Three-dimensional analysis of contrast-filled microvessel diameters

Organ blood flow is controlled, in part, by changes in diameter of resistance vessels. In thick tissue, vessels can be imaged with a microscope using contrast-enhancing methods (e.g., fluorescence) and image analysis techniques can be used for quantitative diameter estimations. However, a change in the position of a vessel with respect to the plane of focus can be misinterpreted as a diameter change. In order to address this problem, a 3D image in a light microscope is obtained by serial optical sectioning, and a 3D deconvolution procedure (Avinash et al., 1991, "Fourteenth Association for Research in Otolaryngology Midwinter Meeting, St. Petersberg, FL," Abstract 156) is used to deblur 3D image data. Deblurred sections are computationally projected onto a 2D plane to give an extended-focus image, from which diameter estimates of microvessels are made using a quantitative, 2D diameter-tracking algorithm (Miles, 1987, "Semiautomatic Quantitative Image Analysis of Dynamic in Vivo Cochlear Microvessel Diameters." Ph.D. dissertation, Univ. Michigan; Miles and Nuttall, 1992, IEEE Trans. Biomed. Eng.). Justification for 3D preprocessing before diameter analysis is provided by absolute and relative error analyses using computer-generated synthetic vessels. The 3D diameter analysis technique is validated using a capillary tube of known diameter, filled with fluorescent solution. Demonstration of its applicability is shown in diameter measurements from the vessels of guinea pig cochlea. Our approach, using extended-focus images, minimizes overestimation of microvascular diameters and underestimation of relative diameter changes. Therefore, unambiguous diameter measurements are possible with extended-focus images.