Photoacoustic flowmetry (PAF) based on time-domain cross correlation of photoacoustic signals is a promising technique for deep tissue measurement of blood flow velocity. Signal processing has previously been developed for single element transducers. Here, the processing methods for acoustic resolution PAF using a clinical ultrasound transducer array are developed and validated using a 64-element transducer array with a −6 dB detection band of 11 to 17 MHz. Measurements were performed on a flow phantom consisting of a tube (580 μm inner diameter) perfused with human blood flowing at physiological speeds ranging from 3 to 25 mm / s. The processing pipeline comprised: image reconstruction, filtering, displacement detection, and masking. High-pass filtering and background subtraction were found to be key preprocessing steps to enable accurate flow velocity estimates, which were calculated using a cross-correlation based method. In addition, the regions of interest in the calculated velocity maps were defined using a masking approach based on the amplitude of the cross-correlation functions. These developments enabled blood flow measurements using a transducer array, bringing PAF one step closer to clinical applicability.
This work demonstrates the first measurements of blood flow velocity using photoacoustic flowmetry (PAF) employing a transducer array. The measurements were made in a flow phantom consisting of a tube (580 μm inner diameter) containing blood flowing steadily at physiological speeds ranging from 3 mm/s to 25 mm/s. Velocity measurements were based on the generation of two successive photoacoustic (PA) signals using two laser pulses with a wavelength of 1064 nm; the PA signals were detected using a 64-element transducer array with a -6 dB detection bandwidth of 11-17 MHz. We developed a processing pipeline to optimise a cross-correlation based velocity measurement method comprising the following processing steps: image reconstruction, filtering, displacement detection, and masking. We found no difference in flow detection accuracy when choosing different image reconstruction algorithms (time reversal, Fourier transformation, and delay-and-sum). High-pass filtering and wallfiltering were however found to be essential pre-processing steps in order to recover the correct displacement information. We masked the calculated velocity map based on the amplitude of the cross-correlation function in order to define the region of interest corresponding to highest signal amplitude. These developments enabled blood flow measurements using a transducer array, bringing PAF one step closer to clinical applicability.