Review on Deep Learning Methods Used for Computer-aided Lung Cancer Detection and Diagnosis

Abstract

Cancer is one of the diseases with high mortality rates in the 21st century, with lung cancer being the first in all cancer morbidity and mortality rates. In recent years, with the large rise of data and artificial intelligence researches, the auxiliary diagnosis of lung cancer based on deep learning has gradually become a hot research topic. As the available and public datasets for lung cancer are mainly CT scans images with lung nodules annotations, the work on the assisted diagnosis of lung cancer using deep learning is mainly based on image data preprocessing, Pulmonary nodule segmentation, and lesion analysis and diagnosis. Computer-aided diagnosis (CAD) tools help radiologists to reduce diagnostic errors such as missing tumors and misdiagnosis. So our aim is to help the development of a new CAD system with higher performance than the existent ones to assist lung cancer detection in early stages. This paper presents an overview of the deep learning methods used for computer-aided lung cancer detection and diagnosis. It is mainly focused on the important processing and analyzing methods for the pulmonary image data obtained by medical instrument imaging, and which we can summarize into these 4 steps: Medical image data preprocessing, Pulmonary nodule segmentation, pulmonary nodule detection, and finally lesion diagnosis A full description of the CAD systems steps is given along with an overview of the state of art deep learning medical image processing methods.