and 31% (P = 0.002) for bile duct cancer, 20% and 15% (P = 1.0) for pancreatic head cancer, and 30% and 67% (P = 0.30) for gallbladder cancer. The CB method for bile cytology showed a higher diagnostic yield than smear cytology. Its diagnostic sensitivity was satisfactory in cases of bile duct cancer.