M Ohana1, A Labani2, F Severac3, M Y Jeung4, S Gaertner5, T Caspar6, C Roy7. 1. iCube Laboratory, Université de Strasbourg/CNRS, UMR 7357, 67400 Illkirch, France; Service de Radiologie B, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg, 1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: mickael.ohana@gmail.com. 2. Service de Radiologie B, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg, 1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: aissam.labani@chru-strasbourg.fr. 3. Département de Biostatistiques et d'Informatique Médicale, Hôpital Civil - Hôpitaux Universitaires de Strasbourg,1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: francois.severac@chru-strasbourg.fr. 4. Service de Radiologie B, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg, 1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: Mi-Young.Jeung@chru-strasbourg.fr. 5. Service de Médecine Vasculaire, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg,1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: Sebastien.Gaertner@chru-strasbourg.fr. 6. Service de Cardiologie, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg,1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: thibault.caspar@chru-strasbourg.fr. 7. Service de Radiologie B, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg, 1 place de l'hôpital, 67000 Strasbourg, France. Electronic address: Catherine.Roy@chru-strasbourg.fr.
Abstract
OBJECTIVE: To determine the optimal monochromatic energy level for lung parenchyma analysis in spectral CT. METHODS: All 50 examinations (58% men, 64.8±16yo) from an IRB-approved prospective study on single-source dual energy chest CT were retrospectively included and analyzed. Monochromatic images in lung window reconstructed every 5keV from 40 to 140keV were independently assessed by two chest radiologists. Based on the overall image quality and the depiction/conspicuity of parenchymal lesions, each reader had to designate for every patient the keV level providing the best diagnostic and image quality. RESULTS: 72% of the examinations exhibited parenchymal lesions. Reader 1 picked the 55keV monochromatic reconstruction in 52% of cases, 50 in 30% and 60 in 18%. Reader 2 chose 50keV in 52% cases, 55 in 40%, 60 in 6% and 40 in 2%. The 50 and 55keV levels were chosen by at least one reader in 64% and 76% of all patients, respectively. Merging 50 and 55keV into one category results in an optimal setting selected by reader 1 in 82% of patients and by reader 2 in 92%, with a 74% concomitant agreement. CONCLUSION: The best image quality for lung parenchyma in spectral CT is obtained with the 50-55keV monochromatic reconstructions.
OBJECTIVE: To determine the optimal monochromatic energy level for lung parenchyma analysis in spectral CT. METHODS: All 50 examinations (58% men, 64.8±16yo) from an IRB-approved prospective study on single-source dual energy chest CT were retrospectively included and analyzed. Monochromatic images in lung window reconstructed every 5keV from 40 to 140keV were independently assessed by two chest radiologists. Based on the overall image quality and the depiction/conspicuity of parenchymal lesions, each reader had to designate for every patient the keV level providing the best diagnostic and image quality. RESULTS: 72% of the examinations exhibited parenchymal lesions. Reader 1 picked the 55keV monochromatic reconstruction in 52% of cases, 50 in 30% and 60 in 18%. Reader 2 chose 50keV in 52% cases, 55 in 40%, 60 in 6% and 40 in 2%. The 50 and 55keV levels were chosen by at least one reader in 64% and 76% of all patients, respectively. Merging 50 and 55keV into one category results in an optimal setting selected by reader 1 in 82% of patients and by reader 2 in 92%, with a 74% concomitant agreement. CONCLUSION: The best image quality for lung parenchyma in spectral CT is obtained with the 50-55keV monochromatic reconstructions.
Authors: Hoon Ko; Heewon Chung; Wu Seong Kang; Kyung Won Kim; Youngbin Shin; Seung Ji Kang; Jae Hoon Lee; Young Jun Kim; Nan Yeol Kim; Hyunseok Jung; Jinseok Lee Journal: J Med Internet Res Date: 2020-06-29 Impact factor: 5.428