Evaluating whether 3D chest extrapulmonary musculature measurements improve prediction performance relative to conventional risk factors for all-cause mortality in a lung cancer screening cohort.
Principal Investigator
Name
Miranda Kirby
Degrees
Ph.D
Institution
Toronto Metropolitain University
Position Title
Associate Professor
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1370
Initial CDAS Request Approval
Dec 16, 2024
Title
Evaluating whether 3D chest extrapulmonary musculature measurements improve prediction performance relative to conventional risk factors for all-cause mortality in a lung cancer screening cohort.
Summary
Recent research has shown that automated deep learning algorithms can analyze skeletal muscle on chest CT scans to improve predictions of lung cancer death and overall mortality [1]. However, these studies typically rely on only a few CT slices and consider the entire muscle area, without distinguishing between individual muscle groups. This study aims to enhance these findings by using full 3D measurements of different muscle groups from chest CT scans to better predict risks for lung cancer, cardiovascular disease (CVD), and all-cause mortality. This study will also explore whether these 3D measurements offer more predictive power than the traditional 2D muscle measures and other known CT-based markers for lung disease, like emphysema and airway thickness [1]. Methods: This project will reanalyze data from the National Lung Screening Trial (NLST), which involved over 26,000 individuals aged 55-74 with a significant smoking history [2]. Among these, around 25,000 participants will be selected based on specific criteria, such as scan quality and length. The previously developed 2D skeletal muscle measurements will be compared with new 3D muscle measurements derived from the entire CT scan. The study will also assess CT-based indicators of lung damage: emphysema (measured by low-density areas in the lungs) and airway disease (measured by the thickness of airway walls). Analysis: Sex-specific models will be created to predict the risks of lung cancer, CVD, and mortality, both with and without the new 3D muscle data. By comparing these models using statistical methods, the added value of including 3D muscle measurements will be determined. It is anticipated that 3D muscle measurements will independently predict a lower risk of lung cancer, CVD, and overall death. These results are expected to outperform models that use only 2D muscle measurements or lung disease markers. The study will highlight the potential of using routine CT scans to gather additional information that can aid in predicting and preventing serious diseases, improving outcomes for at-risk populations like smokers.
Aims
Aim 1: Assess whether the 3D muscle group measurements improve the prediction of lung cancer, CVD, and overall mortality when used alongside established CT-based lung disease markers, such as emphysema and airway wall thickness.
Aim 2: Develop and compare sex-specific risk models for lung cancer, CVD, and mortality that include both 3D muscle measurements and traditional lung disease markers to determine whether anatomical differences between males and females affect the predictive accuracy.
Aim 3: Investigate whether 3D muscle measurements provide independent predictive value for decreased risk of lung cancer, CVD, and all-cause mortality across an at-risk population, with the goal of enhancing disease prevention strategies.
Collaborators
N/A