Abstract Introduction: Identification of metastatic sentinel lymph node (mSLN) is necessary for staging axillary and guiding following treatment in breast cancer patients. mSLNs exhibit significant heterogeneity in morphology and immune microenvironment, and might associate with the risk of recurrence/metastasis. Intraoperatively, frozen sections are utilized to assess SLN’s metastatic status, and to decide whether to perform axillary dissection or not. Previous studies primarily illustrated that near-infrared fluorescence (NIRF) imaging could visualize mSLNs in animal models. Thus, the purpose of this study is to investigate NIRF’s capability to stratify the subtypes of mSLNs, by combining with histopathological findings, and further to explore the mechanism underlying. Here, we report a TROP2-targeted second window of NIRF imaging approach for non-invasive, real-time identification of mSLNs subtypes. Materials and Methods: A novel NIR-II probe, ICG-SG, was synthesized by conjugating Sacituzumab Govitecan (SG) with ICG-Mal. In vivo NIRF imaging was applied to visualize mSLNs in mouse breast cancer model. In addition, mSLNs were stratified into distinct subtypes based on pathological features and mean fluorescence intensity (MFI). FFPE tissue sections from 345 invasive breast cancer patients were retrospectively analyzed, and histological subtypes of human mSLNs were assigned according to the classification established in the preclinical models. Moreover, three FFPE specimens per subtype underwent high-resolution spatial transcriptomics (Stereo-seq) and single-cell RNA sequencing (scRNA-seq) to map subtype-specific molecular features and mechanisms driving secondary lymph node metastasis. The association of subtypes with prognosis was also analyzed. Results and Summary In this study, we developed a TROP2-targeted probe ICG-SG for noninvasive, real-time detection of mSLNs in breast cancer. In both human- and mouse lines-derived models, we observed that ICG-SG exhibited significantly higher MFI than ICG-IgG in bilateral metastatic SLN models. In unilateral metastasis models with ICG-SG, mSLNs also showed markedly higher MFI than non-metastatic SLNs. Based on MFI and pathological features, mSLNs were classified into three subtypes: subcapsular (highest fluorescence), paracortical-medullary (intermediate, yet above normal), and replacement (negative fluorescence, below normal). Paracortical-medullary and replacement types had a higher risk of secondary LNs metastasis, which was also confirmed in a cohort of 345 patients, with odds 5.51- and 9.56-fold higher than subcapsular type, respectively. Whole-transcriptome analysis of representative mSLNs revealed distinct molecular and immune profiles across subtypes. Paracortical-medullary and replacement MSLNs exhibited aggressive tumor traits, impaired immune responses, and fibroblastic reticular cell-driven stromal remodeling, generating a pro-tumor microenvironment. In summary, mSLN subtypes intraoperatively and non-invasively determined predict non-sentinel lymph node metastasis and correlate closely with patient prognosis. The integrative imaging and molecular approach provides a framework for precise, real-time and noninvasive mSLN stratification and potential guidance for clinical decision-making. Citation Format: Y. zhu, W. Chen, J. chen, L. zhang, Z. Wu, Z. Wei, Q. Zhang, X. Duan, T. Guo, S. Guan, S. Song, Z. Guo, Y. Tang, C. Yang, C. Guo, C. Peng, G. Zhang. Integration of NIR-II Imaging and Molecular Profiling Identifies Subtypes of Metastatic Sentinel Lymph Nodes in Breast Cancer 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 PS1-13-28.
zhu et al. (Tue,) studied this question.