Effect of temperature on the longitudinal variability of quantitative ultrasound variables.

Abstract It is unclear whether longitudinal change in phantom measurements bears any relation to the long-term in vivo instrument performance of quantitative ultrasound devices. Longitudinal quantitative ultrasound phantom data were obtained by measuring the manufacturer-provided phantom at ambient temperature and two different sets of Leeds phantoms at either ambient temperature or following a phantom temperature-control protocol. Measurements were performed using the Achilles Plus bone densitometer. Changes in longitudinal phantom data were compared to in vivo quantitative ultrasound data obtained from seven healthy, young volunteers. A cosinor model with linear trend and Hotelling's T 2 -test were used to quantify seasonal rhythms and long-term drift in quantitative ultrasound variables. Temperature effects and marked seasonal rhythms on quantitative ultrasound phantom measurements were evident but were far less apparent in vivo. Longitudinal precision of quantitative ultrasound variables was poorer for the manufacturer-provided phantom than for phantoms that were subjected to a temperature-control protocol or for healthy volunteers. This study has shown that longitudinal precision and longitudinal change differs between in vivo and phantom data. Longitudinal quantitative ultrasound measurements for monitoring change in skeletal status cannot, as yet, be properly controlled.