|ECR 2019 / C-2065|
|Towards radiologist-level malignancy detection on chest CT scans: a comparative study of the performance of convolutional neural networks and four thoracic radiologists|
Aims and objectives
The demonstration of a 20% reduction in lung cancer mortality in the USA National Lung Screening Trial (NLST)  and the subsequent decision by the U.S. Centers for Medicare and Medicaid Services to provide Medicare coverage for lung cancer screening has paved the way for nationwide lung cancer screening in the USA. Additionally, results of the NELSON trial also confirmed the value of low-dose CT screening with decreased mortality by 26% in high-risk men and 61% in high-risk women over a 10-year period. 
Lung cancer screening programs have subsequently initiated big data analysis projects on chest CTs. The National Lung Screening Trial (NLST) dataset, in particular, has longitudinal data of high-risk patients to closely monitor potentially malignant lung nodules and provide the opportunity for the development of Computer-aided Detection (CADe) in detecting lung nodules and Computer-aided Diagnosis (CADx) systems in characterizing lung nodules to assist radiologists in reporting high volumes of chest CT scans.
The purpose of this work is to evaluate the performance of a deep learning system based on convolutional neural networks in predicting the presence of malignant lung nodules on chest CT scans. We also attempt to benchmark its performance against four radiologists.
Thematically related posters
ECR 2019 / C-3685
Evaluation of deep learning software tool for CT based lung nodule growth assessment
ECR 2019 / C-2695
Variability of coronary plaque burden evaluation by low-dose coronary CTA with iterative model reconstruction