CT-based radiomics signature for the stratification of N2 disease risk in clinical stage I lung adenocarcinoma

• DOI: 10.21037/tlcr.2019.11.18
• Publication Date: 2020-01-02T08:40:46Z
• Authors: Minglei Yang; Yunlang She; Jiajun Deng; Tingting Wang; Yijiu Ren; Hang Su; Junqi Wu; Xiwen Sun; Gening Jiang; Ke Fei; Lei Zhang; Dong Xie; Chang Chen
• Keywords: 03 medical and health sciences; 0302 clinical medicine

Risk stratification of N2 disease is vital for selecting candidates to receive invasive mediastinal staging modalities. In this study, we aimed stratify the risk metastasis in clinical stage I lung adenocarcinoma using radiomics analysis.Two datasets patients with who underwent resection were included (training dataset, 880; validation 322). Using PyRadiomics, 1,078 computed tomography (CT)-based features extracted after semi-automated nodule segmentation. order predict status, a signature was constructed optimal feature subset by sequentially applying minimum-redundancy-maximum-relevance and least absolute shrinkage selection operator (LASSO) techniques. Its performance validated dataset.The incidences 8.4% 7.1% training datasets, respectively. Unsupervised cluster analysis revealed that significantly correlated lymph node status pathological subtypes. For prediction, five selected establish signature, which showed better predictive than factors (P<0.001). The area under receiver operating characteristic curve 0.81 (0.77-0.86) 0.69 (0.63-0.75) factors, respectively, 0.82 (0.71-0.92) 0.64 (0.52-0.75), established CT-based could adenocarcinoma, thus assisting clinicians making patient-specific strategy.