Lung Nodule Classification on Computed Tomography Images Using Deep Learning

Lung Cancer is the most fast growing cancer around the world and is mostly diagnosed at an advanced stage. Due to enhancement in medical imaging modalities like Computed Tomography (CT) scans there is a need for computer aided detection system to classify the lung nodule into benign and malignant type with maximum accuracy to prevent delay in diagnosis. Many state-of-art methods used so far classify the images by applying machine learning algorithms on manually extracted features from imaging modalities. But in the recent years many deep learning techniques are being used in classification of lung nodule and have shown promising results when compared to other state-of-art methods. In this paper we have surveyed around 108 research papers to focus on the contribution of deep learning methodologies in detection of malignant tumor in Lung CT scan. This paper discusses variation applied on deep learning architecture to improve the accuracy of the classification system and a comprehensive comparison between various deep learning methods used so far for lung nodule classification. After reviewing each paper, this survey also presents challenges and opportunities in classifying lung nodule by using advanced deep learning strategies. The paper concludes with the need to address new issues in nodule classification with an aim to detect the malignant lesion at an early stage.