The study of GI tissue from surgical or autopsy material has resulted in many important findings. Correlation of the EUS wall structure to histology (Fig. 13.1) is the most significant. 48-53 Accuracy of these studies has depended on use of experimental systems that ensured maximum axial resolution and optimal focusing during sonographic imaging. Generally, specimens are mounted with needles into a container constructed to permit accurate spatial localization of the tissue and corresponding ultrasound image. Tissue are immersed in different media including water, deaerated water, saline, water soluble gels and fat emulsions. The acoustic coupling properties of these media have not been studied systematically. Fresh specimens as well as tissue fixed with different agents have been studied. Microdissection techniques are generally utilized.

Lymph nodes have also been studied with in vitro models.29,54,55 In addition, contrast agents which can enhance grey scale and Doppler signals have been evaluated with various systems. 56-61 These agents include: fluorocarbons, galactose microspheres, oil-water and fat emulsions. They are used for tissue suspension media, for intraluminal media, and for intravenous injection to alter and enhance vessels and lymph node acoustic properties. Substances injected into the GI wall can be located with EUS (Fig. 13. 1 c). With in vitro and in vivo models this technique could be used to investigate disease or the effects of various substances on specific layers of the GI wall.

Swine, canine and sheep animal models have been used to study EUS (Snady H, unpublished observations).20-22,42,62-65 Models Studying gastric ulcers and portal hypertension have been reported.63,64 Experimental systems have been used to test new probes.21-26 Human intraoperative studies have also been performed (Snady H, unpublished observations). Recognizing anatomy correctly for proper image orientation is one of the most difficult aspects of EUS. Consequently, animal models are limited in their applicability to humans because animal anatomy is significantly different. This is particularly true for the pancreas. The normal pig pancreas appears as an inhomogenous structure with scattered hyperechoic, and round or linear foci within the parenchyma. The pancreatic duct is not visibile.62 In contrast, the human pancreas is homogeneous with a visible duct. The pancreatic duct is visible in dogs, but limited surrounding fat and a separate ventral and dorsal pancreas make correlations to human disease difficult (Snady H, unpublished observations). Investigating and formulating an effective diagnostic system for early chronic pancreatitis has been limited by the difficulty of gaining direct one-to-one histological confirmation of EUS findings. The exquisite detail that EUS produces is likely eventually to make it the gold standard against which other methods are cornpared.66