Purpose: To measure the accuracy and variability in Doppler flow rate measurements within a phantom across four identical scanners.
Methods: Spectral Doppler measurements were acquired on four Philips iU22 ultrasound scanners with two transducers, one linear (L12-5) and one curvilinear (C5-1). B-mode image uniformity was ascertained for all transducers prior to data acquisition. A calibrated Sun Nuclear 403 Doppler Flow phantom was used and a laminar flow rate of 4.0 mL/min was set which produced peak and mean velocities of 40.7 cm/s and 20.35 cm/s, respectively. A single physicist repeated the same measurement at a depth of 4cm on the phantom with both transducers three times on each scanner. The peak and mean flow velocities were estimated by system spectral analysis tools and recorded.
Results: For the linear transducer, the average of the four scanners’ peak and mean flow velocity estimates were 36.8 cm/s (range: 36.19-37.66 cm/s) and 22.7 cm/s (range: 22.06-23.37 cm/s), respectively. The curvilinear transducers average peak and mean flow velocity estimates were 41.2 cm/s (range: 37.99-43.45 cm/s) and 19.3 cm/s (range: 18.05-20.45 cm/s), respectively. The L12-5 transducers, on average, exhibited bias of -9.5% ± 0.75% and 11.5% ± 0.66% for peak and mean velocity estimates, respectively. The C5-1 transducers exhibited average bias of -5.2% ± 2.67% and -5.2% ± 1.02% for peak and mean velocities, respectively.
Conclusion: This study sought to quantify the accuracy and variability of flow rate measured across four identical scanners. These results give insight into the accuracy of scanner estimated flow velocities compared to ground truth, as well as fleet performance for one scanner model. The relatively narrow range of flow velocities estimated by the four scanners can help inform passing thresholds for flow velocity accuracy quality assurance and gives reassurance that patient diagnoses are not subject to inter-scanner variability.
Funding Support, Disclosures, and Conflict of Interest: The phantom used in this study was loaned to us by Sun Nuclear Corporation as part of participation in AAPM Task Group 353.