Toward an Expert Level of Lung Cancer Detection and Classification Using a Deep Convolutional Neural Network

• DOI: 10.1634/theoncologist.2018-0908
• Publication Date: 2019-04-17T19:35:32Z
• Authors: Chao Zhang; Xing Sun; Kang Dang; Ke Li; Xiao-wei Guo; Jia Chang; Zong-qiao Yu; Fei-yue Huang; Yun-sheng Wu; Zhu Liang; Zai-yi Liu; Xue-gong Zhang; Xing-lin Gao; Shao-hong Huang; Jie Qin; Wei-neng Feng; Tao Zhou; Yan-bin Zhang; Wei-jun Fang; Ming-fang Zhao; Xue-ning Yang; Qing Zhou; Yi-long Wu; Wen-zhao Zhong
• Keywords: Nodule (geology); Clinical Practice

Abstract Background Computed tomography (CT) is essential for pulmonary nodule detection in diagnosing lung cancer. As deep learning algorithms have recently been regarded as a promising technique medical fields, we attempt to integrate well-trained algorithm detect and classify nodules derived from clinical CT images. Materials Methods Open-source data sets multicenter used this study. A three-dimensional convolutional neural network (CNN) was designed them into malignant or benign diseases based on pathologically laboratory proven results. Results The sensitivity specificity of model were found be 84.4% (95% confidence interval [CI], 80.5%–88.3%) 83.0% CI, 79.5%–86.5%), respectively. Subgroup analysis smaller (<10 mm) demonstrated remarkable specificity, similar that larger (10–30 mm). Additional validation implemented by comparing manual assessments done different ranks doctors with those performed CNN. results show the performance CNN superior assessment. Conclusion Under companion diagnostics, may assist radiologists future providing accurate timely information regular practices. Implications Practice described article both high classifying regardless diameters well superiority compared Although it still warrants further improvement screening cohorts, its application could definitely facilitate practice.