Development and validation of a neural network to detect and describe lung nodules
Principal Investigator
Name
Alexandre Compas
Degrees
MSc
Institution
Gleamer SAS
Position Title
Head of Product
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-878
Initial CDAS Request Approval
Jan 26, 2022
Title
Development and validation of a neural network to detect and describe lung nodules
Summary
Following the NLST, a lung screening program has been adopted in the USA. As a result, the number of pulmonary CT increased. However, thoracic pulmonary nodule detection is one of the most time-consuming components of the assessment of a chest CT examination. Therefore there is room for a computer-aided detection (CAD) software which could detect and assess pulmonary nodules. Indeed, such CAD could result both in reducing workload of the radiologists but more importantly in improved detection accuracy.
In order to create such software a large dataset of normal and pathological images is required to train the algorithm. This is why the lung CT image dataset from the NLST could be used to create and validate such an algorithm.
In order to create such software a large dataset of normal and pathological images is required to train the algorithm. This is why the lung CT image dataset from the NLST could be used to create and validate such an algorithm.
Aims
1) To develop a CAD software which is able to detect lung nodules (lung screening and incidental findings) and characterize them.
2) To outperform specificity and sensitivity of existing CAD solutions
Collaborators
all members of the Gleamer R&D team