Coherent normalization of finger strontium XRF measurements: feasibility and limitations.

A non-invasive in vivo x-ray fluorescence (XRF) method of measuring bone strontium concentrations has previously been reported as a potential diagnostic tool able to detect strontium concentration in the finger and ankle bones. The feasibility of coherent normalization for (125)I-source-based finger bone strontium x-ray fluorescence (XRF) measurements is assessed here by theoretical considerations and Monte Carlo simulations. Normalization would have several advantages, among which are the correction for the signal attenuation by the overlying soft tissue, and intersubject variability in the bone size and shape. The coherent normalization of bone strontium XRF measurements presents several challenges dictated by the behaviour of the coherent cross section and mass attenuation coefficient at the energies involved. It was found that the coherent normalization alone with either 22.1 keV or 35.5 keV photons was not successful in correcting for the overlying soft tissue attenuation. However, it was found that the coherent peak at 35.5 keV was able to correct effectively for variability in the finger bone size between people. Thus, it is suggested that, if the overlying soft tissue thickness can be obtained by means of an independent measurement, the 35.5 keV peak can be used to correct for the bone size, with an overall accuracy of the normalization process of better than 10%.