Brush cytology for the routine microscopic examination of malignant pancreaticobiliary strictures has been shown to have a high degree of specificity but a sensitivity of less than 75%. We wanted to determine whether flow cytometric analysis for DNA content could increase the sensitivity of routine cytology in brushings from these strictures, especially in those patients with atypical cells noted by brush cytology but no definite tumor cells. Fifty-one sets of brushings were obtained from 48 patients in whom ERCP revealed pancreaticobiliary strictures. Specimens were obtained for both routine cytology and flow cytometry. Both studies were shown to have a sensitivity of 42% in diagnosing malignant strictures. The specificity was 92% for routine cytology and 77% for flow cytometry, with false-positive results obtained in one patient by the former technique and in three by the latter. When the studies were combined, so that the presence of either tumor cells or abnormal DNA content was diagnostic of malignancy, the sensitivity rose to 63%. However, the specificity fell to 69%. Routine cytology identified atypical cells in 13 patients with malignant strictures (39%). Flow cytometry identified abnormal DNA content in only six of these patients with atypical cytological specimens (46%). Although flow cytometry succeeded in identifying eight additional strictures as malignant, it was associated with a greater false-positive rate than was routine cytology. For patients with pancreatic carcinoma, increased survival was noted in those with a diploid cell population (mean, 8.9 months) as revealed by flow cytometry compared to those with aneuploid brushings (mean, 3.0 months). We conclude that the addition of flow cytometry for DNA content to routine cytology increases the diagnostic yield of brushings from pancreaticobiliary strictures. However, these results need to be interpreted together with the patient's history, ERCP, and other clinical findings in order to minimize false-positive results.