Category: Central DXA: (DXA, TBS)

28 - Feasibility Study of Quality Control Methodology for TBS

Objective: The aim of this study is to develop a TBS Quality Control (QC) methodology using the new TBS phantom.

Material &
TBS phantom data has been acquired on two GE-Lunar devices (iDXA and Prodigy) during 146 non-consecutive days over 11 months. The phantom is composed of a soft tissue part (14 cm thickness, 100% fat content) and a fractal metallic part which generates 4 different TBS values. During the first 45 days, two sets of 5 acquisitions of the phantom were done each day, with complete repositioning between the two sets. Over the following 9 months, a single scan was performed each day. We arbitrary selected data from the 10 first days to create TBS reference values for this QC. Phantom precision for TBS was computed for both devices. To develop a TBS QC, we have followed the methodology proposed by Lu et al. (JBMR, 1996); i.e. using Shewart 2 and Tabular CuSum techniques. TBS precision threshold for the QC (sdTh) was established from a performance target, computed from TBS L1L4 in-vivo reproducibility: sdTh = in-vivo TBSsd / 2.5. This target is designed to be similar to the BMD QC sensitivity, as 2.5 is the ratio between BMD phantom reproducibility and in-vivo L1L4 BMD reproducibility. TBS phantom precision was compared to our QC precision target.

Phantom precision error is lower than the precision target for both devices (0.007 and 0.003 for Prodigy and iDXA respectively, compared to our target of 0.008). No confirmed alarms were risen for either DXA device over the course of this study. Shewart and CUSUM graphs for Prodigy are presented in Fig. 1.

This study demonstrates the feasibility of developing a QC approach for TBS. These two devices provided stable measurements during the 11 months of follow-up. The phantom TBS precision is sufficient to perform a daily QC. The number of days required to create a stable reference value, and the minimum number scans each day have to be optimized. Further studies are needed to make sure this method can detect changes or drifts of the TBS values produced by DXA devices.

Neil Binkley

Professor of Medicine
University of Wisconsin School of Medicine and Public Health
Madison, Wisconsin

Diane Krueger

Research Program Manager
University of Wisconsin Osteoporosis Clinical Research Program
Madison, Wisconsin

Diane Krueger received her Bachelor of Science degree at the University of Wisconsin-Madison. She is an ISCD-certified clinical densitometrist and a certified clinical research coordinator through the Association of Clinical Research Professionals. She has been program manager of the University of Wisconsin Osteoporosis Clinical Research Program since its inception in 1993. Ms. Krueger has extensive clinical research experience in osteoporosis and bone densitometry, having coordinated multiple industry and investigator-initiated studies. In collaboration with the UW Osteoporosis Program, she has published over 70 manuscripts and authored or presented over 100 abstracts. Her service with ISCD has included chairing the Technologist Bone Densitometry Update and Annual Meeting Committees, participating on the Education Council, serving on the Executive Committee and Board as member and officer (Secretary and four Presidential seats). She has been Technologist Bone Densitometry Course faculty since 2006.

Franck Michelet

Medical imaging scientist
Medimaps SASU, Merignac, France
Merignac, Aquitaine, France