Assessment the Accuracy of Densitometry Measurements Using DMA PP2 Phantom

Abstract

Purpose of the study — to assess the accuracy of dual energy x-ray absorptiometry (DXA ) for measurements of mineral bone density, bone mineral content, area of selected spine zone of examination as well as impact of subcutaneous fat layer and correction of auto-segmenting of the spine on the mentioned parameters. Material and Methods. The study was performed on iDXA scanner using the designed phantom DMA PP2 of the lumber spine with inlays to simulate subcutaneous fat (SF). To ensure correct assessment of measurements (precision and accuracy) the authors performed fivefold repeated scanning. Two modifications of the phantom were used, with and without SF inlays, as well as two methods for selection of spine range for examination – automatic and correction of autosegmentation. Results. Scanning of the phantom without SF inlays demonstrated a systematic understated values of bone mineral density (BMD) and bone mineral content (BMC) along the full measured interval: mean relative error of BMD for L1-L4 interval was 10.62% with automatic segmentation and 7.43% — with  correction of autosegmentation. The least accuracy for BMD and BMC (1.53% and 0.90%, respectively) was observed during SF simulation and with correction of auto-segmentation of the spine. Analysis of variation coefficient for area of examined vertebrae, BMC and BMD demonstrated rather high precision of measurements, namely for BMD without SF in the L1-L4 interval amounted to 1.00% (auto-segmentation) and 0.56% (correction). Variation coefficient for scanning including SF inlays in the interval L1-L4 was 1.00% (auto-segmentation) and 0.68% (correction). Conclusion. The lowest level of accuracy was observed with the SFL object; in this case, the variation coefficient did not exceed 1% for all BMD interval. The mean value of the BMC accuracy also did not exceed 1% with the optimal scan parameters. The study proved the effectiveness of “RSK PK2” phantom when estimating the accuracy of BMD and BMC on iDXA scanner.