Research article

Dynamic morphometric characterization of local connective tissue network structure in humans using ultrasound

Helene M Langevin^1,3 , Donna M Rizzo² , James R Fox¹ , Gary J Badger⁴ , Junru Wu⁵ , Elisa E Konofagou⁶ , Debbie Stevens-Tuttle¹ , Nicole A Bouffard¹ and Martin H Krag³

¹  Department of Neurology, University of Vermont, Burlington, VT, USA

²  Department of Civil & Environmental Engineering, University of Vermont, Burlington, VT, USA

³  Department of Orthopaedics & Rehabilitation, University of Vermont, Burlington, VT, USA

⁴  Department of Medical Biostatistics, University of Vermont, Burlington, VT, USA

⁵  Department of Physics, University of Vermont, Burlington, VT, USA

⁶  Department of Biomedical Engineering, Columbia University, New York, NY, USA

author email corresponding author email

BMC Systems Biology 2007, 1:25doi:10.1186/1752-0509-1-25

Published:	5 June 2007

Abstract

Background

In humans, connective tissue forms a complex, interconnected network throughout the body that may have mechanosensory, regulatory and signaling functions. Understanding these potentially important phenomena requires non-invasive measurements of collagen network structure that can be performed in live animals or humans. The goal of this study was to show that ultrasound can be used to quantify dynamic changes in local connective tissue structure in vivo. We first performed combined ultrasound and histology examinations of the same tissue in two subjects undergoing surgery: in one subject, we examined the relationship of ultrasound to histological images in three dimensions; in the other, we examined the effect of a localized tissue perturbation using a previously developed robotic acupuncture needling technique. In ten additional non-surgical subjects, we quantified changes in tissue spatial organization over time during needle rotation vs. no rotation using ultrasound and semi-variogram analyses.

Results

3-D renditions of ultrasound images showed longitudinal echogenic sheets that matched with collagenous sheets seen in histological preparations. Rank correlations between serial 2-D ultrasound and corresponding histology images resulted in high positive correlations for semi-variogram ranges computed parallel (r = 0.79, p < 0.001) and perpendicular (r = 0.63, p < 0.001) to the surface of the skin, indicating concordance in spatial structure between the two data sets. Needle rotation caused tissue displacement in the area surrounding the needle that was mapped spatially with ultrasound elastography and corresponded to collagen bundles winding around the needle on histological sections. In semi-variograms computed for each ultrasound frame, there was a greater change in the area under the semi-variogram curve across successive frames during needle rotation compared with no rotation. The direction of this change was heterogeneous across subjects. The frame-to-frame variability was 10-fold (p < 0.001) greater with rotation than with no rotation indicating changes in tissue structure during rotation.

Conclusion

The combination of ultrasound and semi-variogram analyses allows quantitative assessment of dynamic changes in the structure of human connective tissue in vivo.