Previous studies using radioactive microspheres (RMS) and fluorescent
microspheres (FMS) to validate fluorescent methods disregarded error caused by
spillover from neighboring fluorescent colors Since spillover is a constant
percentage of peak emission intensity, when relative flow distribution remains
constant, correlations between FMS and RMS are artificially high. New
applications of the FMS method, such as inhalation studies, and availability of
additional colors of FMS with high spillover (up to 30%) exacerbate this problem.
To quantify spillover error, and to evaluate a matrix correction method, we
induced a large reduction in blood flow to one lung in an anes-thetized pig using
a balloon type catheter. 1 RMS and 5 FMS were simultaneously injected into a
femoral vein. After partial balloon inflation, a second RMS and 5 different FMS
were simultaneously injected. The lungs were excised, flushed with saline, dried
at total lung capacity, and cubed into ~2-cm^{3} pieces (n=630). Radioactivity for
each piece was measured and corrected for decay, background and spillover.
Fluorescent dyes were extracted and measured using two analytical methods, (1)
designed to maximize peak intensity and (2) designed to minimize spillover, with
a LS-50B spectrofluorimeter and FAC9 software. Intensities were corrected for
spillover by solving a system of simultaneous linear equations.

Blood flow, normalized to the mean, was determined for all pieces, FMS, and RMS.
The worst case scenario was obtained by using only pieces with decreased flow
relative to baseline (N=325). Correlations between RMS and FMS were determined
for both analytical methods, using signals before and after correction for
spillover. The percent error was defined as 1-R^{2}. The mean % error for method 1
before and after spillover correction was 9.14 ± 7.85 and 5.93 ± 2.40,
respectively. The mean % error for method 2 before and after spillover
correction was 4.59 ± 2.19 and 3.75 ± 1.82, respectively.

For this worst case scenario we observed a 5% error in the FMS method after correcting for spillover. This is an overestimate of the true error because of the small number of microspheres/piece and the assumption of no error in the radioactive measure. All ten fluorescent colors can be used in a single experiment using matrix inversion for spillover correction.