Deriving Bone Density from Thoracic CT Scans

Bdi, Inc., Manhattan Beach CA

Investigators

Abstract

PROJECT SUMMARY/ABSTRACT As people age, they steadily lose bone mass, and their bones become fragile and increasingly prone to hip and other fractures, and this can prove fatal. The American College of Radiology therefore recommends that assessment of bone mineral density (BMD) be performed in post-menopausal women and all people over the age of 50 every two years. If indicated, preventive measures and medications can effectively decrease risk of fractures. BMD is usually measured by DXA (dual-energy x-ray absorptiometry). But very often, BMD is not assessed when indicated for a variety of reasons: 1) patients do not know they need BMD assessment; 2) health care providers have no systematic way to identify and notify patients who need BMD assessment; and 3) BMD assessment usually requires a separate referral and reimbursement. We have a solution that could help to address this need, while presenting an opportunity for both radiologists and our own company (BDI, INC) to profit by sharing reimbursement. We created, patented, and published a software solution that uses an algorithm to accurately measure BMD from previously acquired CT scans. Hence, if a patient is either scheduled to undergo a chest CT scan or has had a CT scan in the last 2 years, we can measure BMD and obviate the time, trouble, radiation exposure, and expense of a DXA scan, which would be a separate procedure. We plan to offer BMD measurement as an add-on service and share the CMS reimbursement of $144 with referring radiologists. Our primary target market will be individuals undergoing low-radiation lung CT scans (LCS) to assess risk of lung cancer. New expert guidelines mandate annual scans for current or former smokers, hence the volume of CT scans will increase by an estimated 11 million per year. However, our current algorithm was created using high-resolution cardiac CT scans. To similarly measure BMD from LCS used for lung cancer screening, we propose: 1) to obtain DXA scans and chest CT scans using parameters that are commonly used with low dose lung imaging; 2) develop algorithms appropriate to these acquisition protocols (Aim 1); 3) validate against DXA as we did previously with higher-resolution cardiac CT scans (Aim 2); and 4) determine retest reproducibility to ensure that our method can be used for serial assessment of BMD (Aim 3), both as a screening strategy and to evaluate the effectiveness of medications that inhibit bone loss. Recruitment and scanning will be performed by our LA BioMed subcontractor (Site PI: Dr. Matthew Budoff), who currently operate 3 CT scanners and a DXA scanner, and image 25 to 30 low-radiation CT scans per day, partly for research protocols. Our feasibility criteria are: 1) create an algorithm for LCS; 2) identify the 3 thoracic vertebrae that produce the best concordance with DXA; and 3) achieve retest reproducibility of 10% or less. A follow-on Phase II application will propose to validate both algorithms in a large, multi-ethnic cohort (the MESA Study) to ensure that our algorithms can accurately and reproducibly assess BMD from either high resolution or low-resolution serial chest CT scans in all ethnic groups, and in a wide range of body sizes.