基于定量影像组学的肺肿瘤良恶性预测方法

发布时间：2018-05-13 22:43

本文选题：影像组学 + 肺癌　；参考：《自动化学报》2017年12期

【摘要】：肺癌是世界范围内致死率最高的癌症之一,肺肿瘤的良恶性诊断对于治疗方式选择意义重大.本文借助影像组学(Radiomics)方法利用LIDC(Lung imaging database consortium)肺癌公开数据库中619例病人的肺癌计算机断层(Computed tomography,CT)影像数据,分割出病变区域,并结合肿瘤医学特性和临床认知,提取反映肿瘤形状大小、强度和纹理特性的60个定量影像特征,然后利用支持向量机(Support vector machine,SVM)构建诊断肺肿瘤良恶性的预测模型,筛选出对诊断肺肿瘤良恶性有价值的20个影像组学特征.为肺肿瘤良恶性预测提供了一种非入侵的检测手段.随着CT影像在肺癌临床诊断中的广泛使用,应用样本量的不断增加,本文方法有望成为一种辅助诊断工具,有效提高临床肺肿瘤良恶性诊断准确率.
[Abstract]:Lung cancer is one of the most fatal cancers in the world. In this paper, by means of radiology), the computed tomographic CT images of 619 patients with lung cancer in the LIDC(Lung imaging database consortium) lung cancer open database were used to segment the lesion area, and combined with the characteristics of tumor medicine and clinical cognition. Sixty quantitative image features reflecting the shape, size, intensity and texture characteristics of lung tumors were extracted, and then a predictive model for the diagnosis of benign and malignant lung tumors was constructed by using support vector machine (SVM). Twenty imaging features were screened for diagnosis of benign and malignant lung tumors. It provides a non-invasive detection method for the prediction of benign and malignant lung tumors. With the extensive use of CT images in the clinical diagnosis of lung cancer and the increasing sample size, this method is expected to become an auxiliary diagnostic tool and to effectively improve the accuracy of diagnosis of benign and malignant lung tumors.
【作者单位】：哈尔滨理工大学自动化学院;中国科学院自动化研究所;东北大学中荷生物医学与信息工程学院;
【基金】：国家自然科学基金(81227901,81527805,61231004,81501616,81301346,61672197) 黑龙江省然科学基金(F201311,12541105) 中国科学院科技服务网络计划(KFJ-SW-STS-160);中国科学院科研设备项目(YZ201502)资助~~
【分类号】：R734.2;TP391.41

【相似文献】