Hologic和Norland骨密度仪的精确度比较和数据换算

目的应用Hologic椎体体模和欧洲体模对不同骨密度仪进行横向及纵向评估,并应用 Shewhart规则进行仪器质量控制,同时比较Hologic及Norland骨密度仪活体精确度差异,探讨两者的数据换算关系。方法 Norland XR-36和Hologic Delphi A骨密度仪分别用欧洲体模连续扫描10次,每次扫描均应重新放置体模,用Shewhart规则对两台机器每日校准值进行质控。另外分别采用两套系统对30名志愿者的脊椎骨和左股骨的骨密度进行了两次重定位测量,计算精确度误差,通过F-检验比较差异。两台骨密度仪分别用Hologic腰椎体模每天测5 次,连续8 d,建立两仪器之间的数据换算关系。结果①Norland XR-36和Hologic Delphi A 骨密度仪用Shewhart规则监控均符合,Norhand XR-36连续3 m变异系数百分比为0．38％～ 0．53％,Hologic Delphi A连续3 m变异系数百分比为0．37％～0．41％,Norland-XR-36所测欧洲体模高、中、低为1．355,0．944,0．582,与真值相差9．6％,5．6％,16．4％;Hologic Delphi A高、中、低均值为1．423,0．940,0．534,与真值相差5．1％,6％,6．8％。②Hologic Delphi A的活体精密度高于Norland XR-36骨密度仪。③两仪器间BMD,BMC,Brea绝对值差异明显,但可用线性回归方程进行数据换算:BMD Hologic=0．970 Norland 0．026(r= 0．980;P<0．01);BMC Hologic=1．005 Norland 0．128(r=0．989;P<0．01);Brea Hologic=1．056 Norland-0．592(r=0．978;P<0．01)。结论以欧洲体模评估Norland XR-36和 Hologic Delphi A骨密度仪均有良好的精确度,但两仪器之间的数据不能直接互用,可用回归方程进行数据校正。质量控制是确保骨密度仪理想的精确度的必要措施。

Comparison of accuracy and precision between two dual-energy X-ray absorptiometers

Objective To assess the cross and longitudinal data from different bone densitometry equipments by Hologic spine phantom (HSP) and European spine phantom (ESP), and monitot the quality control of both with Shewhart rule. We compared the in-vivo precision with Hologic Delphi A and Norland XR-36 bone densitometer of Spine (L2-4) and femur BMD, and established the transcalibrating equations between them. Methods ESP and HSP was used on scanning the Norland XR-36 and Hologic Delphi A instruments, and every scan should replace phantom. Shewhart rule was used everyday to control the quality based on calibration value by themselves. Thirty people were measured twice with both systems at spine and left femur with repositioning between scans. Pooled precisions between the Hologic Delphi A and Norland XR-36 measurements were compared. Results The accuracy and precision of Hologic Delphi A was better than that of Norland XR-36. Significant differences (P< 0. 05) and positive correlation were found in the results of these two instruments. Conversions of values based on the following regression equations: BMD Hologic = 0. 970Norland 0. 026 (r=0. 980: P<0. 01); BMC Hologic=l. 005Norland 0. 128 (r=0.989:P<0. 01); Brea Hologic=1. 056 Norland-0. 592 (r= 0. 978 : P<0. 01) Conclusions Both Hologic Delphi A and Norland XR-36 all have good precision assessed by ESP. Value conversion based on the regression equations should be used for comparing data from one laboratory to another. Quality control is necessary to assure the precision and accuracy of any BMD equipments.