The Identification of EGFR Mutation Based on the CT Image of Lung Adenocarcinoma

Abstract

Abstract-Lung adenocarcinoma is the most common cause of lung cancer in Taiwan. However, it is difficult to detect lung adenocarcinoma at early stage. Most radiologists usually decide the treatment strategies by screening whether the patients have the epidermal growth factor receptor (EGFR) gene mutations or not. However, the screen is expensive and time-consuming. Therefore, the aim of the present paper is to identify whether there are EGFR mutations in tumors with CT image. We segment the whole region of lung adenocarcinoma and extract the texture features in spatial domain and frequency domain using GLCM and Gabor Wavelet. Finally we use Neuro-fuzzy based AdaBoost to classify and identify the relationship between the texture features of lung adenocarcinoma CT images and EGFR mutations. The experimental result in this paper shows that our success rate was 92%, and we will improve our methods to increase the recognition rate in the future.