Detailed identification of epidermal growth factor receptor mutations in lung adenocarcinoma: Combining radiomics with machine learning

Abstract

Purpose To investigate the use of radiomics in the in‐depth identification of epidermal growth factor receptor (EGFR) mutation status in patients with lung adenocarcinoma. Methods Computed tomography images of 438 patients with lung adenocarcinoma were collected in two different institutions, and 496 radiomic features were extracted. In the training set, lasso logistic regression was used to establish radiomic signatures. Combining radiomic index and clinical features, five machine learning methods, and a tenfold cross‐validation strategy were used to establish combined models for EGFR+ vs EGFR−, and 19Del vs L858R, groups. The predictive power of the models was then evaluated using an independent external validation cohort. Results In the EGFR+ vs EGFR− and 19Del vs L858R groups, radiomic signatures consisting of 12 and 7 radiomic features were established, respectively; the area under the curves (AUCs) of the lasso logistic regression model on the validation set was 0.76 and 0.71, respectively. After inclusion of the clinical features, the maximum AUC of combined models on the validation set was 0.79 and 0.74, respectively. Logistic regression analysis showed good performance in the two groups, with AUCs of 0.79 and 0.71 on the validation set. Additionally, the AUC of combined models in the EGFR+ vs EGFR− group was higher than that of the 19Del vs L858R group. Conclusions Our study shows the potential of radiomics to predict EGFR mutation status. There are imaging phenotypic differences between EGFR+ and EGFR−, and between 19Del and L858R; these can be used to allow patients with lung adenocarcinoma to choose more appropriate and personalized treatment options.

Deep learning predicts epidermal growth factor receptor mutation subtypes in lung adenocarcinoma

“…[15][16][17]20 Li et al first predicted the overall EGFR mutation status (wild-type vs. 19DEL+L858R), and then distinguished between EGFR 19DEL and L858R (19DEL vs. L858R), with AUCs of 0.79 and 0.74,respectively. 16 Liu et alused radiomics features to predict mutation statuses of EGFR (wild-type vs. 19DEL+L858R), 19DEL (19Del vs. wild-type+L858R), and L858R (L858R vs. wild-type+19Del), with AUCs of 0.76, 0.70, and 0.66, respectively. 17 In this study, we also attempted to use radiomics to predict EGFR mutation subtypes.…”

Molecular typing of lung adenocarcinoma with computed tomography and CT image-based radiomics: a narrative review of research progress and prospects

“…Moreover, CT image-based radiomics has the capacity to distinguish EGFR subtype mutation exon 19 deletion and exon 21 L858R substitution ( 64 , 68 ). In some studies, the predictive performance of CT image-based radiomics in the identification of EGFR mutations in lung adenocarcinoma was better than that for EGFR subtype mutations ( 64 , 72 ), which may be due to the inclusion of clinical variables in the EGFR mutation groups. In another study, after deep learning of CT image-based radiomic features, the prediction model recognized EGFR mutation status with AUCs of 0.910 and 0.841 for the internal and external test cohorts, respectively ( 73 ), with outstanding performance.…”

“…However, it is worth noting that not all prediction signature models of CT image-based radiomics show particularly good performance. In some studies, the prediction performance of EGFR mutation status is not ideal with CT image-based radiomics alone ( 66 , 69 , 74 ) and the AUC values of some articles are lower than 0.8 ( 64 , 69 , 72 , 75 ), which may be related to the fact that the images were not preprocessed before data extraction. It is consistent that the predictive efficiency for EGFR mutation of the radiomic features combined with the clinical features has been improved in lung adenocarcinoma.…”

Machine Learning-Based Radiomics for Prediction of Epidermal Growth Factor Receptor Mutations in Lung Adenocarcinoma

“…Studies have revealed that somatic mutations, which ultimately lead to tumor phenotype, can be predicted by radiomics in different solid tumors, including lung cancer [ 10 , 13 ]. Based on imaging information extracted from magnetic resonance imaging (MRI), computed tomography (CT), and positron-emission-tomography (PET), radiomics analysis can be performed to identify the presence of EGFR, anaplastic lymphoma kinase (ALK), Kirsten rat sarcoma viral oncogene (KRAS), and Erb-B2 receptor tyrosine kinase 2 (ERBB2) mutations in patients with non-small-cell lung cancer (NSCLC) [ 14 – 18 ]. With specific regard to EGFR mutation, previous studies have documented the potential for radiomics to predict EGFR 19Del and L858R based on the phenotypic appearance [ 14 , 16 , 19 ].…”

“…Based on imaging information extracted from magnetic resonance imaging (MRI), computed tomography (CT), and positron-emission-tomography (PET), radiomics analysis can be performed to identify the presence of EGFR, anaplastic lymphoma kinase (ALK), Kirsten rat sarcoma viral oncogene (KRAS), and Erb-B2 receptor tyrosine kinase 2 (ERBB2) mutations in patients with non-small-cell lung cancer (NSCLC) [ 14 – 18 ]. With specific regard to EGFR mutation, previous studies have documented the potential for radiomics to predict EGFR 19Del and L858R based on the phenotypic appearance [ 14 , 16 , 19 ]. For example, Rossi et al built a machine learning (ML) model to identify EGFR mutant and achieved an area under the receiver operating characteristic curve (AUC) of 0.89 [ 19 ].…”