Osteoporotic changes in the porous microstructure of cancellous bone can be detected using ultrasonic backscatter. Ideally, the incident pulse should be perpendicular to the bone surface, but this is difficult to achieve clinically with a handheld probe. The goal of this study was to assess the sensitivity of multiple backscatter parameters to errors caused by non-normal incidence. Measurements were performed on a polymer foam that simulates the ultrasonic properties of cancellous bone with one surface embedded in ∼3 mm of epoxy to simulate the outer bone cortex. Ultrasonic pulses were transmitted at angles of incidence ranging from 0 to 30 deg using a 3.5 MHz transducer. Six backscatter parameters were tested. Three were based on spectral analyses of the power in the backscatter signal. The other three were based on the rate the amplitude of the signal decreased with time, or equivalently depth, in the tissue. The latter three parameters performed well with errors less than 30% in most cases. Large errors (100%) were observed for other parameters. These results indicate that errors in backscatter measurements of bone caused by non-normal incidence depend strongly on the parameter chosen.
Hoffmeister et al. (Tue,) studied this question.