PhD Scientific Days 2026

Poster Session 1.P - Cardiovascular Medicine and Research

Impact of Novel Reconstruction Algorithm on Image Quality, Reader Agreement, and Diagnostic Confidence in Coronary CT Angiography

Előadó neve

Dr. Nagy, Kristóf

Neptune code

H7BUPL

Előadó munkahelye

Semmelweis University, Medical Imaging Centre

Előadó telefonszáma

06305504988

Előadó e-mail címe

nagy.kristof@semmelweis.hu

Az előadás címe

Impact of Novel Reconstruction Algorithm on Image Quality, Reader Agreement, and Diagnostic Confidence in Coronary CT Angiography

Szerző(k) neve és munkahelye

Kristóf Nagy¹, Krisztián Kocsis¹, Gergő Veréb¹, Miklós Vértes¹, Bartykowszky Andrea², Lili Száraz¹, Sámuel Beke¹, István Csulak¹, Réka Sebestény-Dósa¹, Barnabás Baksa¹, Pál Maurovich Horvat¹

1: Medical Imaging Centre, Semmelweis University
2: Medical Imaging Centre, Semmelweis University; Heart and Vascular Center, Semmelweis University

Bemutatás módja

Poszter

Szekció

Poster Session 1.P - Cardiovascular Medicine and Research

Language of the presentation

Hungarian

Preferred session

Cardiovascular Medicine and Research

Összefoglaló szövege

Aims: The novel ZeeFree (ZF) reconstruction algorithm has been shown to reduce stair-step artefacts and improve motion-related image quality in coronary CT angiography (CCTA) using dual-source photon-counting detector CT [1-2]. Recently, a single-source PCD-CT system was introduced alongside a new software version (VB20). This study aimed to evaluate the algorithm’s performance on this updated system by assessing qualitative image parameters.
Methods: Patients who underwent CCTA on the single-source PCD-CT between October 2024 and March 2025 were retrospectively included. Individuals with a history of prior coronary revascularisation were excluded. The automatically identified best systolic or diastolic phase was selected, and curved multiplanar reformatted (CMPR) images of the three main coronary arteries were generated using both TrueStack (TS) and ZF reconstructions. TS enables sharp delineation of interfaces between stack volumes. Four readers with varying levels of experience rated image quality on a Likert scale. Motion artefacts were rated 1-5, step artefacts 1-4, and diagnostic confidence 1-5, with higher scores indicating better image quality. The effects of reconstruction types, image quality metrics, vessels, and readers were evaluated using linear mixed-effect models. Multiple agreement metrics were used to assess agreement with the most experienced reader.
Results: 102 (52.9% female) patients were included in the final analysis. Significant effects of image quality metrics, reconstruction types, readers, and vessels were found on scores. The interaction between reconstruction type and image quality score was highly significant (F(2.7145) = 181.22, p<0.01). In the post-hoc analysis, ZF reconstruction showed a significant improvement in step artefacts (Δ=0.94, p<0.01) and diagnostic confidence (Δ=0.065, p=0.026). Interrater reliability was good for both reconstructions (TS: ICC = 0.713; ZF: ICC = 0.657). ZF improved agreement with the most experienced reader (Rater 3) with all readers, particularly those with less experience (Exact agreement: Rater 1 TS: 0.598, ZF: 0.673; Rater 2 TS: 0.433, ZF: 0.541; Rater 4 TS: 0.616, TF: 0.687).
Conclusions: ZF significantly decreased stair-step artefact and improved diagnostic confidence on the single-source PCD-CT. ZF also enables better agreement with expert readers.
Funding: PhD students received SE250+.

University

Semmelweis University

Supervisor

Pál Maurovich Horvat

Publication of my abstract

I do not give consent to the publication of my abstract on the website of the congress.

phd.section.field

in doctoral studies after complex exam (PhD)

Kind

Szabad

Status

elfogadva

Accepted presentation method

poszter

Előadás fájl jóváhagyás

nem rendelkezett róla

Előadó

9089

Start

17:24

End

17:27