Intraductal papillary neoplasms (IPNs) of the biliary tract are uncommon lesions that may be solitary or may spread extensively along the biliary tree. Some biliary IPNs are histologically and radiologically similar to intraductal papillary mucinous tumors (IPMNs) of the pancreas and present a risk for progression to invasive cholangiocarcinoma. Unlike pancreatic IPMNs, little is known about their molecular pathogenesis. We studied 14 biliary IPNs (including 5 cases with associated invasive cholangiocarcinoma) for genetic alterations in the APC/beta-catenin pathway, K-ras oncogene mutations, p53/chromosome 17p alterations, and Dpc4/18q alterations. Immunohistochemistry was performed for beta-catenin, p53, and Dpc4, and microdissected tissue was analyzed using direct DNA sequencing for exon 1 of K-ras and exon 3 of beta-catenin and allelic loss assays on chromosomes 5q, 17p, and 18q. Activating mutations in codon 12 of the K-ras oncogene were present in 4 of 14 (29%) biliary IPNs. Of these 4 cases, 2 patients had associated invasive cholangiocarcinoma, and identical K-ras mutations were present in both the intraductal and invasive components. Allelic loss on chromosome 18q was present in 4 of 13 informative cases (31%); however, no loss of normal Dpc4 expression was detected by immunohistochemistry. Nuclear accumulation of beta-catenin protein was demonstrated in 3 of 12 cases (25%); however, there were no beta-catenin gene mutations, and allelic loss on 5q was present in only 1 of 10 informative cases (10%). Both immunohistochemistry for p53 and 17p allelic loss assays were negative. Biliary IPNs therefore demonstrate a K-ras gene mutation frequency that is lower than that previously reported for pancreatic IPMNs, but similar to that reported for hepatic cholangiocarcinomas. The presence of K-ras mutations in 2 purely intraductal neoplasms, and identical K-ras mutations in 2 cases with both intraductal and invasive components, suggests that these mutations arise early in tumorigenesis. Finally, the frequency of allelic loss on 18q suggests that a locus on 18q is involved in the molecular pathogenesis of biliary IPNs, but this locus is not DPC4.