In Japan, R2 lymph node dissection is standard practice for all curatively resected gastric cancer patients. From 1969 to 1984, data were collected prospectively to evaluate this procedure and to provide information for a more rational approach to node dissection for each individual case. A total of 3843 cases was included in this study and an evaluation was made of the prognostic variables and lymph node involvement at various locations. Using a computer program, it is possible to compile a group of treated patients, with prognostic variables exactly matching those of an individual patient. Analysis of this group can then give an indication of the extent of lymph node dissection required for an individual patient. This paper gives a demonstration of the structure of such a system by means of an example.