Abstract

Following the intravenous injection of fluorescein, the concentrations of fluorescein in the ocular aqueous and vitreous are measured to determine the integrity of the blood-retinal barrier. The fluorophotometer has an effective sampling volume formed by the intersection of the excitation beam and the fluorescence detection beam. As the sampling volume is scanned through the ocular media, it produces an instrument spread function that is convolved with the actual distribution of fluorescein. Provided that the spread function can be determined, it is possible to perform a deconvolution to recover an approximation to the original distribution. Several methods of deconvolution can be utilized. Factors that affect the results include sampling interval, alignment of the data, precise knowledge of the spread function, and noise. Simulations and model data demonstrate that substantial improvement can be achieved. With patient data, the principal advantage of deconvolution is the ability to assess fluorescein concentration close to the retina, a region that is otherwise obscured by the presence of strong fluorescence in the retina itself.

© 1985 Optical Society of America