Optimization of DXA lumber spine bone imaging parameters based on Monte Carlo method and CT voxel phantom

Abstract

Objective To study the lumber spine imaging process of dual-energy X-ray absorptiometry (DXA) and parameters used to optimize the image quality. Methods A computational voxel phantom was constructed from patient computed tomography (CT) data. Using the Monte Carlo radiation transport method, a dual energy x-ray beam was simulated to scan the phantom of lumbar spine to generate a bone density image. The Figure of Merit (FOM) of each image was claculated. Parameters including the combination of the high and low energy tube voltage, the thickness of Cu filter, and the ratio of two beam energy incident photon number were optimized, which based on FOM. Results FOM reaches a minimum of 1.59 × 10-2 with the tube voltage combination of 75 and 200 kVp. With the thickness of the Cu filter from 0 mm to 3 mm, FOM decreases from 6.30×10-2 to 1.87×10-2, showing a gradually slow-down trend. With the incident photon number ratio (low energy/high energy) increasing from 1∶3 to 19∶1, FOM decreases firstly and then increases, reaching a minimum of 1.40×10-2 at 3∶1. Conclusions According to the simulation results, the combinations of low tube voltage from 70 kVp to 85 kVp and high tube voltage from 160 kVp to 200 kVp, 0.3 mm Cu filter and beam incident photon number ratio from 1 to 5 can yield the best lumbar spine image quality with the lowest patient dose. Key words: Monte Carlo; Computational voxel phantom; Dual-energy X-ray; Bone density imaging; Imaging parameters optimization