【摘要】 收集114例胃癌病例、60例慢性胃病病例和25例正常对照者,选择胃癌流行病学与病因学上有关指标及临床体征、大便隐血等61项因素,根据模式识别原理建立程序在计算机上经自学后建立识别系数,并对146例胃癌、170例胃病、92例正常人进行判别,其符合率分别为89%、55.3%和90.2%,均较单纯应用线性回归筛选为高。结合线性回归方法对本市某厂1108名职工普查筛选出“阳性”者79人,全部经胃镜检查并作活检,发现1例早期胃癌,29例萎缩性胃炎(5例伴有不典型增生),3例胃息肉,其余为不同程度的浅表性胃炎,胃癌普查检出率为0.9‰。更多还原

【Abstract】 This article reports the application of pattern recognition method in mass screening of gastric cancer. 114 cases of gastric cancer, 60 cases of various chronic gastric diseases and ’5 cases of normal controls were collected. From these cases 61 factors including age, sex and some possible carcinogenic factors in life style, and various symptoms were looked into and recorded. A computer program was designed according to the principle of pattern recognition to establish recognition coeffcient. 146 cases of gastric cancer, 170 cases of various chronic diseases and 92 normal controls were investigated and recognized by this method, the rates of correspondence being 89%, 55% and 91% respectively, which were all higher than those from linear regression method. Using this method we mass screened 1108 employees in a Shanghai factory in August 1986, and found 79 cases "positive", among which one case of gastric cancer, 3 cases of gastric polyps, and 76 cases of superficial gastritis and atrophic gastritis were diagnosed by endoscopic biopsy.更多还原