Ultrasonographic tissue characterization of equine superficial digital flexor tendons by means of gray level statistics.
Authors
van Schie HT, Bakker EM, Jonker AM and van Weeren PR
Publisher
Am J Vet Res 2000; 61(2): 210-219
Publishing detail
PMID: 10685695
Abstract
Objective:
To correlate quantitative analysis of ultrasonographic images of normal (injury-free) equine superficial digital flexor (SDF) tendons and equine SFD tendons that have pathologic changes with corresponding histologic sections.
Sample population: 4 SDF tendons, 2 of which had various stages of tissue integrity. The 2 ipsilateral tendons were used as points of reference.
Procedure:
Tendons were mounted in a custom-made device that permitted sequential scanning, transversely and perpendicular to the tendon long axis. At precise steps of 0.5 mm, transverse ultrasonographic images were collected. Subsequently, tendons were fixed and prepared for histologic examination. The following 8 tissue types were discerned: normal young, normal old, necrotic, early granulation, late granulation, early fibrotic, late fibrotic, and scar tissues. In areas of interest, the corresponding ultrasonographic images were selected for gray level statistical analysis.
Results:
Compared with other tissue types, early-stage granulation tissue was characterized by substantially lower mean gray level and a clearly different histogram. Necrotic tissue had a higher mean gray level, with a virtually normal histogram. In late granulation and early fibrotic tissues, the mean gray level and the histogram could not be discerned from those of normal tendon tissue. The same applied to late fibrotic and scar tissues; mean gray levels were fractionally lower than those of normal tendon tissue with a completely normal histogram.
Conclusions:
Although quantification of the transverse ultrasonographic image by use of first-order gray level statistics may be helpful, the method is not sufficiently sensitive to accurately and unequivocally determine the type of tendon tissue. Quantitative analysis should incorporate transverse and longitudinal information.