Abstract P4-09-23: Gene expression signatures of microcalcifications among Taiwanese breast cancers

Abstract

Abstract Introduction: Microcalcification is one of the most common radiological and pathological features of breast ductal carcinoma in situ (DCIS), and to a lesser extent, invasive breast cancer. In current study, we evaluated the transcriptional profiles associated with the phenomenon of ectopic mammary mineralization and a gene expression signature is derived. Materials and methods: a total of 109 consecutive breast invasive cancers were prospectively collected and assayed with Affymetrix Human Genome U133 Plus 2.0 microarrays. The presence of microcalcification was confirmed by histopathological examinations as well as reviews of pre-operative mammography. The associations of gene expression profiles with microcalcifications and relevant clinical features such as DCIS including both comedo/high-grade and non-comedo subtypes were tested. Results: Microcalcifications were presented in 84 (80%) of the study population as confirmed by pathological examination. Of these 84 patients, 81 (96%) were grown with coexistent DCIS microscopically, while only 8 (38%) of the 21 patients without concurrent microcalcifications, the invasive tumors were accompanied with DCIS (Chi-square test, P<0.001). In addition, high-grade (comedo) type DCIS were presented in 44 (54%) of the 81 cancers with microcalcifications whereas only 15% (n=2) of tumors with DCIS but without microcalcificaitons were of high-grade (comedo) type. There were 69 genes differentially expressed between breast cancers with and without microcalcifications (nominal P<0.001 with 10,000 random permutations), and 11 were associated with high-grade (comedo) type DCIS including APOD, CCDC183, SLMO1, SLC6A5, FMO1, QPRT and CES4A. The enriched Gene Ontology categories encompasses glycosaminoglycan, aminoglycan metabolic processes, Golgi apparatus cellular component and protein ubiquitination, indicating an active secretory process. The intersect (18 probesets) of microcalcificaion and DCIS-associated genes provided the best predictive accuracy of 82% with Bayesian compound covariate predictor. Performance of compound covariate predictor classifier:ClassSensitivitySpecificityPPVNPVWith microcalcification0.8260.2170.7980.25Without microcalcification0.2170.8260.250.798 Compared with mammography alone, the diagnostic accuracy of gene expression-based signature is much improved (cross-validated ROC AUC: 0.738). Discussion: Our study suggested that mammary microcalcification is not only the earlier detectable radiological finding for disease screening but the phenomenon itself resulted from distinct biological processes that constituted the molecular heterogeneity of human breast cancers. Further studies to evaluate the prognostic significance of microcalcifications are warranted. Citation Format: Huang C-C, Huang C-S, Tu S-H, Tsai M-L. Gene expression signatures of microcalcifications among Taiwanese breast cancers. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-09-23.