Improving detection of lung cancer using radiomics and blood-based biomarkers
Principal Investigator
Name
Kate Bloch
Degrees
PhD
Institution
The University of Manchester
Position Title
Research Fellow
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-951
Initial CDAS Request Approval
Sep 6, 2022
Title
Improving detection of lung cancer using radiomics and blood-based biomarkers
Summary
With an increasing number of screening trials and the growing utilization of low dose CT screening, the detection of indeterminate and false positive pulmonary nodules are important clinical problems. A biomarker that will be able to improve discrimination between aggressive and benign nodules would help to accelerate diagnosis and reduce unnecessary and invasive procedures. In this study, we plan to assess the utility of quantitative radiomics features together with clinical patient metadata and blood-based biomarkers to predict malignancy. Here, we will analyse the NLST dataset of low-dose CT scans, including scans with benign, indeterminate and malignant nodules. Radiomics features will be extracted, and algorithms will be developed to analyse the association with benign and malignant status. The top features associated with malignancy will subsequently be applied to our early detection Manchester Health Check community-based pilot study1 with correlation to blood-based biomarker discovery projects in the CRUK Manchester Institute.
References
1. Crosbie PA, Balata H, Evison M, et al. Second round results from the Manchester ‘Lung Health Check’ community-based targeted lung cancer screening pilot, Thorax 2019;74:700-704.
References
1. Crosbie PA, Balata H, Evison M, et al. Second round results from the Manchester ‘Lung Health Check’ community-based targeted lung cancer screening pilot, Thorax 2019;74:700-704.
Aims
1. To extract quantitative image features from LDCT images of patients with confirmed lung nodule diagnosis from the NLST cohort and to identify top radiomics features that distinguish benign from malignant nodules.
2. To train the initial model on NLST dataset and evaluate the model detection performance on a held-out test set from the NLST and with our Manchester community-based screening study.
Collaborators
Prof. Marcel van Herk, University of Manchester, UK
Prof. Caroline Dive, CRUK Manchester Institute Cancer Biomarker Centre, University of Manchester, UK
Prof. Philip Crosbie, University of Manchester, UK