Background: A preoperative tissue diagnosis of pancreatic cancer is desirable but difficult to obtain.
Methods: Pancreatic brush cytology, salvage cytology, and collection of pancreatic juice were attempted prospectively during ERCP in 34 patients with pancreatic cancer and 11 with chronic pancreatitis. K-ras-2 codon 12 was analyzed for presence and type of point mutations.
Results: Brush cytology coupled with salvage cytology had a sensitivity of 74%. The addition of cytologic analysis of pancreatic juice did not substantially improve sensitivity (76%). K-ras-2 was mutated in both cancer (87%) and pancreatitis (40%). The specificity for cytology was 100% and for K-ras-2 mutations 60%. Combining cytology with mutation analysis increased sensitivity to 93% but reduced the positive predictive value. The negative predictive value never exceeded 75%. None of the patients with chronic pancreatitis had cancer develop (median follow-up 60 months).
Conclusions: Pancreatic ductal brushing with salvage cytology is useful in the diagnosis of cancer, whereas cytologic analysis of pancreatic juice can be abandoned. At present, K-ras-2 mutation is not useful for differentiating pancreatic cancer from chronic pancreatitis or the identification of patients with chronic pancreatitis at risk for malignant transformation.