Development and validation of a radiomics nomogram for identifying invasiveness of pulmonary adenocarcinomas appearing as subcentimeter ground-glass opacity nodules. Issue 112 (March 2019)

• Record Type: Journal Article
• Authors: Zhao, Wei; Xu, Ya'nan; Yang, Zhiming; Sun, Yingli; Li, Cheng; Jin, Liang; Gao, Pan; He, Wenjie; Wang, Peijun; Shi, Hongli; Hua, Yanqing; Li, Ming
• Is Part Of: European journal of radiology. Issue 112(2019)
• Page Start: 161
• Page End: 168
• Publication Date: 2019-03
• Subjects: GGN ground-glass nodule -- LDCT low-dose computed tomography -- AAH atypical adenomatous hyperplasia -- AIS adenocarcinomas in situ -- MIA minimally invasive adenocarcinoma -- IAC invasive pulmonary adenocarcinoma -- CEA carcinoembryonic antigen -- SCC squamous cell carcinoma antigen -- VOI volume of interest -- ICC intraclass correlation coefficient -- LASSO least absolute shrinkage and selection operator -- DCA decision curve analysis -- VIF variance inflation factor; Radiomics -- Nomogram -- Subcentimeter -- Pulmonary ground-glass nodules; Medical radiology -- Periodicals; Radiology -- Periodicals; Radiologie médicale -- Périodiques; Medical radiology; Periodicals; 616.075705
• DOI: 10.1016/j.ejrad.2019.01.021
• Languages: English
• ISSNs: 0720-048X

Abstract

Highlights: Identifying the invasiveness of subcentimeter GGNs based on radiomics analysis. Constructing a radiomics signature to identify the invasiveness of subcentimeter GGNs. Building a radiomics nomogram to identify the invasiveness of subcentimeter GGNs. Abstract: The aim of the present study was to develop and validate a radiomics-based nomogram for differentiation of pre-invasive lesions from invasive lesions that appearing as ground-glass opacity nodules (GGNs) ≤10 mm (sub-centimeter) in diameter at CT. A total of 542 consecutive patients with 626 pathologically confirmed pulmonary subcentimeter GGNs were retrospectively studied from October 2011 to September 2017. All the GGNs were divided into a training set (n = 334) and a validation set (n = 292). Researchers extracted 475 radiomics features from the plain CT images; a radiomics signature was constructed with the least absolute shrinkage and selection operator (LASSO) based on multivariable regression in the training set. Based on the multivariable logistic regression model, a radiomics nomogram was developed in the training set. The performance of the nomogram was evaluated with respect to its calibration, discrimination, and clinical-utility and this was assessed in the validation set. The constructed radiomics signature, which consisted of 15 radiomics features, was significantly associated with the invasiveness of subcentimeter GGNs ( P < 0.0001 for both training set and validation set). To build the nomogram model, radiomics signature and mean CT value were used. The nomogram model demonstrated good discrimination and calibration in both training set (C-index, 0.716 [95% CI, 0.632 to 0.801]) and validation set (C-index, 0.707 [95% CI, 0.625 to 0.788]). Decision curve analysis (DCA) indicated that radiomics-based nomogram was clinically useful. A radiomics-based nomogram that incorporates both radiomics signature and mean CT value is constructed in the study, which can be conveniently used to facilitate the preoperative individualized prediction of the invasiveness in patients with subcentimeter GGNs.