Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Deep Learning Model’s Potential to Increase the Accuracy of Lung Cancer Screening

Automated prediction of malignancy risk of pulmonary nodules in chest CT scan
22 May 2019
Staging and Imaging;  Cancer Prevention
Thoracic Malignancies

A group of US authors from the Google AI, Stanford Health Care and Palo Alto Veterans Affairs, Northwestern Medicine in Chicago, and New York University-Langone Medical Center, reported in a letter published on 20 May 2019 in the Nature Medicine that a convolutional neural network performs automated prediction of malignancy risk of pulmonary nodules in chest computed tomography (CT) scan volumes and improves accuracy of lung cancer screening.

The authors wrote in the study background that lung cancer screening using low-dose CT has been shown to reduce mortality by 20–43% and is now included in US screening guidelines. However, existing challenges include intergrader variability and high false-positive and false-negative rates.

The authors propose in their article a deep learning algorithm that uses a patient’s current and prior CT volumes to predict the risk of lung cancer. Their model achieves a state-of-the-art performance, in particular 94.4% area under the curve, on 6716 National Lung Cancer Screening Trial (NLST) cases and performs similarly on an independent clinical validation set of 1139 cases.

The investigators conducted two reader studies. When prior CT imaging was not available, the model outperformed all six radiologists with absolute reductions of 11% in false positives and 5% in false negatives; and when prior CT imaging was available, the model performance was on-par with the same radiologists.

The authors concluded that their findings create an opportunity to optimise the screening process via computer assistance and automation. While the vast majority of patients remain unscreened, the authors draw attention on potentials for deep learning models to increase the accuracy, consistency and adoption of lung cancer screening worldwide.

This study used three datasets that are publicly available: LUng Nodule AnalysisThe Lung Image Database Consortium image collection (LIDC-IDRI) and NLST. The dataset from Northwestern Medicine was used under license for the current study, and is not publicly available.

The authors acknowledge the US NCI and the Foundation for the National Institutes of Health for their critical roles in the creation of the free publicly available LIDC-IDRI/NLST Database used in this study. The authors thank the NCI for access to NCI data collected by the NLST.

The study was funded by Google Inc. 

 

Reference

Ardila D, Kiraly AP, Bharadwaj S, et al. End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography. Nature Medicine; Published online 20 May 2019. doi: 10.1038/s41591-019-0447-x.

Last update: 22 May 2019

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.