Since its inception in the mid-1980s of the 20th century testing for carbohydrate antigen 19-9 (CA 19-9) has raised expectation for an earlier diagnosis and accurate monitoring of several malignant diseases. After almost 30 years, the available evidences have confirmed the appropriateness and usefulness of determining CA 19-9 levels as a prognostic indicator and as a reliable tool for monitoring pancreatic and gastrointestinal cancer, but concerns have been raised about its applications in screening, which is actually not recommended, and in the diagnosis of malignancies, due to several interferences that limit the specificity and to the insufficient sensitivity of this marker. In this paper we aimed to review the basic concepts of CA 19-9 testing and its current applications, with a major focus on the most recent evidences dealing with assay interference, methods comparison and monitoring of malignant diseases. The prognostic value and monitoring recommendations for pancreatic, gastric and colorectal cancers are described in depth.