Deep-learning endomicroscope with large field-of-view and depth-of-field for real-time in vivo imaging of epithelial cancer hallmarks
Menée à l'aide d'échantillons tissulaires d'origine porcine ou humaine et menée sur des adultes, cette étude met en évidence l'intérêt, pour détecter in vivo et en temps réel des marqueurs épithéliaux de lésions cancéreuses, d'un endomicroscope utilisant un algorithme d'apprentissage profond et possédant un large champ de vision ainsi qu'une grande profondeur de champ
Early cancer detection is crucial for improving patient survival, yet current diagnostic tools remain limited in their ability to comprehensively assess large, heterogeneous lesions at the cellular level in vivo. We introduce a compact and affordable AI-powered endomicroscope (PrecisionView) that overcomes fundamental trade-offs in resolution, field-of-view, and depth-of-field that constrain existing in vivo microscopy technologies. By enabling real-time, wide-area visualization of cellular morphology and microvasculature at the point of care, PrecisionView reveals focal microscopic features associated with anatomic structures or pathological abnormalities across extensive epithelial regions. PrecisionView has the potential to improve early cancer detection and expand access to high-quality cancer detection in both high- and low-resource clinical settings. In vivo microscopy (IVM) has shown great promise to improve early detection of epithelial precancer, but it suffers from fundamental trade-offs that limit the resolution, field-of-view (FOV) and depth-of-field (DOF). Here, we present PrecisionView, a compact, deep learning-enabled endomicroscope that breaks these constraints and achieves 20 mm2 FOV and 500 µm DOF with 4 µm resolution, representing approximately 5× increase in FOV and 8× larger DOF compared to conventional IVM with similar resolution. PrecisionView integrates a deep learning-optimized phase mask and real-time reconstruction, enabling rapid in vivo assessment of two key hallmarks of cancer: epithelial cell nuclear morphology and subsurface microvasculature through fluorescence and reflectance imaging. By imaging the oral cavity of healthy volunteers and cervical specimens with precancerous lesions, PrecisionView generates large-scale (1 to 3 cm2) coregistered maps of cellular and vascular structures, revealing distinct microscopic patterns associated with anatomic structures and precancerous lesions. Our results suggest the potential of this computational endomicroscope to address the unmet need for early cancer detection at the point of care.
Proceedings of the National Academy of Sciences , article en libre accès, 2026