Factors related to outcome were investigated in a consecutive series of 201 gastric cancer patients identified by endoscopy from January 1973 to May 1984 in this institution. Five-year survival rates were found to correlate significantly with tumor stage, lymph node spread, and also with disease localization. Tumors in the distal or middle third of the stomach were associated with a significantly better outcome than tumors in the upper third, extensive tumors or tumors after previous gastric resections. The choice of operative procedures was likewise related to outcome. Patients who had undergone curative resections had by far the best prognosis with a 5-year survival rate of 44%, but this was primarily taken to represent a reflection of a less advanced disease stage rather than an independent prognostic factor. The histologic tumor type did not influence outcome in this series. We confirmed earlier observations that a long history of symptoms cannot be taken as a poor prognostic sign. A more liberal approach to upper gastrointestinal tract endoscopy in patients with unexplained abdominal pain or weight loss should contribute to improve the prognosis. Thus education of patients and primary care physicians rather than diagnostic or therapeutic modalities represent at present a major clinical concern in gastric cancer disease.