Advancements in Cancer Detection Using Machine Learning Technique: A Systematic Review of Decades, Comparisons, and Challenges

Lung cancer is a type of dangerous cancer and it is difficult to detect. It usually causes death for both men and women, so it is more necessary to take care to immediately and properly examine the nodules. Accordingly, several techniques have been implemented to detect lung cancer in its early stages. In this paper, a comparative analysis of different machine learning-based techniques for lung cancer detection was presented. In recent years, too many methods have been developed to diagnose lung cancer, most of them use CT images and some of them use X-rays. In addition, several classification methods are paired with many segmentation algorithms to use image recognition to identify lung cancer nodules. From this study, it was found that CT images are more suitable for accurate results. Therefore, CT scans are mostly used to detect cancer. Also, marker-driven watershed segmentation provides more accurate results than other segmentation techniques. In addition, the results obtained from methods based on deep learning techniques achieved higher accuracy than methods that were implemented using classical machine learning techniques.