In vitro measurement accuracy of three-dimensional ultrasound

OBJECTIVES: We sought to validate distance and volume measurements in three-dimensional (3-D) ultrasound images. BACKGROUND: Even with the latest equipment, it is not known how accurate 3-D echocardiographic measurements are. METHODS: Six models were imaged in ethanol solution and two within a tissue phantom using a mechanical rotation device rotating in 1 degrees intervals and a real-time 3-D scanner. Distance and volume measurements (n = 60) were performed in two-dimensional (2-D) and 3-D images using TomTec and InViVo software. RESULTS: Distance measurements had a mean total error between 1.12% and 2.31% for Acuson (2.5 MHZ, 3 MHZ, and 4 MHZ) and Hewlett Parkard (HP) fusion frequencies h and m, HP fusion harmonic B in the axial, and between 3.5% and 4.9% in the lateral dimension. HP Harmonic A and B, Volumetrics (2.5 MHZ), and HP fusion Harmonic A exhibited significantly higher differences to reality with a mean difference between 5.1% and 8.9% in the axial and between 6.2% and 7.9% in the lateral direction. Axial 2-D measurements were not different from real dimensions except Volumetrics model 1. In the lateral axis, all imaging modalities were different from reality except the fusion harmonic modus B. Using the HP fusion frequency h and HP fusion Harmonic B-mode, volume measurements in 3-D images significantly underestimated reality, while Acuson's fundamental frequency 3.5 MHZ was not different from real volumes. CONCLUSION: Three-dimensional visualization using different ultrasound settings results in different accuracy.