[Development and clinical research of computer aided multivariate pattern analysis system (CAMPAS) OV-1 for diagnosis...

A computer aided multivariate pattern analysis system (CAMPAS) OV-1, which consisted of 10 discriminant functions based on eight tumor markers including CA125, IAP, TPA, LDH, CRP, CEA, amylase and alkaline phosphatase was developed to effectively diagnose patients with epithelial ovarian carcinoma. One hundred twenty-two patents with epithelial ovarian carcinoma and 215 patients with benign ovarian tumors were examined by using CAMPAS OV-1 retrospectively or prospectively. The clinical significance of CAMPAS OV-1 was compared with CA125 alone, and with a combined assay employing the eight tumor markers used in CAMPAS OV-1. The following results were obtained. 1. When CAMPAS OV-1 was applied to patients in which the value for each tumor marker was used to make the discriminant functions, the sensitivity, specificity and accuracy were 84.5%, 97.5% and 91.3%, respectively. The accuracy of CAMPAS OV-1 (91.3%) was significantly better than those of CA125 (80.0%) and combined assay (74.0%) [p < 0.001]. CAMPAS OV-1 showed relatively better sensitivity (63.3%) than CA125 (50.0%) in patients with stage I disease. Also CAMPAS OV-1 showed relatively better sensitivity (85.7%) than both CA125 (64.3%) and combined assay (78.6%) in patients with mucinous type tumors. Furthermore, the specificity of CAMPAS OV-1 (94.4%) was significantly better than those of CA125 (66.7%) and combined assay (55.6%) in patients with endometrial cysts [p < 0.05]. 2. When CAMPAS OV-1 was applied to the patients prospectively, whose values for each tumor marker were not used to make the discriminant functions, the sensitivity, specificity and accuracy were 88.2%, 83.8% and 85.0% respectively.(ABSTRACT TRUNCATED AT 250 WORDS)