Acute phase serum amyloid A in ovarian cancer as an important component of proteome diagnostic profiling.

In the context of serum amyloid A (SAA) identification as ovarian cancer marker derived by SELDI-MS, its serum levels were measured by immunoassay in different stages of ovarian cancer, in benign gynecological tumors, and in healthy controls. In addition, SELDI-TOF-MS spectra were obtained by protocol optimized for the SAA peak intensity. SELDI data on small proteins (5.5-17.5 kDa) and SAA immunoassay data were combined with cancer antigen (CA)125 data in order to study the classification accuracy between cancer and noncancer by support vector machine (SVM), logistic regression, and top scoring pair classifiers. Although an addition of SAA immunoassay data to CA125 data did not significantly improve cancer/noncancer discrimination, SVM applied to combined biomarker data (CA125 and SAA immunoassay variables plus 48 SELDI peak variables) yielded the best classification rate (accuracy 95.2% vs. 86.2% for CA125 alone). Notably, most of discriminatory peaks selected by the classifiers have significant correlation with the major known peaks of SAA (11.7 kDa) and transthyretin (13.9 kDa). Acute phase serum amyloid A (A-SAA) was proved to be an important member of cancer discriminatory protein profile. Among the eight known ovarian cancer SELDI profile components, A-SAA is the most relevant to molecular pathogenesis of cancer and it has the highest degree of up-regulation in disease.