Abstract
Purpose: To test the influence of different levels of model-based iterative
reconstruction on observed image quality in high-resolution chest CT, acquired
at 150 kVp with spectral tin-filtration.
Methods or Backround: Informed consent was obtained for the use of 44
inspiratory HRCT scans for processing and analysis. Scans were performed on
a 3rd generation Dual Source CT, with 150 kVp tin-filtered X-ray spectrum. All
datasets were reconstructed using filtered back-projection (FBP) and 3
different levels of model-based iterative reconstruction (ADMIRE). All other
recon parameters were identical. Two radiologists reviewed images blinded,
independently, and in random order. Images were scored on a 5-point Likert
scale according to 4 pre-defined criteria: (A) visualization of interlobular septae,
(B) pleural and subpleural regions, (C) centrilobular region, and (D) bronchi
and vessels. Results were analysed with visual grading characteristics (VGC),
using a VGC analyser. The 3 ADMIRE levels were compared to FBP, with area
under the curve (AUC) and 95% confidence intervals. Interreader
reproducibility was assessed with weighted kappa statistics.
Results or Findings: For ADMIRE level 1, AUC did not differ significantly from
0.5 for all criteria. For ADMIRE levels 3 and 5, AUC were 0.64 and 0.91 for
criterion A, 0.68 and 0.93 for criterion B, 0.65 and 0.86 for criterion C, and 0.62
and 0.83 for criterion D. For all criteria, 95% confidence intervals did not
overlap, indicating the statistically significant difference. Weighted kappa was
between 0.2 and 0.3 for all criteria.
Conclusion: ADMIRE level 5 provided superior visualization of structures,
compared to FBP and other ADMIRE levels.
Limitations: We found a relatively low agreement and not all available
ADMIRE levels were tested.
Ethics committee approval: No information provided by the submitter.
Funding for this study: No funding was received for this study.
Author Disclosures:
reconstruction on observed image quality in high-resolution chest CT, acquired
at 150 kVp with spectral tin-filtration.
Methods or Backround: Informed consent was obtained for the use of 44
inspiratory HRCT scans for processing and analysis. Scans were performed on
a 3rd generation Dual Source CT, with 150 kVp tin-filtered X-ray spectrum. All
datasets were reconstructed using filtered back-projection (FBP) and 3
different levels of model-based iterative reconstruction (ADMIRE). All other
recon parameters were identical. Two radiologists reviewed images blinded,
independently, and in random order. Images were scored on a 5-point Likert
scale according to 4 pre-defined criteria: (A) visualization of interlobular septae,
(B) pleural and subpleural regions, (C) centrilobular region, and (D) bronchi
and vessels. Results were analysed with visual grading characteristics (VGC),
using a VGC analyser. The 3 ADMIRE levels were compared to FBP, with area
under the curve (AUC) and 95% confidence intervals. Interreader
reproducibility was assessed with weighted kappa statistics.
Results or Findings: For ADMIRE level 1, AUC did not differ significantly from
0.5 for all criteria. For ADMIRE levels 3 and 5, AUC were 0.64 and 0.91 for
criterion A, 0.68 and 0.93 for criterion B, 0.65 and 0.86 for criterion C, and 0.62
and 0.83 for criterion D. For all criteria, 95% confidence intervals did not
overlap, indicating the statistically significant difference. Weighted kappa was
between 0.2 and 0.3 for all criteria.
Conclusion: ADMIRE level 5 provided superior visualization of structures,
compared to FBP and other ADMIRE levels.
Limitations: We found a relatively low agreement and not all available
ADMIRE levels were tested.
Ethics committee approval: No information provided by the submitter.
Funding for this study: No funding was received for this study.
Author Disclosures:
| Originalsprog | Engelsk |
|---|---|
| Artikelnummer | RPS 414-9 |
| Tidsskrift | Insights into Imaging |
| Vol/bind | 14 |
| Udgave nummer | Suppl. 4 |
| Sider (fra-til) | 47 |
| Antal sider | 1 |
| ISSN | 1869-4101 |
| DOI | |
| Status | Udgivet - 2023 |
| Begivenhed | European Congress of Radiology - VIC, Wien, Østrig Varighed: 1. mar. 2023 → 5. mar. 2023 http://myesr.org |
Konference
| Konference | European Congress of Radiology |
|---|---|
| Lokation | VIC |
| Land/Område | Østrig |
| By | Wien |
| Periode | 01/03/2023 → 05/03/2023 |
| Internetadresse |