免疫组化图像的计算机辅助定量分析研究

Immunohistochemistry ( IHC ) is a modern biological imaging technology, which plays an important role in early detection, early diagnosis and early treatment of tumors. Currently, pathologists usually resort to visual scoring to discriminate the positiveness of IHC. Vulnerable to the experience of operators, the subjectiveness and nonreproducibility of human visual assessment extremely affect the power of IHC in clinical applications .This calls for new computer-aided technologies to facilitate the quantification of IHC images. This project aims to provide new quantitative technologies for computer-aided discrimination of membrane staining IHC image. The main research contents are as follows: (1) Construct membrane IHC image models for automated segmentation combining membrane IHC protein expression and microscopic image processing technologies such as staining separation by color deconvolution, virtual membrane detection by Voronoi diagram prior model, and color edge detection.(2) Establish quantitative analysis model of membrane IHC image and quantitative information database based on parameters measurement and scoring function.(3) Evaluate and improve the automatic segmentation and quantification models. This project will provide key algorithms for the computer-aided diagnosis of membrane IHC and for the quantitative analyses of IHC images both theoretically and practically.

免疫组织化学（IHC）是生物成像技术的杰出代表，它对肿瘤的早期发现、早期诊断、早期治疗意义重大。目前临床上仍然采用视觉评分的手段判别IHC的阳性程度，其结果的主观性和操作的不可重复性限制了IHC的应用，因此计算机辅助IHC图像定量分析已成为业界关注的焦点。本项目旨在提出膜染IHC的阳性程度定量评分方法，主要研究内容包括：（1）根据膜染IHC的蛋白表达方式，深入研究基于颜色去卷积的染色分离技术、基于Voronoi图先验模型的虚拟膜检测方法、彩色边缘检测等显微图像处理技术，构建高精度的膜染IHC图像自动分割模型；（2）设计定量参数和评分函数，建立膜染IHC定量分析模型和定量信息数据库；（3）对膜染IHC的自动分割和定量分析模型进行质量评估和临床验证。本项目开展将为计算机辅助的膜染IHC诊断与量化分析提供理论支撑与关键算法。

免疫组化（IHC）是现代生物成像技术的杰出代表，它对肿瘤的早期发现、早期诊断、早期治疗意义重大。随着IHC定量标准化的制定和普及，计算机辅助IHC图像定量分析的重要性日益凸显。本项目提出结合抗体与DAB先验形状的IHC类型自动判别方法，组合染色分离、彩色边缘检测、区域分割方法，构建IHC图像定量分析模型，主要实现：（1）初步建立免疫组化定量信息数据库；（2）提出颜色滤波方法实现IHC图像的染色分离；（3）提出基于块色彩梯度和颜色熵的边缘检测方法，用于膜染IHC图像的DAB阳性产物分割提取；（4）提出基于色调惩罚的彩色聚类算法，用于核染IHC图像的DAB阳性产物分割；（5）给出定量评分函数，与视觉评分结果吻合度较好。本项目的研究成果为计算机辅助IHC诊断与量化分析提供了理论支撑与关键算法。