Carbohydrate antigen 19.9 (CA19.9) is used as a
tumor marker for clinical and research purposes assuming that it is abundantly produced by
gastrointestinal cancer cells due to a
cancer-associated aberrant glycosylation favoring its synthesis. Recent data has instead suggested a different picture, where immunodetection on tissue sections matches biochemical and molecular data. In addition to CA19.9, structurally related
carbohydrate antigens Lewis a and Lewis b are, in fact, undetectable in
colon cancer, due to the down-regulation of a
galactosyltransferase necessary for their synthesis. In the pancreas, no differential expression of CA19.9 or cognate
glycosyltransferases occurs in
cancer. Ductal cells only express such
Lewis antigens in a pattern affected by the relative levels of each
glycosyltransferase, which are genetically and epigenetically determined. The elevation of circulating
antigens seems to depend on the obstruction of neoplastic ducts and loss of polarity occurring in malignant ductal cells. Circulating Lewis a and Lewis b are indeed promising candidates for monitoring
pancreatic cancer patients that are negative for CA19.9, but not for improving the low diagnostic performance of such an
antigen. Insufficient
biological data are available for gastric and
bile duct cancer. Studying each patient in a personalized manner determining all
Lewis antigens in the surgical specimens and in the blood, together with the status of the tissue-specific glycosylation machinery, promises fruitful advances in translational research and clinical practice.