Department of Radiology, The Third Xiangya Hospital of Central South University, Changsha 410013, China; Department of Radiology, The People's Hospital of Hunan Province, The First Hospital Affiliated of Hunan Normal University, Changsha 410005, China.
Department of Radiology, The Third Xiangya Hospital of Central South University, Changsha 410013, China.
Eur J Radiol. 2019 Sep;118:32-37. doi: 10.1016/j.ejrad.2019.06.025. Epub 2019 Jun 28.
To explore the feasibility and performance of machine learning-based radiomics classifier to predict the cell proliferation(Ki-67)in non-small cell lung cancer (NSCLC).
245 histopathological confirmed NSCLC patients who underwent CT scans were retrospectively included. The Ki-67 proliferation index (Ki-67 PI) were measured within 2 weeks after CT scans. A lesion volume of interest (VOI) was manually delineated and radiomics features were extracted by MaZda software from CT images. A random forest feature selection algorithm (RFFS) was used to reduce features. Six kinds of machine learning methods were used to establish radiomics classifiers, subjective imaging feature classifiers and combined classifiers, respectively. The performance of these classifiers was evaluated by the receiver operating characteristic curve (ROC) and compared with Delong test.
103 radiomics features were extracted and 20 optimal features were selected using RFFS. Among the radiomics classifiers established by six machine learning methods, random forest-based radiomics classifier achieved the best performance (AUC = 0.776) in predicting the Ki-67 expression level with sensitivity and specificity of 0.726 and 0.661, which was better than that of subjective imaging classifiers (AUC = 0.625, P < 0.05). However, the combined classifiers did not improve the predictive performance (AUC = 0.780, P > 0.05), with sensitivity and specificity of 0.752 and 0.633.
The machine learning-based CT radiomics classifier in NSCLC can facilitate the prediction of the expression level of Ki-67 and provide a novel non-invasive strategy for assessing the cell proliferation.
探索基于机器学习的放射组学分类器预测非小细胞肺癌（NSCLC）细胞增殖（Ki-67）的可行性和性能。
回顾性纳入 245 例经 CT 扫描证实的 NSCLC 患者。在 CT 扫描后 2 周内测量 Ki-67 增殖指数（Ki-67 PI）。通过 MaZda 软件手动勾画病变感兴趣区（VOI），并从 CT 图像中提取放射组学特征。采用随机森林特征选择算法（RFFS）进行特征降维。分别采用 6 种机器学习方法建立放射组学分类器、主观影像特征分类器和联合分类器。通过受试者工作特征曲线（ROC）评估这些分类器的性能，并与 Delong 检验进行比较。
共提取 103 个放射组学特征，通过 RFFS 选择 20 个最优特征。在 6 种机器学习方法建立的放射组学分类器中，基于随机森林的放射组学分类器在预测 Ki-67 表达水平方面表现最佳（AUC=0.776），其敏感性和特异性分别为 0.726 和 0.661，优于主观影像分类器（AUC=0.625，P<0.05）。然而，联合分类器并未提高预测性能（AUC=0.780，P>0.05），其敏感性和特异性分别为 0.752 和 0.633。
基于机器学习的 NSCLC CT 放射组学分类器有助于预测 Ki-67 的表达水平，为评估细胞增殖提供了一种新的非侵入性策略。
