Deep learning reconstruction for detection of liver lesions at standard-dose and reduced-dose abdominal CT - Takeaways - MDSpire

Deep learning reconstruction for detection of liver lesions at standard-dose and reduced-dose abdominal CT

  • By

  • Tormund H. Njølstad

  • Kristin Jensen

  • Hilde K. Andersen

  • Audun E. Berstad

  • Gaute Hagen

  • Cathrine K. Johansen

  • Kjetil Øye

  • Jan Glittum

  • Anniken Dybwad

  • Emma Thingstad

  • Marianne G. Guren

  • Johann Baptist Dormagen

  • Anselm Schulz

  • April 19, 2025

  • 0 min

Share

  • 1

    Deep learning reconstruction (DLR) enhances image quality in low-dose abdominal CT while reducing image noise compared to traditional methods.

  • 2

    Iterative reconstruction (IR) improves image noise but may compromise low-contrast task performance, appearing more 'blotchy' than filtered back projection.

  • 3

    Clinical studies on DLR show improved subjective image quality, but results on its effectiveness for low-contrast lesion detection remain conflicting.

  • 4

    The study aims to evaluate diagnostic performance for liver lesion detection using DLR and IR across varying radiation dose levels.

  • 5

    A total of 235 liver lesions were required for the study, ensuring statistical power to assess non-inferiority in detection rates.

Original Source(s)

Deep learning reconstruction for detection of liver lesions at standard-dose and reduced-dose abdominal CT

  • by Tormund H. Njølstad, Kristin Jensen, Hilde K. Andersen, Audun E. Berstad, Gaute Hagen, Cathrine K. Johansen, Kjetil Øye, Jan Glittum, Anniken Dybwad, Emma Thingstad, Marianne G. Guren, Johann Baptist Dormagen, Anselm Schulz

  • April 19, 2025

Related Content

Spectral computed tomography characterization of coronary atherosclerotic plaque: principles, imaging biomarkers, and clinical significance

TOPOS: target organ prediction on scout views for automated CT scan planning

Combining radiomics and machine learning for enhanced localization of premature ventricular contractions