Computer recognition of gastric (pre)malignancy with the aid of some biochemical, immunological and cytological data

17 patients with gastric pre-malignancy and 26 without it were correctly classified on the basis of the data from gastric juice (amount, pH, diphenylamine extinction, immunoglobulin A concentration, and total, parietal and surface epithelial cell counts) and from serum (immunoglobulin A concentration) using a kybernetic method of pattern recognition. The most decisive indices appear to be the diphenylamine extinction and immunoglobulin A concentration of gastric juice. Although X-rays, endoscopy and biopsy are the substantial diagnostic methods for gastric (pre-) malignancy a wide screening is desirable to recognize those people who are still asymptomatic but need the special examinations. We propose for this purpose a set of a few simple tests from blood serum and gastric juice, evaluated by method of pattern recognition.