Osteoporosis screening using areal bone mineral density estimation from diagnostic CT images

Area of Science:

• Radiology
• Osteoporosis Research
• Medical Imaging Analysis

Background:

• Aging populations are experiencing an increase in osteoporotic fractures.
• Diagnostic computed tomography (dCT) images may offer valuable densitometric data for osteoporosis screening.
• A reliable and cost-effective screening method is crucial for managing this growing health concern.

Purpose of the Study:

• To investigate the correlation between volumetric information from dCT and areal bone mineral density (aBMD).
• To assess the suitability of dCT imaging for developing an osteopenia screening system.
• To estimate aBMD and predict bone disease status from lumbar spine dCT images.

Main Methods:

• Dual-energy x-ray absorptiometry (DXA) and CT images were acquired from 44 male patients.
• aBMD from CT (aBMD(CT)) was calculated using Hounsfield units to bone density relationships.
• Statistical methods were employed to estimate DXA-derived aBMD (aBMD(DxA)) and diagnose osteopenia.

Main Results:

• A high correlation (r = 0.852) was found between aBMD(CT) and aBMD(DxA).
• The strongest correlation utilized all dCT slices and both bone components (trabecular and cortical).
• The osteopenia classifier achieved 80.1% accuracy and an AUC of 0.894.

Conclusions:

• Routine CT data can be used to determine aBMD(DxA).
• This method facilitates the development of a dCT imaging-based opportunistic screening system.
• The system aids in the detection and management of osteopenia.