We agree with Gutiirrez and colleagues that pancreatic cancer is a multistep process involving several oncogenes and tumor suppressor genes. Pancreatic cancer, however, is the human tumor with the highest incidence of ras mutations identified to date [1]. Furthermore, the finding of mutations exclusively affecting codon 12 in the K-ras gene was unexpected, given the results of previous experimental carcinogenesis studies. Until recently, our knowledge of the molecular mechanisms of pancreatic cancer was poor compared with our knowledge of the mechanisms of colon cancer. At least two different factors make this research difficult. First, most animal models, except perhaps the Syrian golden hamster model, are not directly comparable to human duct carcinoma; second, pathologic studies are often limited by the late stage of the tumor sample. Nevertheless, several reports address the activation of other oncogenes and deletion or inactivation of tumor suppressor genes. For example, the reported frequencies for erb2 and Rb1 alterations are less than 20%. A recently reported high incidence of somatic mutations and homozygous deletions of the p16MTS1 gene is also of interest. This discovery documents the role of a cyclin-dependant kinase inhibitor as a tumor suppressor gene [2]. The highest frequency of inactivation of the tumor suppressor gene p53 is found in invasive tumors and metastases [3].

These data support a genetic model of pancreatic ductal cell tumorigenesis initiated by almost universal mutation of a dominant oncogene, K-ras, associated with deregulation of cell-cycle control by frequent mutations of p53 and MTS1 [4].

We believe that the close follow-up of patients who have no pancreatic mass but have K-ras mutations in pancreatic juice will yield further information on the profile of other molecular genetic alterations. Furthermore, the identification of the precise type of mutation affecting codon 12 might aid in the identification of risk factors.

In our opinion, the K-ras mutation study, although associated with other well-chosen molecular markers, should not be recommended for a large-scale risk-assessment program. However, we do recommend the test for patients requiring endoscopic retrograde cholangiopancreatography for therapeutic or diagnostic purposes. The identification of a subgroup of patients harboring an activation of this early oncogene will probably be essential for the definition of the panel of genetic tests needed to improve the treatment of patients with this form of cancer.