Value of pre-therapy 18F-FDG PET/CT radiomics in predicting EGFR mutation status in patients with non-small cell lung cancer

• Published in: European journal of nuclear medicine and molecular imaging Vol. 47; no. 5; pp. 1137 - 1146
• Main Authors: Zhang, Jianyuan, Zhao, Xinming, Zhao, Yan, Zhang, Jingmian, Zhang, Zhaoqi, Wang, Jianfang, Wang, Yingchen, Dai, Meng, Han, Jingya
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2020; Springer Nature B.V
• Subjects: Advanced Image Analyses (Radiomics and Artificial Intelligence); Algorithms; Cardiology; Computed tomography; Diagnostic systems; Epidermal growth factor; Epidermal growth factor receptors; Feature extraction; Fluorine isotopes; Growth factors; Heterogeneity; Imaging; Lung cancer; Medicine; Medicine & Public Health; Mutation; Nomograms; Non-small cell lung carcinoma; Nuclear Medicine; Oncology; Original Article; Orthopedics; Performance prediction; Phenotypes; Positron emission; Positron emission tomography; Radiology; Radiomics; Small cell lung carcinoma; Therapy; Tomography; Training; F-FDG; Epidermal growth factor receptor; Non-small cell lung cancer; PET/CT; Radiomics
• ISSN: 1619-7070; 1619-7089
• DOI: 10.1007/s00259-019-04592-1

Purpose To assess the predictive power of pre-therapy 18 F-FDG PET/CT-based radiomic features for epidermal growth factor receptor (EGFR) mutation status in non-small cell lung cancer. Methods Two hundred and forty-eight lung cancer patients underwent pre-therapy diagnostic 18 F-FDG PET/CT scans and were tested for genetic mutations. The LIFEx package was used to extract 47 PET and 45 CT radiomic features reflecting tumor heterogeneity and phenotype. The least absolute shrinkage and selection operator (LASSO) algorithm was used to select radiomic features and develop a radiomics signature. We compared the predictive performance of models established by radiomics signature, clinical variables, and their combinations using receiver operating curves (ROCs). In addition, a nomogram based on the radiomics signature score (rad-score) and clinical variables was developed. Results The patients were divided into a training set ( n = 175) and a validation set ( n = 73). Ten radiomic features were selected to build the radiomics signature model. The model showed a significant ability to discriminate between EGFR mutation and EGFR wild type, with area under the ROC curve (AUC) equal to 0.79 in the training set, and 0.85 in the validation set, compared with 0.75 and 0.69 for the clinical model. When clinical variables and radiomics signature were combined, the AUC increased to 0.86 (95% CI [0.80–0.91]) in the training set and 0.87 (95% CI [0.79–0.95]) in the validation set, thus showing better performance in the prediction of EGFR mutations. Conclusion The PET/CT-based radiomic features showed good performance in predicting EGFR mutation in non-small cell lung cancer, providing a useful method for the choice of targeted therapy in a clinical setting.