Purpose: We aimed to develop noninvasive and early detection breast cancer biomarkers panel that may serve as assistant diagnostic method.Methods: 61 biomarkers were detected in sea of 101 healthy controls, 46 benign breast diseases and 77 breast cancer patients in the training group. A metropolis algorithm with Monte Carlo simulation was used for choosing the model. 444 individuals were used for validation. Serum from 245 female cancer patients including 5 kinds of cancers were also collected to evaluate cancer selectivity. Results: Panel consisting of Apolipoprotein A І (ApoA І ), ApopB, C-reactive protein (CRP) and interleukin (IL)-8 had the highest value for discriminating between breast cancer and healthy control. The sensitivity (SN) was 98.70% for all-stage, 100.00% for early-stage and 97.92% for advanced-stage with 90% specificity (SP). In the validation group, the sensitivities were 96.43%, 100.00% and 94.21% at 90% SP. This panel identified 14.29% cervical cancer, 0% lung cancer, 20.29% pancreatic cancer, 25.00% gastric cancer, and 17.50% colorectal cancer as non-breast cancer. Panel consisting of Pepsinogen (PG) І /II, CRP, Superoxide dismutase, Tumor necrosis factor α had the highest value for discriminating between breast cancer and benign breast diseases. The SN was 88.31% for all-stage, 72.41% for early-stage and 97.92% for advanced-stage with 90% SP. In the validation group, the sensitivities were 81.25%, 69.77% and 88.41% at 90% SP.Conclusions: The biomarker panels showed an improved performance when compared to CA153. It may serve as assistant tools for breast cancer screening and early detection to improve the clinical outcome.

No abstract

No abstract