Body composition with dual energy X-ray absorptiometry: from basics to new tools

Carmelo Messina, Domenico Albano, Salvatore Gitto, Laura Tofanelli, Alberto Bazzocchi, Fabio Massimo Ulivieri, Giuseppe Guglielmi, Luca Maria Sconfienza

Research output: Contribution to journalReview articlepeer-review


Dual-energy X-ray absorptiometry (DXA) in nowadays considered one of the most versatile imaging techniques for the evaluation of metabolic bone disorders such as osteoporosis, sarcopenia and obesity. The advantages of DXA over other imaging techniques are the very low radiation dose, its accuracy and simplicity of use. In addition, fat mass (FM) and lean mass (LM) values by DXA shows very good accuracy compared to that of computed tomography and magnetic resonance imaging. In this review we will explain the technical working principles of body composition with DXA, together with the possible limitations and pitfalls that should be avoided in daily routine to produce high-quality DXA examinations. We will also cover the current clinical practical application of whole body DXA values, with particular emphasis on the use of LM indices in the diagnostic workup of reduced muscle mass, sarcopenia and osteosarcopenic obesity according to the most recent guidelines. The possible use of adipose indices will be considered, such as the fat mass index (FMI) or the android/gynoid ratio, as well as lipodystrophy indices and the evaluation of visceral adipose tissue (VAT). Whenever available, we will provide possible cut-off diagnostic values for each of these LM and FM indices, according to current literature and guidelines.

Original languageEnglish
Pages (from-to)1687-1698
Number of pages12
JournalQuantitative Imaging in Medicine and Surgery
Issue number8
Publication statusPublished - Aug 2020


  • Dual-energy X-ray absorptiometry (DXA)
  • fat mass index
  • obesity
  • sarcopenia


Dive into the research topics of 'Body composition with dual energy X-ray absorptiometry: from basics to new tools'. Together they form a unique fingerprint.

Cite this