OBJECTIVES: This study compared simultaneous regional myocardial blood flow (RMBF) measurements using fluorescent microspheres (FM) and radiolabeled microspheres (RM). The utility of an internal standard during processing was also examined. METHODS: Paired FM and RM were injected into the left atrium of 9 anaesthetised rabbits. RMBF was altered by use of either regional ischaemia or (-)-N6-(2-phenylisopropyl)-adenosine. Radioactivity of blood reference and tissue samples was quantitated using standard methods. Samples were then digested with potassium hydroxide and microspheres recovered by vacuum filtration, with an additional label of FM as the internal standard. FM labels were extracted using Carbitol acetate and quantitated using fluorescence spectroscopy. Agreement between the fluorescent and radioactive methods was assessed using both orthogonal regression and difference-against-mean analyses. RESULTS: Using recovery-uncorrected data, the slope of the orthogonal regression of RM and FM-determined RMBF was not statistically different from 1, but the intercept was statistically different from 0 [-0.03(0.01), P = 0.005] and the mean RMBF by each method differed from one another [1.24(0.08) vs. 1.17(0.08) ml.min-1.g-1, P = 0.0002]. The mean +/- 2 s.d. of the differences of RMBF (RM minus FM) was +0.07 +/- 0.30 ml.min-1.g-1. Although recovery of FM from tissue averaged 97.6(1.2)%, use of the internal standard to correct for losses substantially improved the agreement between RM and FM-determined RMBF: the orthogonal regression slope was not statistically different from 1, the intercept was not statistically different from 0, and the means of the flows were not different. The mean +/- 2 s.d. of the differences of RMBF was -0.01 +/- 0.22 ml.min-1.g-1. The internal standard also improved RMBF estimates from samples with simulated large spillage during processing. CONCLUSION: Fluorescent microspheres are an equivalent alternative to radiolabeled microspheres for the estimation of RMBF. Although the overall recovery of microspheres using this technique was high, use of an internal standard is recommended for correction of random losses.