Comparative CEA-Specific Near-Infrared Probes Incorporated with Confocal Laser Endomicroscopy for In Situ Imaging of Colorectal Cancer Cells

Abstract Optimizing the targeting ligand is a critical strategy to enhance the diagnostic specificity of near-infrared (NIR) confocal laser endomicroscopy (CLE) for colorectal cancer. However, it remains unclear how distinct recognition elements, specifically antibodies versus aptamers, differ in their CLE imaging performance and capacity to distinguish tumors from inflammation-associated backgrounds. To address this, we developed a targeted NIR-CLE paradigm by integrating a customized video-rate CLE platform with two carcinoembryonic antigen (CEA)-targeted NIR probes synthesized for comparative evaluation: an antibody-based probe (Insi-ACI) and an aptamer-based probe (Insi-ACIapt), representing protein-recognition and nucleic-acid-recognition probe formats, respectively, and both utilizing indocyanine green (ICG) as the fluorophore. Following covalent synthesis and systematic spectroscopic characterization, the imaging performance of these two probes was systematically evaluated across CEA-high, CEA-low, and CEA-negative cell models. Quantitative analyses of fluorescence intensity, signal-to-noise ratio (SNR), and spatial contrast were subsequently performed. While both probes selectively labeled CEA-positive cells, Insi-ACIapt produced superior fluorescence brightness, whereas Insi-ACI yielded a higher imaging contrast. CLE imaging further revealed stable, membrane-associated boundary patterns across multiple fields of view. Importantly, spatial analysis effectively differentiated tumor-associated signals from inflammation-induced false positives based on texture and line-scan profiles rather than relying solely on absolute intensity. Positive tumor cells consistently displayed a defined membrane-associated ring-like distribution, whereas inflammatory macrophages exhibited disorganized fluorescence patterns lacking structural boundaries. These results establish a pattern-recognition-based NIR-CLE framework for cell-level molecular imaging toward targeted in situ endoscopic diagnosis and indicate that antibody and aptamer probes may be tailored to diverse clinical imaging scenarios.