Personalizing screening intervals for lung cancer screening using artificial intelligence

Principal Investigator

Name
Scott Adams

Degrees
MD, PhD, FRCPC

Institution
University of Saskatchewan

Position Title
Assistant Professor

Email
scott.adams@usask.ca

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-1064

Initial CDAS Request Approval
May 30, 2023

Title
Personalizing screening intervals for lung cancer screening using artificial intelligence

Summary
Personalized screening intervals hold significant promise to increase the efficiency and cost-effectiveness of lung cancer screening programs. While various multivariate risk prediction models have been developed for lung cancer risk prediction, artificial intelligence (AI) may provide a level of comprehension that maximizes the number of lung cancers detected at an earlier stage while minimizing the number of delayed diagnoses. This project will compare a state-of-the-art AI model for lung cancer risk prediction using CT images to conventional risk prediction tools, such as PLCOm2012. Additionally, we will develop and assess the performance of an ensemble model which combines image and non-image inputs for lung cancer risk prediction, and assess its utility to inform personalized screening intervals.

Aims

1. Compare the performance of a state-of-the-art AI lung cancer risk prediction model to traditional risk prediction models such as PLCOm2012;

2. Develop and evaluate an ensemble AI model which combines image and non-image inputs for lung cancer risk prediction; and

3. Assess the performance of an AI model to stratify individuals for annual vs. biannual screening.

Collaborators

Dr. Scott J. Adams - University of Saskatchewan
Evan Seebach - University of Saskatchewan
Samuel Boctor - University of Saskatchewan