Noninvasive Characterization of Indeterminate Pulmonary Nodules Detected on Chest High-Resolution Computed Tomography
Abstract
Lung cancer remains a devastating disease that is often recognized late when the chances for a cure are limited. Fortunately, a recent large study, the National Lung Screening Trial (NLST), showed that annual high resolution computed tomography (HRCT) of the chest can detect lung cancer early and save lives. However, there are a number of open questions regarding this type of screening. Specifically, while HRCT identified suspicious lung nodules in approximately 40% of individuals, the vast majority of these nodules (over 95%) turned out to be benign (non-cancerous). The identification of a lung nodule on a screening HRCT typically triggers multiple follow-up CT (computed tomography) scans to monitor for changes in size or density. Alternatively, patients are evaluated with different types of scans such as positron emission tomography (PET) scans or invasive procedures such as tissue sampling of the nodule by a needle or surgical biopsy. Consequently, the diagnosis of a new pulmonary nodule typically results in significant patient anxiety, and the invasive procedures have the potential for physical harm and increased healthcare costs. As such, an objective diagnostic tool that could to determine accurately whether a lung nodule represents cancer or is benign at the time of its detection would be invaluable to physicians to individualize the care of these individuals. We, a multidisciplinary team of scientists and clinicians working at the Mayo Clinic and Brown University, are developing a new computer-aided diagnostic tool to distinguish benign from malignant lung nodules. This tool analyzes multiple HRCT parameters in a novel fashion. Our preliminary data are very encouraging. Herein we are proposing to use our computer-aided approach in conjunction with statistical modeling (combining the power of all promising parameters) to develop and validate a diagnostic tool that can reliably classify lung nodules as benign or malignant. We will develop the tool using data from the NLST (Aim 1). Afterwards, we will independently confirm (validate) our findings in a military/Veteran population (DECAMP-1 [Detection of Early lung Cancer Among Military Personnel Study 1] study) (Aim 2). "The stronger one s ability to see the end in the beginning, the easier it is to walk through the middle." If our research succeeds, patients with lesions suspicious for lung cancer on HRCT of the chest (as part of a screening strategy or not) could benefit from our imaging software analysis. Such analysis would allow the distinction between malignant and benign cancers, and the management could be individualized to each patient. Consequently, unnecessary imaging, invasive diagnostic procedures, and surgeries carrying substantial risks could be avoided in some of these patients. We estimate that our research could yield significant results within 2 years. We believe that funds awarded to pursue this research will refine, validate, and confirm the diagnostic and predictive power of this approach and will herald a major advance in providing optimal patient care and individualized medicine based on quantitative imaging.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 29, 2016
- Source ID
- W81XWH1510110
Entities
People
- Fabien Maldonado
Organizations
- United States Army
- Vanderbilt University