Skip to main content

12-04-2012 | Bone health | Article

Ultrasound ‘outperforms DXA for fracture prediction’


Free abstract

MedWire News: Researchers in Finland have taken the first step toward developing a simple, ultrasound-based imaging model to predict fracture risk in the primary prevention setting.

Their study reports the first-ever in vivo use of ultrasound "backscatter" measurements which, in combination with clinical data, outperformed the current "gold standard" in fracture risk assessment, dual-energy X-ray absorptiometry (DXA).

"Prospective studies will be needed for clinical validation of fracture risk prediction and refinement of the present method," write Janne Karjalainen (University of Eastern Finland) and co-authors in Osteoporosis International.

Axial DXA is routinely used for osteoporosis diagnosis and, when inserted into FRAX (the World Health Organization fracture risk assessment tool), predicts fracture risk with moderate accuracy.

In this study, Karjalainen's team hypothesized that an ultrasound multi-site method might improve on the accuracy of DXA with the added advantage of being suitable for use in the primary care setting.

"Ultrasound backscatter parameters, which are obtained from pulse-echo measurements, have been shown to relate to variations in the trabecular bone structure, as well as its composition and mechanical properties," the researchers explain.

They recruited 30 elderly women (mean age 74.1 years), 14 of whom had a previous hip fracture, and assessed them using DXA and quantitative ultrasound.

As anticipated, mean values for bone mineral density (BMD, obtained with DXA) and apparent integrated backscatter (AIB, obtained with ultrasound) differed significantly between women with and without a hip fracture. This was true for total as well as site-specific values, the authors remark.

The team then compared the performance of two fracture prediction models: FRAX, which uses BMD data, and a new model that included AIB at the femoral neck, age, and weight.

The new AIB-based model offered the highest discriminatory ability, with an area under the receiver operating curve (AUC) of 0.88. This was significantly better than the best BMD-based model, which had an AUC of 0.62.

Karjalainen et al conclude: "The results from qualitative ultrasound could be used to estimate successfully the femoral neck BMD, which had been measured by DXA. Furthermore, the subjects with previous hip fractures could be effectively discriminated from the age-matched healthy control subjects."

MedWire ( is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2012

By Joanna Lyford

Related topics