胸部CT图像中孤立性肺结节良恶性快速分类

Fast classification of benign and malignant solitary pulmonary nodules in CT image

为突破医学影像诊断学依据医学征象进行定性诊断准确度不高的瓶颈,针对胸部CT图像中孤立性肺结节(SPN)定性诊断问题,提出了能够用图像特征有效表示SPN病理特性,快速准确诊断SPN良恶性的计算机辅助诊断系统。采取交互式分割方法从胸部CT图像中提取出SPN;直接计算SPN图像的多分辨率直方图得到768维空间信息特征样本集;然后,充分利用具有处理高维数据集优势的支持向量机(SVM)构造SPN良恶性分类器;最后,通过测试样本集对经训练后的SVM分类器进行测试以评价分类性能。对214例病例进行实验,结果表明:240个SPN图像的768维特征计算所用时间为4.83s,SVM分类器训练测试所用时间为2.24s,敏感性为73.33%,特异性为70%,准确度达71.67%,接受者操作特性曲线(ROC)下面积(AUC)为0.7864。该系统提取的高维图像空间信息特征能够有效表示SPN特性;没有考虑医学征象进行SPN定性诊断的准确度即可达到71.67%,同时分类速度比传统纹理算法提高了近50倍,为医学影像学解决SPN定性诊断问题提供了便捷、客观的辅助手段。

In order to improve the accuracy of the diagnosis with medical signs in medical imaging diagnostics,a computer-aided diagnosis system is developed quickly and accurately to find out the differentce of benign and malignant Solitary Pulmonary Nodwles（SPNs） in chest CT images based the image features of SPNs. Firstly, SPNs are extracted from chest CT images using the interactive segmentation,and the multi-resolution histograms of SPNs are directly calculated to receive a high-dimensional feature sample set with spatial information of SPNs. Then, the classifier for differentiating benign and malignant SPN is constructed by using a Support Vector Mechine（SVM）. Finally, the performance of classification is evaluated by testing the trained SVM with the test sample set. The test results of 214 eases show that it takes 4.83 s for computing 768 dimensional features of 240 SPNs and 2.24 s for training and testing the SVM classifier. The receiver operating characteristic （ROC） analysis of classification performance of the proposed approach shows that the sensitivity is 73.33%, specificity is 70%,accuracy is 71.67% ,and the area under curve （AUC） is nearly 0. 786 4. Obtained results show that the image spatial information can effectively express the characteristics of SPNs. The system classification accuracy of benign and malignant SPNs is up to 71.67% without medical signs, and the classification speed is about 50 times faster than that of traditional texture methods. It provides a feasible, simple and objective method for solving the problem in medical imaging diagnosis of the SPNs.