Role of brush cytology in the diagnosis of neoplastic and non-neoplastic upper gastrointestinal lesions

Objective: To evaluate the utility of brush cytology (BC) in diagnosing non-neoplastic and neoplastic lesions of the upper gastrointestinal (GI) tract. Methods: The study comprised 106 brushings from 104 patients with clinically suspected lesions in the upper GI tract. Diagnoses were rendered on correlating brush cytological findings with clinicoradiological findings and compared with endoscopic biopsies (EB). An attempt was made to grade oesophageal squamous cell carcinomas (SCC) and to subcategorise gastric adenocarcinomas (AC). The sensitivity, specificity and predictive values of BC were calculated with respect to clinico-radiological outcome. Results: Malignant and benign lesions on BC comprised 74.5% and 25.5%, respectively, and on histology, 64.2% and 28.3%, respectively. Twenty-one of 32 (65.6%) gastric ACs were accurately sub-typed (16 intestinal and five diffuse) and 25/36 (69.4%) oesophageal SCCs (7, 15 and three well, moderately and poorly differentiated, respectively) were accurately graded. On correlation with the final clinico-radiological diagnosis, there was one false-positive and no false-negative cases, respectively, on BC; three false-negative and no false-positive cases, respectively, on EB; eight EB were non-representative in which a diagnosis was not rendered, out of which seven were malignant. Thus, BC had a sensitivity and specificity of 100% and 96.4%, respectively, with a positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy (DA) of 98.7%, 100% and 99.1% (P < 0.05), respectively. BC was superior to EB in detecting Candida species and EB in detecting Helicobacter pylori. Conclusions: BC is a useful, routinely applicable and reliable test for diagnosing upper GI lesions.