Breast cancer, now the fourth leading cause of cancer-related mortality worldwide, necessitates early detection for improved clinical outcomes. Conventional histopathology, though widely used, is invasive and subjective, limiting its utility in early-stage diagnosis. Here, we integrated confocal Raman spectroscopy with metabolomics to analyze biochemical and morphological features of breast tissues, including normal, fibroadenoma, ductal carcinoma in situ (DCIS), and invasive ductal carcinoma (IDC). Using spectral analysis and spectral unmixing, we mapped key biochemical components-proteins, lipids, and nucleic acids-across distinct tissue types. Cancerous tissues displayed heightened signals for proteins and nucleic acids but reduced lipids and carotenoids, reflecting profound metabolic alterations. Validation through metabolomic profiling revealed upregulated glycolytic and lipid synthesis pathways in tumor regions. Raman imaging further enabled precise classification of breast cancer subtypes, such as mucinous carcinoma and phyllodes tumors. These findings underscore the potential of Raman spectroscopy as a non-invasive diagnostic modality for early detection and classification of breast cancer. The integration of Raman imaging with machine learning presents a promising avenue for advancing precision oncology.
X et al. (Mon,) studied this question.