3603

It is now well established that early detection of breast cancer improves survival rates and increases the probability of successful treatment. To date, mammographic screening is the most reliable method to detect breast cancer in asymptomatic women. However, mammography has significant limitations: it often fails to detect very small tumors in absence of microcalcifications and is difficult to interpret with dense breast in young women. These limitations lead us to evaluate an alternative method for screening breast cancers based on the identification of a molecular signature in peripheral blood cells. ExonHit’s gene profiling technology DATAS (Differential Analysis of Transcripts with Alternative Splicing) allowed the isolation of a set of messengers and alternatively spliced mRNA sequences that were differentially expressed between the blood of healthy women and early stage breast cancer patients. A customised Affymetrix Genechip array has then been set up for the validation of those DATAS clones. Total RNA extracted from whole blood samples of 37 controls and 55 breast cancer stage I/II (i.e. train group) were amplified, labeled and hybridized on the custom array in order to identify the most significant diagnostic markers. Monoparametric and multiparametric approaches have been carried out, leading to the identification of a 54-gene signature capable of clearly distinguishing healthy women from early stage breast cancers. By applying this method to the train group, 86.5% of the controls and 92.7% of the breast cancers were correctly classified. A first, limited validation with 21 independent blood samples (5 healthy controls vs 16 early stage-breast cancers) showed that this expression profile in blood kept its diagnostic power. The confirmation of the robustness of the blood-molecular signature is underway using a larger independent cohort of 1000 patients. In parallel, blood samples of patients suffering from benign breast diseases, inflammatory diseases or other types of cancer are processed in order to evaluate the specificity of the gene signature for breast cancer. Taken together, our results show that a gene expression-based assay can be developed for the early detection of breast cancer through the use of 5ml of whole blood and that such assay will rely on transcripts very unlikely to be expressed from circulating tumor cells.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]