In vivo measurement of corneal angiogenesis with video data acquisition and computerized image analysis.
BACKGROUND: Measurement of corneal angiogenesis is useful for quantitating the effects of angiogenic stimuli and for evaluating the efficacy of potential inhibitors of neovascularization. Because accurate methods to record the entire pattern of corneal neovascularization over time in individual living animals do not exist, we have developed a noninvasive method to achieve this goal. EXPERIMENTAL DESIGN: The technique couples video data acquisition methods with computerized analysis of the video images. A stereotactic holding and positioning device allows alignment of the cornea such that it is viewed in a known and repeatable way. Contrast between blood vessels and corneal stroma in the images is enhanced by illuminating the cornea with monochromatic light centered on the peak absorption of hemoglobin. For each observation, multiple overlapping images of the peripheral cornea are recorded on videotape and subsequently digitized with a computer image analysis system. Overlapping regions are found by either statistical cross-correlation or common object identification methods. A montage of nonoverlapping adjacent images is made. Background electronic signals are reduced and contrast is enhanced in each montage with the aid of image processing. Finally, vessel area is calculated by pixel counting after establishing the density range for vessel identification. To demonstrate the utility of the method, we measured over the course of 18 days, the total area of neovascularization in rabbit corneas cauterized with silver/potassium nitrate, and treated topically with either normal saline (control) or 1% prednisolone acetate (100 microliters four times daily for 10 days). The corneal angiogenic response was measured at the time of cautery and at selected intervals thereafter. RESULTS: The amount of angiogenesis in each control cornea increased progressively during the entire observation period. In contrast, the prednisolone-treated corneas manifested less neovascularization than controls during the treatment interval. After treatment ended, the amount of corneal angiogenesis increased slightly in this experimental group. This method provided multiple data points from each animal. CONCLUSIONS: To date, accurate measurement of corneal neovascularization has been a time-consuming process yielding few data points for each animal studied. The new method described, accurately measures even small changes in the area of corneal neovascularization, and allows for multiple observations of the same animal. The technique as currently developed, however, is not applicable to corneas that are markedly opaque or associated with intracorneal hemorrhage.