VALIDATION OF AUTOMATED SPECTROFLUORIMETRY FOR MEASUREMENT OF REGIONAL LUNG PERFUSION USING FLUORESCENT MICROSPHERES.
C. Schimmel, D. Frazer, S. Bernard, R. Glenny. University of Washington, Division of Pulmonary and Critical Care Medicine, Seattle, WA 98195-6522 USA

Fluorescent microspheres are an alternative to radio-labeled microspheres for measurement of regional organ blood flow. However, sample preparation and analysis using manual spectrofluorimetry is comparatively time and labor intensive. A spectrofluorimeter interfaced with an automated sample delivery system is now commercially available and was used to determine relative blood flow in lung tissue. These results were correlated with those obtained using the radio-labeled microsphere method. Additionally, we determined intra-method reliability by simultaneous injection of different types of fluorescent and radio-labeled microspheres.

Two different radio-labeled, and four different fluorescent-labeled microspheres (15-µm diameter, ~106 per color) were simultaneously injected via the femoral vein of an anesthetized pig. The lungs were excised, flushed with saline, dried at total lung capacity, and cubed into ~2-cm3 pieces (n=600). Radioactivity for each piece was measured using a gamma counter with count rates corrected for decay, background and spillover. Fluorescent dye was extracted from lung tissue with 2-ethoxyethyl acetate and measured with a Perkin-Elmer LS-50B spectrofluorimeter equipped with an AS-91 robotic probe and sample tray unit in line with a 40-µl flow cell and DS 6 diluter station.

Blood flow, normalized to the mean flow, was determined for each piece for four fluorescent colors and two radio-labeled nuclides. The correlation between Scandium46 and Tin113 was R=1.00, slope=1.00, and intercept=0.00. Mean intra-method correlations for all six combinations of colors was R=0.98 ± 0.0091, mean slope=0.99 ± 0.012 and mean intercept=0.011 ± 0.019. The mean inter-method correlation for all eight combinations of colors and radioactive nuclides was R=0.98 ± 0.0076, mean slope=0.99 ± 0.0098 and mean intercept=0.0092 ± 0.011.

Analysis time for 100 samples is 5 hours, about twice the time required for manual analysis. However, the autosampler can analyze up to 280 samples unattended. We conclude that automated spectrofluorimetry is a reliable and labor-saving method for measurement of fluorescence in large samples sets.

Participant list or Abstract list