Breast Cancer Incidence by Age at Finding of Mammographic Abnormality in Women Participating in French Organized Screening Campaigns.

Purpose: Using reduced samples, statistical modelling was already predicted the occurrence of Breast-cancer or its prognosis from previous radiological findings. This study aims to predict breast-cancer risk by mammographic abnormalities finding age in the French breast-cancer screening campaign. Methods: The study involved 261,083 women aged 50-74 living in French Departments (Ain, Doubs, Haute-Saône, Jura, Territoire-de-Belfort, Yonne). These women had at least two screening mammograms between Jan-1999 and Dec-2017 of which the first was classified as “normal/benign”. The incidence of mammographic-abnormality (microcalcification, spiculated-mass, obscured-mass, architectural-distortion, asymmetric-density) and the incidence of breast-cancer after abnormality detection were estimated abnormalities finding age, using an actuarial life-table method. Breast-cancer risk was predicted in a Cox multivariate model.Results: The incidence of mammographic-abnormality was 95.4[94.9; 95.9]/1000 person-years. Breast-cancer (6,326 cases) incidence was 3.3[3.0; 3.1]/1000 person-years. That incidence was 5 times higher in women who showed a speculated-mass vs. those who did not (6.9[6.4; 7.4] vs. 1.3[1.2; 1.3]). Whatever the abnormality, the incidence of cancer was higher when it was present in only one breast. Depending on the spiculated-mass finding age, the risk increased by at least 40% between the age groups 55-59years (1.4[1.0; 1.8]) and ≥70years (2.4[1.9; 3.3]).Conclusion: The study showed the increased risk of cancer with the abnormalities finding age and the low risk related to the presence of the same mammographic-abnormality in both breasts compared to the isolated mammographic-abnormality in one of the breasts. This should alert radiologists to the relevance of certain diagnostic procedures in the management of a bilateral mammographic abnormality.

Biomarkers for predicting cancer in women with a suspicious mammogram.

e15533 Background: Biomarker identification for early breast cancer diagnosis is confounded by comparing healthy control patients to patients undergoing surgical procedures and stress of a potential cancer diagnosis. We implemented a clinical research protocol that combines biomarker harvesting and identification with Breast Imaging-Reporting and Data System (BIRADS) results, within a cohort of women with a suspicious mammogram who donated samples prior to biopsy. The primary goals were to discover candidate novel plasma markers for stage I breast cancer versus benign lesions, and validate the markers by mass spectrometry and immunohistochemistry. Methods: 150 women found on screening mammography to have a BIRADS IV or V mammographic abnormality were enrolled in the IRB approved study, with one year follow-up. After informed consent, serum, plasma, and saliva specimens were obtained and frozen. The patient underwent image guided biopsy of the mammographic abnormality. Hydrogel nanoparticles were used to harvest and concentrate low abundance protein biomarkers from plasma. Proteins were identified by mass spectrometry. The BIRADS score and biopsy outcome were blinded to the laboratory researchers. Results: 37/150 women (median age 64, 73% ER+, 70% PR+) were diagnosed with biopsy-proven breast cancer. 15/37 had a family history of breast cancer. Within the context of stress of an abnormal mammogram and invasive biopsy, we identified 5478 plasma peptides. A model to predict endpoints that discriminate cancer vs no cancer was developed using cross-validation and lasso shrinkage method. The best fit multi-analyte ROC/AUC model of peptide spectral matches revealed 10 candidate peptides, including PLAA, TRAPPC9, PROS1, DDX41, ANKRD63, EGFLAM (AUC = 0.81), that discriminated cancer versus no cancer. The functional mechanisms of these proteins are calcium metabolism, GPI anchor biosynthesis, neural-immune crosstalk, DNA repair, and ubiquitin-mediated protein trafficking. Conclusions: Molecular profiling of blood can potentially complement imaging to improve diagnostic specificity in the setting of a suspicious mammogram. This unique trial design, enhanced by nanotechnology protein harvesting, identified potential novel cancer biomarkers in the presence of a suspicious mammogram. A confirmation trial is underway.