Facebook Twitter YouTube


Introducing a New Ultrasound-Based Technique for Quantitative Assessment of Infant Bone

Friday, October 19, 2012
Room R02-R05 (Morial Convention Center)
Yasaman Fatemi, BA1, Armen Sarvazyan, DSc2, Garth F. Asay, MD1 and Azra Alizad, MD3, (1)Mayo Medical School, Mayo Clinic College of Medicine, Rochester, MN, (2)Artann Laboratories, Lambertvill, NJ, (3)Physiology and Biomedical Engineering, Mayo Clinic College of Medicine, Rochester, MN

Purpose: To develop a non-invasive screening method with sufficient sensitivity and specificity to provide evidence of developing osteopenia of prematurity over the first several weeks of life in premature infants.

Methods: The proposed method uses ultrasound energy to remotely exert a minute vibrating force on the bone and detects the acoustic response of the bone to such vibration by a small sensor. The method of remote excitation is particularly suitable for the small and sensitive body of premature infants. Excitation in a wide range of frequencies provides rich data for spectral analysis of the acoustic signal.  The result of this analysis is used to characterize bone mineralization. Experiments are conducted on bone samples in two groups of normal bones and those chemically treated to simulate osteopenia of prematurity. The acoustic signals are processed in time and frequency to obtain quantitative parameters, such as the temporal extension of select frequency components and mean frequency shift, which are highly correlated with bone condition.

Results: The spectrograms in Figure 1 show consistent changes in the spectrogram of the signal due to demineralization; the signal energy shifts to higher frequencies as demineralization increases. The plot of mean frequency versus demineralization time (Figure 2) shows that the mean frequency shifts from 90 kHz for intact bone to about 275 kHz for the fully demineralized bone.

Conclusion: This experiment demonstrates that features of the spectrogram can be used as indicators of demineralization. Although the frequency range may different for different bone sizes, but a 3-fold shift in mean frequency can be easily detected and used to discriminate between mineralized and demineralized bone. Results of this study provide a strong impetus for further development of the proposed technique for clinical applications.

Description: C:\Users\axa04\Desktop\AAP abstract files\image001.png

Figure 1. Spectrogram of intact (left), partially demineralize (middle), and fully demineralized (right) rabbit femur. Changes in bone mineralization are clearly reflected in consistent changes in the spectrograms.

Description: C:\Users\axa04\Desktop\AAP abstract files\image002.jpg

Figure 2. Weighted mean frequency versus demineralization time. (Day 1 is intact bone).