Abstract
Introduction: Dual Energy x-ray absorptiometery (DXA) is a useful tool for the assessment of peri-prosthetic bone mineral changes following total joint replacement. In order to assess these changes the precision of the DXA technique must be optimised. While patient positioning is an important factor, the role of the analysis software should also be considered. We developed and applied a new image analysis method to data from the EXPERT-XL fan beam densitometer (LUNAR GE, USA) aiming to improve reproducibility of bone region and tissue type determination by the analysis software.
Methods: 60 patients with cemented THR received repeat same day DXA examinations. These were initially analysed strictly according to the manufacturer’s femoral peri-prosthetic protocol. A modification of this protocol was attempted allowing further small corrections to the tissue typing by the operator. The scans were then reanalysed using locally developed image analysis to accurately determine the bone, prosthesis and Gruen zone boundaries. The coefficient of variation (CV) was calculated from the differences of the repeat examinations for each of the seven Gruen zones and for the whole peri-prosthetic region.
Results: The average zone CV was 5%. The poorest was zone 1 (10%) and best zone 4 (2%). With the operator corrections there was an overall 4% improvement. With our method there was an overall 40% reduction in variation (average CV 3%, maximum 4%, minimum 2%). The whole region CV was 3.1% for the standard method 2.7% modified and 1.3% for our method.
Discussion: Our method significantly improved the reproducibility of EXPERT analysis. This study demonstrates the high dependency of DXA precision on robust regional analysis.
Correspondence should be addressed to The Secretary, BHS, c/o BOA, The Royal College of Surgeons, 35–43 Lincoln’s Inn Fields, London WC2A 3PE.
