Zhuoqing Chen1, Kenichi Nakajima2, Ken-Ichi Hirano3, Takashi Kamiya4, Shohei Yoshida5, Shintaro Saito1, Seigo Kinuya1. 1. Department of Nuclear Medicine, Kanazawa University, Kanazawa, Ishikawa, Japan. 2. Department of Functional Imaging and Artificial Intelligence, Kanazawa University, Kanazawa, Ishikawa, 920-8640, Japan. nakajima@med.kanazawa-u.ac.jp. 3. Department of Triglyceride Science, Graduate School of Medicine, Osaka University, Suita, Osaka, Japan. 4. Department of Medical Technology, Osaka University Hospital, Suita, Osaka, Japan. 5. Department of Cardiology, Kanazawa University Hospital, Kanazawa, Ishikawa, Japan.
Abstract
OBJECTIVE: This study aimed to optimize various methods of calculating washout rates (WRs) of 123I-β-methyl-p-iodophenyl-pentadecanoic (BMIPP), as they are essential to diagnose triglyceride deposit cardiomyovasculopathy (TGCV) which is a rare disease entity identified in Japan and has been encoded in Orphanet (ORPHA code 565612). METHODS: We calculated WRs of 123I-BMIPP from early (20 min) and delayed (200 min) images. We evaluated six methods of calculating WRs to discriminate TGVC patients (age, 56.8 ± 14.6 y; male, n = 13; female, n = 4) and 21 123I-BMIPP studies were involved including 4 follow-up studies. Washout rates were calculated by two planar methods using anterior images with cardiac and background regions of interest (ROIs) and by four SPECT methods using either array and polar plots or summed short-axis images. The final diagnoses of TGCV were confirmed according to the 2020 diagnostic criteria, and the diagnostic accuracy of WRs calculated using the six methods was analyzed using the area under receiver-operating characteristics curves (ROC-AUC). Multiple scatter-plot matrix methods were evaluated with correlations for comparison. RESULTS: All six methods were useful for diagnosis and did not significantly differ. The four SPECT methods showed excellent diagnostic accuracy (AUC 1.0), whereas the planar methods with and without background correction could be acceptable (AUC 0.857 and 0.964, respectively). The WRs were relatively lower for patients with CAD and remarkable metabolic defects than for patients with TGCV but without defects. CONCLUSIONS: For the diagnosis of TGCV, the WR cutoff of 10% of 123I-BMIPP functioned well in planar and SPECT discrimination based on computational methods as a classifier. However, calculation optimization should improve TGCV diagnoses.
OBJECTIVE: This study aimed to optimize various methods of calculating washout rates (WRs) of 123I-β-methyl-p-iodophenyl-pentadecanoic (BMIPP), as they are essential to diagnose triglyceride deposit cardiomyovasculopathy (TGCV) which is a rare disease entity identified in Japan and has been encoded in Orphanet (ORPHA code 565612). METHODS: We calculated WRs of 123I-BMIPP from early (20 min) and delayed (200 min) images. We evaluated six methods of calculating WRs to discriminate TGVC patients (age, 56.8 ± 14.6 y; male, n = 13; female, n = 4) and 21 123I-BMIPP studies were involved including 4 follow-up studies. Washout rates were calculated by two planar methods using anterior images with cardiac and background regions of interest (ROIs) and by four SPECT methods using either array and polar plots or summed short-axis images. The final diagnoses of TGCV were confirmed according to the 2020 diagnostic criteria, and the diagnostic accuracy of WRs calculated using the six methods was analyzed using the area under receiver-operating characteristics curves (ROC-AUC). Multiple scatter-plot matrix methods were evaluated with correlations for comparison. RESULTS: All six methods were useful for diagnosis and did not significantly differ. The four SPECT methods showed excellent diagnostic accuracy (AUC 1.0), whereas the planar methods with and without background correction could be acceptable (AUC 0.857 and 0.964, respectively). The WRs were relatively lower for patients with CAD and remarkable metabolic defects than for patients with TGCV but without defects. CONCLUSIONS: For the diagnosis of TGCV, the WR cutoff of 10% of 123I-BMIPP functioned well in planar and SPECT discrimination based on computational methods as a classifier. However, calculation optimization should improve TGCV diagnoses.