Using deep learning approaches for detection lung cancer

Cancer is one of the leading causes of death and poor quality of life in all countries in the world. According to the estimations made by the World health organization in 2019, cancer ranks first or second among the causes of death over the age of 70 in 112 of 183 countries, while it ranks third or fourth as the cause of death in 23 countries. The decline in the rates of deaths from coronary heart disease and stroke has led to an increase in the importance of cancer. With 2.2 million new diagnoses (about 11.4% among all cancers) and 1.8 million deaths in 2020, lung cancer was the second most diagnosed cancer and the first cause of death. Medical imaging techniques are of great importance in the diagnosis and treatment process of lung cancer. Computed tomography, chest X-ray, magnetic resonance imaging, molecular imaging techniques and positron emission tomography play an important role in detecting lung cancer. Among the biggest shortcomings of these techniques used are problems such as unsuitability for patients with other pathological conditions and inability to classify cancer type. New approaches are needed for early diagnosis and classification for this cancer. Deep learning method has started to be used more and more as one of the topics that has an increasing value in the field of imaging in medicine. The development of medical devices including artificial intelligence and deep learning is extremely important for doctors to make clearer and faster decisions for lung cancer. In this study, VGG16, VGG19 and Xception methods were used for detection and classification of lung cancer. In studies using deep learning-based methods, 94.19%, 95.24% and 90.14% success rates were obtained with the proposed approach and VGG16, VGG19 and Xception methods.