Abstract Background: Nipple discharge is a common clinical manifestation of breast disease, yet current diagnostic tools such as ultrasound, mammography, MRI, ductography, and cytology have limited accuracy in early cancer detection. Copy number variation (CNV) profiling of DNA from body fluids has shown promise in early diagnosis of cancers like bladder and cervical, but its application to nipple discharge remains unexplored. This study aimed to establish a noninvasive, CNV-based diagnostic approach that reflects primary breast tumor biology with high specificity and sensitivity. Methods: We analyzed CNV data from 1,086 tumor and 942 normal samples in the TCGA-BRCA cohort using GISTIC2.0, identifying recurrent CNV regions with high frequency and statistical significance (q0.0001). Selection criteria included CNV frequency 50%, biological interpretability (oncogenes in amplifications, tumor suppressors in deletions), consistency with gene expression changes, and prognostic relevance. A final panel of 20 CNV features (12 amplifications, 8 deletions) was established and validated across multiple breast cancer cohorts and their matched normal tissues, when available. Diagnostic performance was further tested using nipple discharge DNA an in-house cohort of patients with benign and malignant breast disease. Additionally, single-cell RNA sequencing was performed on tumor tissues from five breast cancer patients in the cohort. Using inferCNV, the CNV features detected in nipple discharge were traced back to distinct epithelial subclusters using inferCNV, and the functional characteristics of the dominant CNV-contributing subpopulations were further analyzed. Results: In the TCGA-BRCA dataset, only 8 of 118 adjacent normal tissues were panel-positive, yielding a specificity of 93.22%. External validation demonstrated consistent diagnostic performance: GSE87048 (sensitivity 77.0% in tumor tissue, specificity 99.0% in matched peripheral blood mononuclear cells), GSE118527 (sensitivity 92.8%, specificity 100%), METABRIC (sensitivity 95.7%), and GSE31427 (Stage I-II; sensitivity 66.9%). In the MSK-CHORD PanCancer cohort, the panel showed the highest detection rate in breast cancer (78.3%) compared to colorectal, prostate, lung, and pancreatic cancers, supporting tissue specificity. In nipple discharge samples (n = 39 malignant, n = 18 benign), the panel achieved a sensitivity of 71.79% and a specificity of 94.44%, highlighting its potential for accurate, noninvasive early detection. In five patients from the in-house cohort, single-cell RNA sequencing of tumor tissues revealed that CNV features detected in nipple discharge originated from specific epithelial subclusters. These CNV-contributing subpopulations exhibited enrichment in pathways related to cell junction assembly and tissue morphogenesis. Conclusion: This study provides the first evidence that CNV profiling of nipple discharge DNA is a feasible and tumor-specific approach for early breast cancer detection. The selected CNV panel demonstrates robust diagnostic performance across independent datasets and early-stage tumors, supporting its clinical utility as a novel, minimally invasive screening tool. Citation Format: W. Wen, P. Pu, H. Xu, T. Shang, L. Cong, G. Qiao, Z. Jia, Y. Liu, Y. Huang, R. Zhou, Y. Li, J. Liu. Noninvasive Early Detection of Breast Cancer via Copy Number Variation Profiling of Nipple Discharge DNA abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS4-01-25.
Wen et al. (Tue,) studied this question.