Dual energy x-ray absorptiometry (DXA) extended femur scans to support opportunistic screening for incomplete atypical femoral fractures: A short term in-vivo precision study.

Background
Atypical femoral fracture (AFF) is documented as a known but rare complication of bisphosphonate use for the treatment of osteoporosis. These present in an incomplete form prior to failure, which results in a complete fracture requiring surgical intervention. Dual energy x-ray absorptiometry (DXA) is the gold standard for the diagnosis of Osteoporosis and for monitoring the response to therapeutic interventions. This provides an opportunity to use routine DXA scans to identify incomplete atypical fractures, which can subsequently be monitored for progression and pre-fracture intramedullary nailing undertaken where necessary. DXA manufacturers have developed extended femur scans to assess and measure the femoral cortex for incipient atypical femoral fractures. The aim of this study was to evaluate the precision errors related to the cortical measurements and for hip bone mineral density using the extended femur setting. Methodology A single operator performed duplicate same day in-vivo measurements of the femur in 30 consenting participants, with repositioning between scans, during their visit for routine DXA scanning. The study was performed on a single GE Lunar Prodigy scanner (GE Lunar, Bedford, UK). Root mean squared standard deviation (RMS SD) and coefficient of variation (RMS CV%) were calculated for the cortex measurements known as beaking index (BI) and hip bone mineral density (BMD) measurements. Results The use of the extended femur scan software yielded an RMS SD (RMS CV%) of 0.011 (1.43%) for the total hip and 0.015 (2.05%) for the femoral neck. The BI measurement RMS SD (RMS CV%) was 0.473 (38.10%) Visual assessment of the femoral cortex discounted all positive BI anomalies as software generated in this dataset. Conclusions The use of extended femur scan software did not affect the precision errors of the BMD measurements at the hip when compared to the literature on focused hip scans. The BI precision errors were much greater and therefore unreliable unless accompanied by visual assessment which is recommended to avoid unnecessary investigation in around one fifth of the scan population.