Fast multimodal imaging combined with machine learning identifying taurine as a potential marker for breast cancer margin assessment

Sensitive and accurate determination of margins between breast-conserving surgery (BCS) is challenging. This study provided a novel approach and biomarker for margin detection of BC through multi-modality imaging of femtosecond label-free imaging (FLI) microscopy and imaging mass spectrometry (IMS) plus machine learning (ML). The regions of interest (ROI) identified by FLI microscopy closely match tissue micro-regions diagnosed by pathologists. Additionally, a biomarker panel comprising taurine, threonate, and glutamate was established for classifying BC and tumor-adjacent noncancerous breast (TANB) tissues. Among these, taurine with higher abundance was confirmed as a potential biomarker to assess positive margins. Elevated level of taurine was associated with poor overall survival in BC patients. Functional analysis validated pro-tumorigenic effect of taurine in BC cell lines. Overall, the integration of FLI microscopy and IMS enables rapid visualization of cellular structures and metabolites in unlabeled tissues, with further analysis identifying taurine as a potential biomarker for margin evaluation.