Data published during the latter part of 1980s have shown that the natural history of gastrointestinal cancers can be influenced by treatment. This was further confirmed in esophageal cancers, for which many phase II studies have shown that chemotherapy with and without radiotherapy can induce major responses before surgery. Trials demonstrating a survival benefit are needed. Gastric cancers were confirmed to be sensitive to chemotherapy. Treatment based on combinations of fluorouracil, methotrexate, doxorubicin, etoposide, and cisplatin have shown high response rates (FAMTX [fluorouracil, doxorubicin, and methotrexate], EAP [etoposide, doxorubicin, and cisplatin], ELF [etoposide, leucovorin, and fluorouracil]) and a survival benefit (FAMTX). Adjuvant treatment after surgery using the FAM (fluorouracil, epirubicin, methotrexate) regimen has failed to improve survival. The prognosis for other cancers remains dismal. When all types of gastrointestinal cancers except colon cancers are reviewed, current data indicate that chemotherapy should not be recommended outside the confines of prospective studies.