RESUMO
OBJECTIVE: The purpose of this study was to compare volumetric image processing techniques for reducing noise and speckle while retaining tissue structures in 3-dimensional (3D) gray scale ultrasound imaging. METHODS: Eighty subjects underwent a clinically indicated abdominal or obstetric 3D ultrasound examination (20 hepatic, 20 renal, and 40 obstetric cases). Volume data were processed on a pixel ("2-dimensional [2D] processing") or a voxel ("3D processing") basis using commercially available image enhancement software (ContextVision AB, Linköping, Sweden). Randomized, side-by-side comparisons of the image processing techniques were performed for each subject. An independent and blinded reader scored the volumes for image quality on a 3-point scale from 1 (worst) to 3 (best) and compared the results using a nonparametric Wilcoxson signed rank test. RESULTS: The 40 subjects with abdominal 3D imaging received a mean score (+/- 1 SD) of 1.52 +/- 0.51, 2.45 +/- 0.60, and 2.75 +/- 0.44 for the original, the 2D processed, and the 3D processed volumes, respectively. The differences between the unprocessed and the processed volumes were highly statistically significant (P < .0001), as was the difference between the 2D and 3D processing methods (P = .002). Similar results were obtained for the obstetric data sets (n = 39 due to an acquisition problem) with a mean score of 1.03 +/- 0.16 for the original, 2.33 +/- 0.48 for the 2D processed, and 2.79 +/- 0.47 for the 3D processed volumes (P < .003). CONCLUSIONS: A new volumetric ultrasound image enhancement technique has been assessed in abdominal and obstetric applications. Compared to unprocessed volumes and volumes processed with 2D image enhancement software, the new 3D processing technique performed best.
