传统的宫颈癌筛查方法主要有TBS分类筛查法和基于细胞DNA的定量分析筛查法, TBS分类筛查法诊断率高但需要经验丰富的医生参与且敏感度低, 难以实现宫颈癌早期筛查; 基于细胞DNA的定量分析筛查法仅染色细胞核, 实现了定量化与自动化分析, 敏感度高但特异性差。 因此实现宫颈癌TBS与DNA定量分析联合筛查十分必要, 但目前TBS与DNA定量分析联合筛查方法均是使用两张不同的细胞涂片分别进行, 费时费力费材极为不便, 国内外还没有在一张宫颈细胞涂片上联合使用这两种方法的宫颈癌筛查方法。 为此提出了一种在一张细胞涂片上同时使用TBS分类法和细胞DNA定量分析法的宫颈癌筛查方法: 在同一张细胞涂片上对细胞进行巴氏染色和Feulgen染色, 为解决多重染色带来的DNA物质的吸光度干扰问题, 建立一套多光谱成像系统, 并提出基于线性多元回归的DNA吸光度剥离模型, 通过该模型解算出DNA物质的真实吸光度从而实现DNA物质的定量分析; 选取接近RGB波长的3个波段细胞图像合成伪彩色图像进行TBS分类筛查。 实现了一张细胞涂片上的宫颈癌TBS与DNA定量分析联合筛查, DNA定量分析模型稳定度高、 误差小, 诊断率高, 用于TBS筛查的伪彩色图像颜色明亮、 细胞核清晰可见、 细胞质边缘明确, 该方法在宫颈癌的诊断和筛查中具有很强的实用性。Technology

The traditional screening method of cervical cancer mainly involves TBS classification method and quantitative analysis method based on DNA, TBS screening method which has high diagnostic rate, but it needs experienced doctors to participate in the process with low sensitivity. Therefore, it is difficult to achieve early cervical cancer screening; cell’s DNA quantitative analysis, only stained the nucleus, achieving a quantitative and automated analysis. Even with high sensitivity, the specificity is poor. So it is extremely necessary to realize the combination of screening for TBS and cell’s DNA quantitative analysis, but the current TBS and DNA quantitative analysis combined screening method for the use of two different cell smear, time-consuming, laborious and very inconvenient, there is no screening method for cervical cancer on a combination of TBS and DNA quantitative analysis at home and abroad. This paper presents a method using TBS classification and DNA quantitative analysis on the same cell smear which was stained with Pap and Feulgen in order to solve the problem of the interference of the absorbance of DNA substance caused by multiple staining. a set of multi spectral imaging system and DNA absorbance peeling model are established based on linear multiple regression. With model solution, the real absorbance of the substance DNA was calculated, and the quantitative analysis of DNA was carried out. A pseudo color image is synthesized from 3 - band cell images with the close wavelength of RGB for TBS classification, so the organic combination of TBS and cell’s DNA quantitative analysis is realized. Experiments show that the DNA quantitative analysis model of this method is stable, with small error and high diagnostic rate due to the fact that the pseudo color images used for TBS screening were bright, clear, and clear cytoplasm. Therefore, this method is very useful in the diagnosis and screening of cervical cancer.