Radiomic Detection of EGFR Mutations in NSCLC

Abstract

Radiomics is defined as the use of automated or semi-automated post-processing and analysis of multiple features derived from imaging exams. Extracted features might generate models able to predict the molecular profile of solid tumors. The aim of this study was to develop a predictive algorithm to define the mutational status of EGFR in treatment-naïve patients with advanced non–small cell lung cancer (NSCLC). CT scans from 109 treatment-naïve patients with NSCLC (21 EGFR-mutant and 88 EGFR-wild type) underwent radiomics analysis to develop a machine learning model able to recognize EGFR-mutant from EGFR-WT patients via CT scans. A “test–retest” approach was used to identify stable radiomics features. The accuracy of the model was tested on an external validation set from another institution and on a dataset from the Cancer Imaging Archive (TCIA). The machine learning model that considered both radiomic and clinical features (gender and smoking status) reached a diagnostic accuracy of 88.1% in our dataset with an AUC at the ROC curve of 0.85, whereas the accuracy values in the datasets from TCIA and the external institution were 76.6% and 83.3%, respectively. Furthermore, 17 distinct radiomics features detected at baseline CT scan were associated with subsequent development of T790M during treatment with an EGFR inhibitor. In conclusion, our machine learning model was able to identify EGFR-mutant patients in multiple validation sets with globally good accuracy, especially after data optimization. More comprehensive training sets might result in further improvement of radiomics-based algorithms. Significance: These findings demonstrate that data normalization and “test–retest” methods might improve the performance of machine learning models on radiomics images and increase their reliability when used on external validation datasets.

The BRAF P.V600E Mutation Status of Melanoma Lung Metastases Cannot Be Discriminated on Computed Tomography by LIDC Criteria nor Radiomics Using Machine Learning

“…However, the variations in the acquisition protocols were small, making it unlikely this significantly affected the results of the current study. Feature selection methods based on feature test-retest reproducibility could be investigated in future work [52,53]. The difference in tube current between BRAF-mt and BRAF-wt almost reached statistical significance and could have been implicitly used by the model to distinguish these lesions.…”

CT-Based Peritumoral and Intratumoral Radiomics as Pretreatment Predictors of Atypical Responses to Immune Checkpoint Inhibitor Across Tumor Types: A Preliminary Multicenter Study

“…Of note, the glszm_ ZoneEntropy feature has been selected from both intra-and peritumoral regions in PsP vs. HPD, which was often appeared in tumor grading or staging and differentiation diagnosis (37,38). Gldm_SmallDependenceEmphasis was used for predicting the genetic mutation status in NSCLC patients (39). Moreover, firstorder and shape features quantify the range of gray values in the ROI which reflect the degree of heterogeneity of the tumor (40,41).…”

Artificial Intelligence Applications to Improve the Treatment of Locally Advanced Non-Small Cell Lung Cancers

“…Several attempts have been made to use radiomics signatures to predict EGFR status and other molecular targets in NSCLC patients [ 58 ]. For example, a ML model using CT radiomics and clinical features achieved a diagnostic accuracy of 88.3% in the external validation dataset for predicting EGFR mutant NSCLC [ 59 ]. Additionally, PET-imaging derived radiomic features have also been used to predict EGFR mutation status with accuracies around 75–78% [ 60 , 61 ].…”

“…The development of the T790M mutation in EGFR, which can occur during treatment with first-generation EGFR tyrosine kinase inhibitors (gefitinib and erlotinib) is an important mechanism of resistance. Radiomics signatures have also been developed to predict the development of this mutation [ 59 ].…”