Routine cytologic (RC), fluorescence in situ hybridization (FISH), digital image analysis (DIA), and quantifiable morphometric results from 284 pancreatobiliary stricture brushings were compared. We chose specific DIA nuclear features assessed by pathologists in evaluating RC specimens, such as area and shape. A visual nuclear morphometric score (VNMS) was calculated. There was a difference (P < .001) in the mean VNMS when RC results were classified as negative (11.5), atypical (12.5), suspicious (13.8), and positive (16.5). The mean VNMS of specimens diagnosed as disomy (11.3), trisomy 7 (12.1), and polysomy (14.7) by FISH was also different (P < .001). There was no difference in the VNMS of false-negative and true-negative cytologic specimens (P = .225). Our findings substantiate the relationship between cell nuclear visual alterations and genetic FISH abnormalities. The low sensitivity of cytologic examination for pancreatobiliary carcinoma is due to an absence of tumor cells or the presence of well-differentiated tumor lacking recognizable nuclear atypia.