Evaluation of Abdominal Computed Tomography Image Quality Using a New Version of Vendor-Specific Model-Based Iterative Reconstruction

Abstract

Purpose To qualitatively and quantitatively compare abdominal CT images reconstructed with a new version of model-based iterative reconstruction (Veo 3.0; GE Healthcare) to those created with Veo 2.0. Materials & Methods This retrospective study was approved by our IRB and was HIPPA compliant. The raw data from 29 consecutive patients who had undergone CT abdomen scanning was used to reconstruct 4 sets of 3.75mm axial images: Veo 2.0, Veo 3.0 standard, Veo 3.0 5% resolution preference and Veo 3.0 20% resolution preference. A slice thickness optimization of 3.75 mm and texture feature was selected for Veo 3.0 reconstructions. The images were reviewed by three independent readers in a blinded, randomized fashion using a 5-point Likert scale and 5-point comparative scale. Multiple 2D circular regions of interest were defined for noise and contrast-to-noise ratio (CNR) measurements. Line profiles were drawn across the 7 lp/cm bar pattern of the CatPhan 600 phantom for spatial resolution evaluation. Results The Veo 3.0 standard image set was scored better than Veo 2.0 in terms of artifacts (mean difference 0.43, 95% CI 0.25-0.6, P<0.0001), overall image quality (mean difference 0.87, 95% CI 0.62-1.13, P<0.0001) and qualitative resolution (mean difference 0.9, 95% CI 0.69-1.1, P<0.0001). While the Veo 3.0 standard and RP05 presets were preferred across most categories, the Veo 3.0 RP20 series ranked best for bone detail. Image noise and spatial resolution increased along a spectrum with Veo 2.0 the lowest and RP20 the highest. Conclusion Veo 3.0 enhances imaging evaluation relative to Veo 2.0; readers preferred Veo 3.0 image appearance despite the associated mild increases in image noise. These results provide suggested parameters to be used clinically and as a basis for future evaluations such as focal lesion detection in the oncology setting.

Third version of vendor-specific model-based iterativereconstruction (Veo 3.0): evaluation of CT image quality in the abdomen using new noise reduction presets and varied slice optimization

“…A 5-point Likert scale was used by the readers to rate overall image quality without specific comparison to other series. 20 The following scores were used: 1, excellent; 2, good; 3, acceptable; 4, suboptimal and 5, very poor. The four image sets were also ranked against one another on a comparative scale with regard to overall image quality, image noise and resolution/detail.…”

“…Previous studies of Veo 3.0 with standard and resolution preference presets have demonstrated improved image quality over Veo 2.0 in both the chest and the abdomen. 14,19,20 Jensen et al demonstrated reduced artefacts with Veo 3.0 compared to Veo 2.0 in the abdomen and Li et al detailed a more isotropic noise power spectrum through the use of the texture function. 14,20 Furthermore, Yasaka et al demonstrated the higher spatial resolution RP20 clinical preset to improve detail during lung evaluation when compared to Veo 2.0 and ASIR.…”

Deep Learning Reconstruction at CT: Phantom Study of the Image Characteristics

“…Its value for organ recognition (segmentation), lesion detection, and lesion characterization has been explored (11) and its ability to detect and reduce the image noise has been demonstrated (12À16). Unlike the conventional noise reduction methods, which involve a trade-off between spatial resolution and noise reduction (6,17), deep learning lowers the image noise and increases spatial resolution simultaneously (9). In this phantom study we examined the noise characteristics, spatial resolution, and task-based detectability of DLR images.…”

Tradeoff between noise reduction and inartificial visualization in a model-based iterative reconstruction algorithm on coronary computed tomography angiography

“…We believe that the optimized balance between spatial resolution and image noise behavior achieved by FIRST-CS provided better subjective image quality despite the increased image noise. Jensen et al [ 13 ] reported similar results in that the newer model-based IR (Veo 3.0, GE Healthcare) enhanced the imaging evaluation relative to the prior-generation model-based IR (Veo 2.0), and readers awarded higher scores for the imaging appearance of Veo 3.0 despite the increased image noise. These findings suggest that the image noise and image appearance may have a tradeoff relationship.…”

“…Furthermore, we estimated the effective radiation dose of the chest with the following equation: effective dose = (CTDI vol × anatomical range for the chest) × 0.014. [ 13 ]…”