Background & Aims: Prostate cancer (PCa) is the second most common cancer in men worldwide. The identification of sensitive and specific biomarkers in tissue and serum is of utmost importance to reduce the mortality of prostate cancer. Since that, early detection of cancer has an important role in treatment, in this study we tried to identify genes that could potentially effective in early screening for prostate cancer. Using logistic regression, suitable model to screen tumor samples from normal samples was designed.
Materials & Methods: In this study, gene expression data of metastatic and non-metastatic cancer samples and normal prostate samples were collected from the NCBI database. By examining the expression level of genes in these samples, valuable genes for screening were identified.
Results: Using logistic regression two model were designed based on the increase or decrease in the expression of genes. The Area under the curve, sensitivity and specificity for the first model were, respectively, 0.968, 0.911 and 0.914 and for the second model 0.991, 0.951 and 0.956, respectively.
Conclusion: Due to the high value of sensitivity and specificity in the designed models, studied Genes have the potential for screening prostate cancer in the early stages and metastasis stages of cancer