Application of an image analysis system to the quantitation of tumor perfusion and vascularity in human glioma xenografts.

A semiautomatic method based on a computerized digital image analysis system was developed to quantitate the perfused fraction of blood vessels and the relative vascular area in cross-sections of human glioma xenografts, implanted subcutaneously in athymic mice or intracerebrally in nude rats. The fluorescent dye Hoechst 33342 was injected intravenously to detect perfused tumor vessels. An immunofluorescent staining of Collagen type IV visualized the vascular structures in the same tumor section. Whole tumor sections were automatically scanned twice on a computer-controlled motorized stage of a fluorescence microscope under two different settings of the image analysis system. At the beginning of a scanning session an interactive routine was used to determine the threshold value for segmentation of vascular structures from the darker background. After the first scan a composite image was created, from the individually processed microscopic images, containing the detected vascular structures. The second scan yielded another composite image with objects representing the perfused areas. When both composite images were combined the overlapping structures showed the perfused vessels. Differences in perfused fractions and relative vascular areas were found between different tumors. The reproducibility of this analysis system was tested and evaluated. The method developed here provides a fast and accurate technique for simultaneous quantitative analysis of tumor perfusion and vasculature.