Tumors expressing a high level of certain types of tumor-associated carbohydrate antigens (TACAs) exhibit greater metastasis and progression than those expressing low level of TACAs, as reflected in decreased patient survival rate. Well-documented examples of such TACAs are: (i) H/Le(y)/Le(a) in primary non-small cell lung carcinoma; (ii) sialyl-Le(x) (SLe(x)) and sialyl-Le(a) (SLe(a)) in various types of cancer; (iii) Tn and sialyl-Tn in colorectal, lung, breast, and many other cancers; (iv) GM2, GD2, and GD3 gangliosides in neuroectodermal tumors (melanoma and neuroblastoma); (v) globo-H in breast, ovarian, and prostate cancer; (vi) disialylgalactosylgloboside in renal cell carcinoma. Some glycosylations and TACAs suppress invasiveness and metastatic potential. Well-documented examples are: (i) blood group A antigen in primary lung carcinoma; (ii) bisecting beta1 --> 4GlcNAc of N-linked structure in melanoma and other cancers; (iii) galactosylgloboside (GalGb4) in seminoma. The biochemical mechanisms by which the above glycosylation changes promote or suppress tumor metastasis and invasion are mostly unknown. A few exceptional cases in which we have some knowledge are: (i) SLe(x) and SLe(a) function as E-selectin epitopes promoting tumor cell interaction with endothelial cells; (ii) some tumor cells interact through binding of TACA to specific proteins other than selectin, or to specific carbohydrate expressed on endothelial cells or other target cells (carbohydrate-carbohydrate interaction); (iii) functional modification of adhesive receptor (integrin, cadherin, CD44) by glycosylation. So far, a few successful cases of anti-cancer vaccine in clinical trials have been reported, employing TACAs whose expression enhances malignancy. Examples are STn for suppression of breast cancer, GM2 and GD3 for melanoma, and globo-H for prostate cancer. Vaccine development canbe extended using other TACAs, with the following criteria for success: (i) the antigen is expressed highly on tumor cells; (ii) high antibody production depending on two factors: (a) clustering of antigen used in vaccine; (b) choice of appropriate carrier protein or lipid; (iii) high T cell response depending on choice of appropriate carrier protein or lipid; (iv) expression of the same antigen in normal epithelial tissues (e.g., renal, intestinal, colorectal) may not pose a major obstacle, i.e., these tissues are not damaged during immune response. Idiotypic anti-carbohydrate antibodies that mimic the surface profile of carbohydrate antigens, when administered to patients, elicit anti-carbohydrate antibody response, thus providing an effect similar to that of TACAs for suppression of tumor progression. An extension of this idea is the use of peptide mimetics of TACAs, based on phage display random peptide library. Although examples are so far highly limited, use of such "mimotopes" as immunogens may overcome the weak immunogenicity of TACAs in general.