The fluorescent microsphere (FM) method for determination of regional organ blood Row is labor intensive, requiring processing of 100 s to 1000 s of samples per organ. To save time and reduce handling errors, we developed an automated fluorescence analysis system by interfacing a commercially available spectrofluorimeter and sample delivery unit to a PC, operated by a Windows95(R)-based program, WINFAC. WINFAC allows versatile analysis setup and instrument control for fluorescent intensity acquisition at fixed wavelengths or by synchronous scanning. Data is presented on screen, in real-time, and stored in text format. Reference analyses are performed at specified intervals and the coefficient of variation is continuously updated to monitor instrument performance. The automated system was validated against radioactive microspheres (RM) for lung perfusion in a pig and sheep and evaluated for analysis reproducibility. Fluorescent and radioactive microspheres were delivered by simultaneous injection into a femoral vein. Lungs were excised, flushed, dried at total lung capacity, and cubed into approximate to 2cm(3) pieces (n = 833 and 1560, pig and sheep, respectively). Radioactive count rates were determined for each lung piece (corrected for decay, background and spillover). Fluorescent dyes were extracted in solvent and intensities were determined at fixed wavelengths, using the automated spectrofluorimerer (corrected for background and spillover). Multi-color reference solutions were measured every 50 samples to monitor instrument reproducibility. Blood flow estimates for each piece determined by RM and FM methods were highly correlated: R-2 = 0.98 +/- 0.017, Slopes = 1.00 +/- 0.007 and Intercepts = 0.00 +/- 0.006 (mean +/- SD). The CV of repeat reference analyses was 0.71% +/- 0.16, a 30% to 50% reduction relative to manual analysis. Automated spectrofluorimetry reduces measurement errors and is a reliable and time saving advancement. With this technology, use of FM to measure regional lung perfusion approaches the ease and accuracy of the RM method. (C) 2000 Elsevier Science Ireland Ltd. All rights reserved.