Development of CADx for pulmonary nodules
Principal Investigator
Name
Lyndsey Pickup
Degrees
MEng, DPhil
Institution
Optellum Ltd.
Position Title
Senior Research Scientist
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-251
Initial CDAS Request Approval
Oct 13, 2016
Title
Development of CADx for pulmonary nodules
Summary
With the growing interest in Lung Cancer Screening in the US and beyond, the volume of chest CT images is growing fast. Being able to handle the reading and reporting for these scans in an efficient and consistent way is therefor a major concern for care providers.
Optellum is a new UK-based company whose vision is to enable cancer to be diagnosed earlier and with more confidence. Our team consists of award-winning machine learning and imaging experts who met at Oxford (UK), and we have a track record of bringing innovation into clinical use. Our PI, who joined Optellum in 2016, was previously the technical lead on the project that won the NIH's 2015 LUNGx challenge, and has worked extensively with NLST data previously.
We are currently developing deep learning solutions for the characterisation and stratification of pulmonary nodules in CT, and our performance to date matches that seen in state-of-the-art systems. We currently have access to the Oxford Lung Nodule database, which is composed primarily of nodules from screening and oncology follow-up, and we now seek access to the NLST data to complement this in our ongoing research and development work.
Optellum is a new UK-based company whose vision is to enable cancer to be diagnosed earlier and with more confidence. Our team consists of award-winning machine learning and imaging experts who met at Oxford (UK), and we have a track record of bringing innovation into clinical use. Our PI, who joined Optellum in 2016, was previously the technical lead on the project that won the NIH's 2015 LUNGx challenge, and has worked extensively with NLST data previously.
We are currently developing deep learning solutions for the characterisation and stratification of pulmonary nodules in CT, and our performance to date matches that seen in state-of-the-art systems. We currently have access to the Oxford Lung Nodule database, which is composed primarily of nodules from screening and oncology follow-up, and we now seek access to the NLST data to complement this in our ongoing research and development work.
Aims
1) Validate Optellum's current pulmonary nodule software on screening data.
2) Develop and test new models for pulmonary nodule analysis and stratification on screening data.
3) Improve Optellum's robustness to variations in scanner manufacturer, kernel type, patient geographic location etc, since differences in these factors are well-represented in the NLST data.
Collaborators
Dr. Vaclav Potesil, Optellum Ltd.
Related Publications
-
Developing an understanding of artificial intelligence lung nodule risk prediction using insights from the Brock model.
Chetan MR, Dowson N, Price NW, Ather S, Nicolson A, Gleeson FV
Eur Radiol. 2022 Mar 3 PUBMED -
Lung cancer prediction by Deep Learning to identify benign lung nodules.
Heuvelmans MA, van Ooijen PMA, Ather S, Silva CF, Han D, Heussel CP, Hickes W, Kauczor HU, Novotny P, Peschl H, Rook M, Rubtsov R, von Stackelberg O, Tsakok MT, Arteta C, Declerck J, Kadir T, Pickup L, Gleeson F, Oudkerk M
Lung Cancer. 2021 Jan 31; Volume 154: Pages 1-4 PUBMED -
Assessing the Accuracy of a Deep Learning Method to Risk Stratify Indeterminate Pulmonary Nodules.
Massion PP, Antic S, Ather S, Arteta C, Brabec J, Chen H, Declerck J, Dufek D, Hickes W, Kadir T, Kunst J, Landman BA, Munden RF, Novotny P, Peschl H, Pickup LC, Santos C, Smith GT, Talwar A, Gleeson F
Am. J. Respir. Crit. Care Med. 2020 Apr 24 PUBMED