In recent years, a new method to measure transverse blood flow based on the decorrelation of the radiofrequency (RF) signals of intravascular ultrasound (IVUS) rotating single-element scanners was introduced. We report here in vitro, animal and patient testing to evaluate the correlation-based method using an IVUS array catheter. A new correlation-based method to dynamically correct the correlation coefficients for noise is implemented. The decorrelation due to noise was estimated from the correlation coefficients from flowing blood obtained at increasing time lags. First, blood flow experiments were carried out with different catheters in a tissue-mimicking flow phantom with an inner diameter ranging from 3.0 to 5.0 mm. A calibrated electromagnetic flow meter (EMF, range: 0 to 250 cc/min) was used as a reference. Good linear relationships were found between the IVUS-derived flow and the calibrated EMF (all R(2)> 0.96). The catheter position within the flow phantom and the size of the ring-down were theoretically analyzed. These elements, and noise in the RF signals, have an important influence on the IVUS blood flow measurements reflected by the offset and the slope of the linear relationships. By placing the IVUS catheter outside the flow phantom, parabolic blood flow profiles were also measured. Second, IVUS blood flow measurements were performed in the carotid artery of two Yorkshire pigs, which showed linear relationships (all R(2)> 0.85) between the IVUS-derived flow and the calibrated EMF. Experimentally, the offset was lower than 3 mL/min and the slope was close to 1. Third, IVUS blood flow measurements were performed in coronary arteries in patients. Preliminary results for the coronary flow reserve (CFR = high flow/baseline flow) in patients using the decorrelation method of RF signals of an array IVUS scanner were comparable with CFR based on Doppler measurements.