基于纳米等离激元芯片彩色显微成像和人工智能图像分析的肿瘤特征分子功能可视化检测

等离激元共振（SPR）生物传感器可以实时、无需标记、高分辨率以及无背景干扰的动态测量生物分子如肿瘤标志物蛋白分子间的互相作用过程，并已经作为重要的科学实验手段广泛应用于肿瘤药物筛选，免疫治疗和肿瘤细胞通路分子互作基础研究中。然而由于受到常规SPR成像技术的制约限制了其在实际生物传感应用中的灵敏度、特异性和重复性。本项目将研究基于三维多层复合材料纳米结构SPR芯片和显微成像分析的新器件理论和实验检测方法，来实现性能增强的表面肿瘤蛋白分子互作的动态成像检测和数据分析。本项目将通过结合新型三维纳米等离激元与光学谐振耦合器件，通用明场显微彩色成像，数字化动态图像处理和人工智能机器学习图像分析相结合的方法和理论来进行并行多重肿瘤蛋白互作的动态定量可视化检测和深度的信息挖掘，从而达到在通用光学测量仪器上大幅度（几个数量级）的提高SPR生物分子动态成像灵敏度和准确性的目标。

Owing to the unique capability of real-time, label-free, high-resolution, and localized detections in measuring the kinetic interaction of biomolecules such as cancer biomarker proteins, surface plasmon resonance (SPR) sensing and imaging, as a powerful scientific tool, has been widely applied in cancer drug discovery and cancer biomarker studies. However conventional SPR detections including localized SPR sensors have moderate sensitivity, specificity and repeatability which prevent practical cancer molecular sensing applications. In this project we will study new theories and experimental methods for 3D nanoplasmonic devices and microscopy imaging analysis, in order to achieve enhanced dynamic detection and analysis of cancer protein interaction kinetics. By combining novel nanoplasmonic-photonic resonance coupling, generic bright-field microscopy color imaging, digital dynamic image analysis and artificial intelligence machine learning, the proposed methods permit the deep information mining of multiplexed cancer protein binding kinetics on device surface and enable significant improvement (a few orders of magnitude) in SPR biodetection sensitivity and accuracy on ubiquitous generic microscopy systems.