The Feasibility of a Model-Based Iterative Reconstruction Technique Tuned for the Myocardium on Myocardial Computed Tomography Late Enhancement
Hidetaka Toritani,Kazuki Yoshida,Takaaki Hosokawa,Yuki Tanabe,Yuta Yamamoto,Hikaru Nishiyama,Tomoyuki Kido,Naoto Kawaguchi,Megumi Matsuda,Shota Nakano,Shigehiro Miyazaki,Teruyoshi Uetani,Shinji Inaba,Osamu Yamaguchi,Teruhito Kido
DOI: https://doi.org/10.1097/RCT.0000000000001652
2024-08-02
Abstract:Objectives: This study evaluated the feasibility of a model-based iterative reconstruction technique (MBIR) tuned for the myocardium on myocardial computed tomography late enhancement (CT-LE). Methods: Twenty-eight patients who underwent myocardial CT-LE and late gadolinium enhancement (LGE) magnetic resonance imaging (MRI) within 1 year were retrospectively enrolled. Myocardial CT-LE was performed using a 320-row CT with low tube voltage (80 kVp). Myocardial CT-LE images were scanned 7 min after CT angiography (CTA) without additional contrast medium. All myocardial CT-LE images were reconstructed with hybrid iterative reconstruction (HIR), conventional MBIR (MBIR_cardiac), and new MBIR tuned for the myocardium (MBIR_myo). Qualitative (5-grade scale) scores and quantitative parameters (signal-to-noise ratio [SNR] and contrast-to-noise ratio [CNR]) were assessed as image quality. The sensitivity, specificity, and accuracy of myocardial CT-LE were evaluated at the segment level using an American Heart Association (AHA) 16-segment model, with LGE-MRI as a reference standard. These results were compared among the different CT image reconstructions. Results: In 28 patients with 448 segments, 160 segments were diagnosed with positive by LGE-MRI. In the qualitative assessment of myocardial CT-LE, the mean image quality scores were 2.9 ± 1.2 for HIR, 3.0 ± 1.1 for MBIR_cardiac, and 4.0 ± 1.0 for MBIR_myo. MBIR_myo showed a significantly higher score than HIR (P < 0.001) and MBIR_cardiac (P = 0.018). In the quantitative image quality assessment of myocardial CT-LE, the median image SNR was 10.3 (9.1-11.1) for HIR, 10.8 (9.8-12.1) for MBIR_cardiac, and 16.8 (15.7-18.4) for MBIR_myo. The median image CNR was 3.7 (3.0-4.6) for HIR, 3.8 (3.2-5.1) for MBIR_cardiac, and 6.4 (5.0-7.7) for MBIR_myo. MBIR_myo significantly improved the SNR and CNR of CT-LE compared to HIR and MBIR_cardiac (P < 0.001). The sensitivity, specificity, and accuracy for the detection of myocardial CT-LE were 70%, 92%, and 84% for HIR; 71%, 92%, and 85% for MBIR_cardiac; and 84%, 92%, and 89% for MBIR_myo, respectively. MBIR_myo showed significantly higher image quality, sensitivity, and accuracy than the others (P < 0.05). Conclusions: MBIR tuned for myocardium improved image quality and diagnostic performance for myocardial CT-LE assessment.