Comparison of virtual monoenergetic imaging between a rapid kilovoltage switching dual-energy computed tomography with deep-learning and four dual-energy CTs with iterative reconstruction

Greffier, Joël; Guiu, Boris; Frandon, Julien; Loisy, Maeliss; Oliveira, Fabien de; Douek, Philippe; Beregi, Jean‐Paul; Dabli, Djamel

doi:10.21037/qims-21-708

Cited by 20 publications

(37 citation statements)

References 33 publications

(65 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the Canon Medical DECT platforms, a DL-based spectral reconstruction (DLSR) is available, based on the creation of deep learning views generated by the trained neural network for opposite and for the same energy views [ 31 , 33 ]. These energy views are generated by transforming the views from one energy to the other [ 26 , 31 ]. The measured views are completed by DL views at each energy to create a full sinogram for each kV.…”

Section: Introductionmentioning

confidence: 99%

“…The spectral reconstruction was trained on complete sinograms obtained at each energy using different patient and phantom data (more details are provided in Supplementary material). This DLSR was shown to have the capacity to generate low-noise spectral images [ 26 , 31 ].…”

Section: Introductionmentioning

confidence: 99%

“…The first version of DLSR developed by Canon Medical Systems Corporation in 2019 allowed one slice thickness reconstruction of 0.5 mm and a body spectral kernel [ 31 ]. A new release of this algorithm was introduced in 2021, also allowing reconstruction with slice thickness of 0.5 mm.…”

Section: Introductionmentioning

confidence: 99%

“…This training with high quality and high quantity datasets led to the improvement of the algorithm performance and thus of the quality of the spectral images obtained. Several studies have assessed the performance of the first version of this DLSR [ 26 , 31 ] and one study has compared its latest version with four other DECT platforms equipped with iterative reconstruction algorithms [ 31 ]. However, to our knowledge, no study has yet compared the performance of the two versions of this DLSR and their potential for reducing the iodined contrast media doses.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Comparison of image quality of two versions of deep-learning image reconstruction algorithm on a rapid kV-switching CT: a phantom study

et al. 2023

Self Cite

View full text Add to dashboard Cite

Background To assess the impact of the new version of a deep learning (DL) spectral reconstruction on image quality of virtual monoenergetic images (VMIs) for contrast-enhanced abdominal computed tomography in the rapid kV-switching platform. Methods Two phantoms were scanned with a rapid kV-switching CT using abdomen-pelvic CT examination parameters at dose of 12.6 mGy. Images were reconstructed using two versions of DL spectral reconstruction algorithms (DLSR V1 and V2) for three reconstruction levels. The noise power spectrum (NSP) and task-based transfer function at 50% (TTF50) were computed at 40/50/60/70 keV. A detectability index (d') was calculated for enhanced lesions at low iodine concentrations: 2, 1, and 0.5 mg/mL. Results The noise magnitude was significantly lower with DLSR V2 compared to DLSR V1 for energy levels between 40 and 60 keV by -36.5% ± 1.4% (mean ± standard deviation) for the standard level. The average NPS frequencies increased significantly with DLSR V2 by 23.7% ± 4.2% for the standard level. The highest difference in TTF50 was observed at the mild level with a significant increase of 61.7% ± 11.8% over 40−60 keV energy with DLSR V2. The d' values were significantly higher for DLSR V2 versus DLSR V1. Conclusions The DLSR V2 improves image quality and detectability of low iodine concentrations in VMIs compared to DLSR V1. This suggests a great potential of DLSR V2 to reduce iodined contrast doses.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Comparison of image quality of two versions of deep-learning image reconstruction algorithm on a rapid kV-switching CT: a phantom study

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…The photoelectric effect depends on the atomic number and dominates at energies up to 100 keV, while Compton scattering is largely dependent on physical density and dominates at higher energies [6]. There are several different technical approaches to obtaining spectral CT information, including dual-source [7], split-filter [8], kVp-switching [9,10], dual-layer [11,12], and photon-counting CT [13,14]. The various technological solutions make it possible to obtain spectral data of varying quality [15][16][17], and generally there is a trend towards detector-based solutions.…”

Section: Introductionmentioning

confidence: 99%

Spectral performance evaluation of a second-generation spectral detector CT

Liu

Shapira

Halliburton

et al. 2022

Preprint

View full text Add to dashboard Cite

Objective To characterize a second-generation wide-detector dual-layer spectral computed tomography system for material quantification accuracy, acquisition parameter and patient size dependencies, and tissue characterization capabilities. Methods A phantom with multiple tissue-mimicking and material inserts was scanned with a dual-layer spectral detector CT using different tube voltages, collimation widths, radiation dose levels, and size configurations. Accuracy of iodine density maps and virtual monoenergetic images (MonoE) were investigated. Additionally, differences between conventional and MonoE 70 keV images were calculated to evaluate acquisition parameter and patient size dependencies. To demonstrate material quantification and differentiation, liver-mimicking inserts with adipose and iron were analyzed with a two-base decomposition utilizing MonoE 50 and 150 keV, and root mean square error (RMSE) for adipose and iron content was reported. Results Spectral accuracy was high for the measured inserts across a wide range of MonoE levels. MonoE 70 keV demonstrated reduced dependence compared to conventional image for phantom size (3 vs. 29 HU) and acquisition parameters, particularly tube voltage (5 vs. 43 HU) and noise-dose (9 vs. 11 HU). Iodine density quantification was successful with errors ranging from 0.25 to 0.5 mg/mL. Similarly, inserts with different amounts of adipose and iron were easily differentiated, and the small deviation in values within inserts corresponded to a RMSE of 1.78 ± 0.37% and 0.87 ± 0.37 mg/mL for adipose and iron content, respectively. Conclusion The second-generation dual-layer CT enables acquisition of quantitatively accurate spectral data without compromises from differences in patient size and acquisition parameters.

show abstract

Technical performance of a dual‐energy CT system with a novel deep‐learning based reconstruction process: Evaluation using an abdomen protocol

et al. 2022

View full text Add to dashboard Cite

Background: A new tube voltage-switching dual-energy (DE) CT system using a novel deep-learning based reconstruction process has been introduced. Characterizing the performance of this DE approach can help demonstrate its benefits and potential drawbacks. Purpose: To evaluate the technical performance of a novel DECT system and compare it to that of standard single-kV CT and a rotate/rotate DECT, for abdominal imaging. Methods: DE and single-kV images of four different phantoms were acquired on a kV-switching DECT system, and on a rotate/rotate DECT. The dose for the acquisitions of each phantom was set to that selected for the kV-switching DE mode by the automatic tube current modulation (ATCM) at manufacturerrecommended settings. The dose that the ATCM would have selected in single-kV mode was also recorded. Virtual monochromatic images (VMIs) from 40 to 130 keV, as well as iodine maps, were reconstructed from the DE data. Single-kV images, acquired at 120 kV, were reconstructed using body hybrid iterative reconstruction. All reconstructions were made at 0.5 mm section thickness. Task transfer functions (TTFs) were determined for a Teflon and LDPE rod. Noise magnitude (SD), and noise power spectrum (NPS) were calculated using 240 and 320 mm diameter water phantoms. Iodine quantification accuracy and contrast-to-noise ratios (CNRs) relative to water for 2, 5, 10, and 15 mg I/ml were determined using a multi-energy CT (MECT) phantom. Low-contrast visibility was determined and the presence of beam-hardening artifacts and inhomogeneities were evaluated. Results:The TTFs of the kV-switching DE VMIs were higher than that of the single-kV images for Teflon (20% TTF: 6.8 lp/cm at 40 keV, 6.2 lp/cm for single-kV), while for LDPE the DE TTFs at 70 keV and above were equivalent or higher than the single-kV TTF. All TTFs of the kV-switching DECT were higher than for the rotate/rotate DECT. The SD was lowest in the 70 keV VMI (12.0 HU), which was lower than that of single-kV (18.3 HU). The average NPS frequency varied between 2.3 lp/cm and 4.2 lp/cm for the kV-switching VMIs and was 2.2 lp/cm for single-kV. The error in iodine quantification was at maximum 1 mg I/ml (at 5 mg I/ml). The highest CNR for all iodine concentrations was at 60 keV, 2.5 times higher than the CNR for single-kV. At 70-90 keV, the number of visible low contrast objects was comparable to that in single-kV, while other VMIs showedThis is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

show abstract

Comparison of virtual monoenergetic imaging between a rapid kilovoltage switching dual-energy computed tomography with deep-learning and four dual-energy CTs with iterative reconstruction

Cited by 20 publications

References 33 publications

Comparison of image quality of two versions of deep-learning image reconstruction algorithm on a rapid kV-switching CT: a phantom study

Comparison of image quality of two versions of deep-learning image reconstruction algorithm on a rapid kV-switching CT: a phantom study

Spectral performance evaluation of a second-generation spectral detector CT

Technical performance of a dual‐energy CT system with a novel deep‐learning based reconstruction process: Evaluation using an abdomen protocol

Contact Info

Product

Resources

About