Purpose: To evaluate image quality, image noise and potential dose reduction of low-dose CT scans of the abdomen with deep-learning image reconstruction (DLIR) and to compare with images reconstructed with the adaptive statistical iterative reconstruction at a level of 50% (ASIR-V50%) and 100% (ASIR-V100%). Material and Methods: Consecutive patients who underwent abdominal enhanced CT were prospectively enrolled from 1st August to 30th September, 2021. Exclusion criteria were: contraindication to CT and severe motion artifacts on CT. Datasets were acquired with ASiR-V50% and then reconstructed with ASiR-V100% and DLIR at high levels (DLIR-H), and data regarding radiation dose were collected. Two radiologists in consensus performed the objective image quality analysis of images using standardized regions of interest (ROIs) to record mean attenuation value and standard deviation (SD) in Hounsfield units (HU) for the liver, aorta, portal vein and muscle. Contrast-to-noise (CNR) and signal-to-noise (SNR) were assessed. Two radiologists independently evaluated subjective image contrast, image noise and conspicuity of structures using a 5-point Likert scale. Results: Sixty patients were enrolled (39 male, mean age 67±13y). SNR of liver parenchyma in arterial phase was significantly higher for DLIR compared to ASIR-V100% (p=0.04), whereas no significant differences were observed for SNR of the liver in the portal phase (p=0.09). No significant differences were observed in terms of CNR of the liver parenchyma in arterial phase between DLIR and ASIR-V100% (p=0.52), whereas DLIR CNR was significantly higher in portal phase (p=0.04). According to the subjective analysis, DLIR had higher image contrast, lower image noise and better conspicuity of structures than ASIR-V100% (all p<0.001), with excellent inter-rater agreement (k=0.81). With respect to ASIR-V100%, radiation dose was significantly lower in DLIR (p=0.031). Conclusion: DLIR significantly improved the image quality and reduced image noise compared to ASIR-V100% while reducing the radiation dose.

Deep-learning image reconstruction for low-dose liver CT / Del Gaudio, A; Guido, G; Caruso, D; Ubaldi, N; Valanzuolo, D; Bona, G; Pugliese, D; Laghi, A. - (2022). (Intervento presentato al convegno ESGAR Annual meeting and Postgraduate Course 2022 tenutosi a Lisbon; Portugal).

Deep-learning image reconstruction for low-dose liver CT

Del Gaudio A;Guido G;Caruso D;Ubaldi N;Valanzuolo D;Bona G;Pugliese D;Laghi A
2022

Abstract

Purpose: To evaluate image quality, image noise and potential dose reduction of low-dose CT scans of the abdomen with deep-learning image reconstruction (DLIR) and to compare with images reconstructed with the adaptive statistical iterative reconstruction at a level of 50% (ASIR-V50%) and 100% (ASIR-V100%). Material and Methods: Consecutive patients who underwent abdominal enhanced CT were prospectively enrolled from 1st August to 30th September, 2021. Exclusion criteria were: contraindication to CT and severe motion artifacts on CT. Datasets were acquired with ASiR-V50% and then reconstructed with ASiR-V100% and DLIR at high levels (DLIR-H), and data regarding radiation dose were collected. Two radiologists in consensus performed the objective image quality analysis of images using standardized regions of interest (ROIs) to record mean attenuation value and standard deviation (SD) in Hounsfield units (HU) for the liver, aorta, portal vein and muscle. Contrast-to-noise (CNR) and signal-to-noise (SNR) were assessed. Two radiologists independently evaluated subjective image contrast, image noise and conspicuity of structures using a 5-point Likert scale. Results: Sixty patients were enrolled (39 male, mean age 67±13y). SNR of liver parenchyma in arterial phase was significantly higher for DLIR compared to ASIR-V100% (p=0.04), whereas no significant differences were observed for SNR of the liver in the portal phase (p=0.09). No significant differences were observed in terms of CNR of the liver parenchyma in arterial phase between DLIR and ASIR-V100% (p=0.52), whereas DLIR CNR was significantly higher in portal phase (p=0.04). According to the subjective analysis, DLIR had higher image contrast, lower image noise and better conspicuity of structures than ASIR-V100% (all p<0.001), with excellent inter-rater agreement (k=0.81). With respect to ASIR-V100%, radiation dose was significantly lower in DLIR (p=0.031). Conclusion: DLIR significantly improved the image quality and reduced image noise compared to ASIR-V100% while reducing the radiation dose.
2022
ESGAR Annual meeting and Postgraduate Course 2022
04 Pubblicazione in atti di convegno::04d Abstract in atti di convegno
Deep-learning image reconstruction for low-dose liver CT / Del Gaudio, A; Guido, G; Caruso, D; Ubaldi, N; Valanzuolo, D; Bona, G; Pugliese, D; Laghi, A. - (2022). (Intervento presentato al convegno ESGAR Annual meeting and Postgraduate Course 2022 tenutosi a Lisbon; Portugal).
File allegati a questo prodotto
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/1645029
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact