Detection of lung cancer from computed tomography (CT) scans using artificial intelligence techniques

Lung cancer stands out as a deadly disease with high mortality rates worldwide. Early diagnosis is critical for effective treatment of the disease. This study investigates the role of artificial intelligence techniques in early diagnosis of lung cancer and highlights the advantages they provide. In this context, it is predicted that AI-assisted diagnostic systems can significantly improve the methods in lung cancer diagnosis by reducing the workload of radiologists and increasing accuracy rates. In this study, results obtained in lung cancer diagnosis using Convolutional Neural Networks (CNN) and YOLO algorithm have been evaluated. The success of both CNN and YOLO model on medical images was compared and analyzed over 6 different datasets. Among all datasets, the highest F1-Score for the CNN model was obtained on the dataset that was subjected to histogram equalization and contrast limited adaptive histogram equalization (CLAHE) with 99.85%. For the YOLO model, the highest F1-Score was obtained at 96% on the dataset where the same preprocessing technique was applied.