K24 Mid-Career Investigator Award in Patient-Oriented Research

Medical University Of South Carolina, Charleston SC

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Career development goals: To provide sufficient time for mentoring and research activities. Long term research objectives: To study lung cancer screening and its implications for patients. To perform this research, the program will focus on better ways to differentiate benign from malignant nodules discovered on Computed Tomography (CT). This project has three components. The first is the development of improved Computer Aided Detection (CAD) software to aid in identifying lesions and differentiating between benign and malignant nodules based on characteristic CT findings. The second component will test the CAD software against independent readers in clinically meaningful scenarios. The third component will utilize clinical/radiologic findings from patients presenting with indeterminate CT detected nodules to produce a prediction model to improve the ability to differentiate benign from malignant disease. Methods: For project 1 the investigators will collect CT data on lesion phantoms and from patients with characteristic or known presentations. This will lead to the development of CT reconstruction techniques for the specific purposes of automated detections and analysis of small lung nodules. Then, calculation of the precision of nodule volume estimates as a function of the CT reconstruction techniques will be performed. For project 2, CAD can be tested by comparing reader confidence in CT findings before and after CAD is used as a second reader. For project 3 to develop a multivariate statistical model of malignancy potential, clinical, radiographic, and CAD data on patients with CT detected nodules will be collected. Multinomial logistic regression will be used to calculate odds ratios for malignancy. A subset of the patient data will be used as test data to validate the model. The test data will be disjoint from the data used to derive the prediction model. In lay terms, this project has important implications for public health because lung cancer is the leading cause of cancer death and a screening program could be very beneficial in reducing the number of deaths from this disease, this can only occur if there is a better way to differentiate the many benign from the few malignant lesions discovered on screening studies. This study proposes to accomplish that goal which will lead to less unnecessary tests and surgeries for non-cancerous lesions and thus improve the efficacy of the lung cancer screening test.