Using deep learning algorithms for detecti̇on of lung cancer

Of around 163 500 new cases of cancer are diagnosed annually in Turkey. Lung cancer is the leading cause of cancer deaths in our country as in most developed countries. Therefore, successful and early diagnosis is vital. There are very few studies in the field of deep learning in the diagnosis of lung cancer, with the use of deep learning methods for the diagnosis of cancerous tissues on lung cancer CT images. For this reason, in the study, a deep learning model for lung cancer tumor tissue detection via PET / CT images was proposed. 1968 CT images of 38 different patients with benign and malignant tumors were used as the data set in the deep learning models. In the graduate study, deep learning models were used for classification before the diagnosis of lung cancer. The data set was first labeled according to the presence of benign and malignant tumors. Histogram thresholding method was applied to the CT images obtained from the patients. The histogram thresholding data were labeled as benign and malignant, 80% of which was divided into training and 20% of which were tested. When the proposed CNN deep learning model was applied to this separated data set, the accuracy rate was 98.48%. The accuracy of the data set in the DNN deep learning model was 93.65%. Compared to other studies in the literature, the proposed deep learning models have higher accuracy.