Noninvasive Characterization of Indeterminate Pulmonary Nodules Detected on Chest High-Resolution Computed Tomography
Abstract
The widespread implementation of lung cancer screening, following favorable results of the National Lung Screening Trial (NLST) and more recently the European NELSON trial, will likely continue to exacerbate the widespread clinical problem of indeterminate pulmonary nodules, which were detected in 40% of high-risk individuals screened by low dose high-resolution computed tomography (HRCT) in the NLST. Because 96% of these nodules were benign, the issue of diagnostic resolution of incidentally and screen-identified lung nodules will become increasingly important. Current clinical and radiological risk prediction models, which allow risk-stratification of patients and individualize management of pulmonary nodules, are commonly used, but remain suboptimal, and optimization of the clinical management of larger (> or = 7 mm) screen-detected nodules is urgently needed to avoid unnecessary diagnostic interventions leading to unwarranted mortality, morbidity and healthcare costs. In our project, we explore the utility of a conventional radiomic approach to the classification of screen-detected indeterminate nodules, leveraging unexploited large datasets contained on digital HRCT images to estimate the probability of malignancy based on selected predictive quantitative radiologic features.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 2019
- Accession Number
- AD1086022
Entities
People
- Fabien Maldonado
Organizations
- Vanderbilt University Medical Center