Creating A Molecular Classification Of Ovarian Serous Carcinoma By Tissue Microarray And Multi-protein Screening

BackgroundHistopathologic diagnosis is the gold standard of tumor diagnosis and the foundation of clinical treatments. But the patients with OSC which have the same clinicopathologic stage, pathological grade and therapy have different therapeutic reaction and prognosis. The reason of this difference is the individual heterogeneity of OSC on molecular level. Molecular classification distinguishing molecular heterogeneity is the foundation of individualized diagnosis and treatment. This study was undertaken to creat a prognostic molecular classification and a platinum-sensitivity molecular classification of OSC by tissue microarray and multi-protein screening.MethodsA total of122cases were taken from the surgical patients with advanced OSC in PUMCH from January2000to December2009. Follow-up study closed on31December2012. Two pathologists confirmed and choosed ovarian carcinoma tissue samples without calcification and necrosis. Specialists made tissue microarraies for the next step of this study.21proteins were investigated by combination of tissue microarraies and immunohistochemistry with the same physical, chemical and time parameters. Then microscopic images of theses samples were taken with the same hardware software parameters. Mean density of samples were gotted by grey values analysis with image pro plus6.0. Expression level of proteins were defined by mean density.Cases were divided by OS and DFS/PFS, poor prognosis group(OS<3years) and good prognosis group(OS≥3years), platinum-resistant group (DFS/PFS <1years) and platinum-sensitive group (DFS/PFS≥1years), cross analysis with expression level of21proteins. Statistic models of prognostic molecular classification and platinum-sensitivity molecular classificat-ion of OSC were gotted with Pearson’s chi square test, Spearman’s rank-correlation test and Logistic regression on spss17.0. ResultsCombined detection of Pim-1, HER-2and WT-1has the highest total rate of accuracy on progonosis of OSC. Compared to single detection of Pim-1, the total rate of accuracy on progonosis of OSC raised from77.9%to82.8%. Removing any protein in this model of prognostic molecular classification, effect of prediction would be changed markedly.Combined detection of HER-2, P-gp and Survivin has the highest total rate of accuracy on platinum-sensitivity of OSC. Compared to single detection of HER-2, the total rate of accuracy on progonosis of OSC raised from66.7%to70.4%. Removing any protein in this model of platinum-sensi-tivity molecular classification, effect of prediction would be changed markedly.For clinical convenience, low expression of protein is defined as0point, high expression of protein is defined as1point. Each molecular classification has three proteins, so the score of molecular classification is between0and3.0or1point means good prognosis or platinum-sensitivity,2or3points means poor pronosis or platinum-resistant.ConclusionsCombined detection of Pim-1, HER-2and WT-1can predict progonosis of OSC more accurately. Combined detection of HER-2, P-gp and Survivin can predict platinum-sensitivity of OSC more accurately.