Simultaneous assessment of heart and lungs with gated high-pitch ultra-low dose chest CT using artificial intelligence-based calcium scoring

Florian Andre; Sebastian Seitz; Philipp Fortner; Thomas Allmendinger; André Sommer; Matthias Brado; Roman Sokiranski; Joana Fink; Hans-Ulrich Kauczor; Claus P Heussel; Felix Herth; Norbert Frey; Johannes Görich; Sebastian J Buss

doi:10.1016/j.ejro.2023.100481

Simultaneous assessment of heart and lungs with gated high-pitch ultra-low dose chest CT using artificial intelligence-based calcium scoring

Eur J Radiol Open. 2023 Feb 15:10:100481. doi: 10.1016/j.ejro.2023.100481. eCollection 2023.

Authors

Florian Andre^{1

2}, Sebastian Seitz², Philipp Fortner², Thomas Allmendinger³, André Sommer², Matthias Brado², Roman Sokiranski², Joana Fink⁴, Hans-Ulrich Kauczor⁵, Claus P Heussel^{6

7}, Felix Herth^{8

7}, Norbert Frey¹, Johannes Görich², Sebastian J Buss²

Affiliations

¹ University of Heidelberg, Department of Cardiology, Angiology and Pneumology, Heidelberg, Germany.
² MVZ-DRZ Heidelberg, Heidelberg, Germany.
³ Siemens Healthcare, Forchheim, Germany.
⁴ AstraZeneca, Wedel, Germany.
⁵ University of Heidelberg, Department of Diagnostic and Interventional Radiology, Heidelberg.
⁶ University of Heidelberg, Thoraxklinik, Department of Diagnostic and Interventional Radiology with Nuclear Medicine, Heidelberg, Germany.
⁷ Translational Lung Research Centre Heidelberg, Member of the German Centre for Lung Research (DZL), Heidelberg, Germany.
⁸ University of Heidelberg, Thoraxklinik, Department of Pneumology and Critical Care Medicine, Heidelberg, Germany.

Abstract

Purpose: The combined testing for coronary artery and pulmonary diseases is of clinical interest as risk factors are shared. In this study, a novel ECG-gated tin-filtered ultra-low dose chest CT protocol (GCCT) for integrated heart and lung acquisition and the applicability of artificial intelligence (AI)-based coronary artery calcium scoring were assessed.

Methods: In a clinical registry of 10481 patients undergoing heart and lung CT, GCCT was applied in 44 patients on a dual-source CT. Coronary calcium scans (CCS) with 120 kVp, 100 kVp, and tin-filtered 100 kVp (Sn100) of controls, matched with regard to age, sex, and body-mass index, were retrieved from the registry (n_total=176, 66.5 (59.4-74.0) years, 52 men). Automatic tube current modulation was used in all scans. In 20 patients undergoing GCCT and Sn100 CCS, Agatston scores were measured both semi-automatically by experts and by AI, and classified into six groups (0, <10, <100, <400, <1000, ≥1000).

Results: Effective dose decreased significantly from 120 kVp CCS (0.50 (0.41-0.61) mSv) to 100 kVp CCS (0.34 (0.26-0.37) mSv) to Sn100 CCS (0.14 (0.11-0.17) mSv). GCCT showed higher values (0.28 (0.21-0.32) mSv) than Sn100 CCS but lower than 120 kVp and 100 kVp CCS (all p < 0.05) despite greater scan length. Agatston scores correlated strongly between GCCT and Sn100 CCS in semi-automatic and AI-based measurements (both ρ = 0.98, p < 0.001) resulting in high agreement in Agatston score classification (κ = 0.97, 95% CI 0.92-1.00; κ = 0.89, 95% CI 0.79-0.99). Regarding chest findings, further diagnostic steps were recommended in 28 patients.

Conclusions: GCCT allows for reliable coronary artery disease and lung cancer screening with ultra-low radiation exposure. GCCT-derived Agatston score shows excellent agreement with standard CCS, resulting in equivalent risk stratification.

Keywords: Artificial intelligence; Chest CT; Coronary artery calcium scoring; Coronary artery disease; Dose reduction; Tin-filter.