Quantitative imaging in medicine and surgery · 2025
Comparison of image quality in 40 keV virtual monoenergetic images of dual-energy CT pulmonary angiography using deep learning and iterative reconstruction algorithms under optimized low dose scanning protocols.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Zhang Dapeng, Zhang Lulu, Long Juan, Wu Yang, Zhang He, Wang Chong, Sun Bo, Wang Chenzi, Zhang He, Sun Xiaonan, Sun Aiyun, Meng Yankai, Hu Chunfeng, Xu Kai
Аннотация
BACKGROUND: Pulmonary embolism is a potentially fatal cardiovascular condition that demands prompt and accurate diagnostic imaging. Traditional single-energy computed tomography pulmonary angiography (CTPA), while widely used, is associated with high radiation doses and substantial volumes of contrast agents, which may increase the risks of radiation-induced tissue damage and contrast-induced nephropathy (CIN), respectively. Dual-energy CTPA (DE-CTPA) presents a promising alternative, though challenges, including elevated image noise at low kilo-electron volt (keV) levels (e.g., 40 keV), persist. The primary aim of this study is to evaluate and compare the image quality of 40 keV virtual monoenergetic images (VMI) reconstructed using deep learning image reconstruction (DLIR) and Adaptive Statistical Iterative Reconstruction-V (ASIR-V) algorithms within the context of low-dose DE-CTPA protocols.
METHODS: This prospective study enrolled patients who underwent DE-CTPA between January and April 2025. Using a Revolution CT scanner, 40 keV VMI were reconstructed with four distinct algorithms: ASIR-V 50%, ASIR-V 70%, Deep learning image reconstruction with medium setting (DLIR-M), and deep learning image reconstruction with high setting (DLIR-H). Iodixanol (350 mgI/mL) was administered at a dose of 0.4 mL/kg. The image quality was assessed through both objective measures [image noise, contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR)] and subjective evaluation via a Likert scale. Statistical analysis was conducted using SPSS 27.0, employing analysis of variance (ANOVA) for normally distributed data and the Kruskal-Wallis test for non-normally distributed data.
RESULTS: A total of 75 patients with clinical suspicion of pulmonary embolism were included in the study. The mean effective dose (ED) was 3.76±1.02 mSv, with a mean CT volume dose index (CTDIvol) of 6.13±1.69 mGy and a mean dose-length product (DLP) of 221.12±59.85 mGy·cm. The mean contrast agent volume was 26.0±5.0 mL. Statistical analysis of image quality revealed significant differences between the four groups in terms of image noise, CNR, and SNR, measured at the levels of the main pulmonary artery, left pulmonary artery, and right pulmonary artery (P<0.001). Post-hoc analysis demonstrated that the DLIR-H algorithm provided the highest image quality, significantly reducing noise while enhancing CNR and SNR relative to both ASIR-V and DLIR-M (P<0.001). Compared with ASIR-V 50%, DLIR-H reduced image noise by 45% at the PA [24.25±16.18 vs. 44.49±18.18 Hounsfield unit (HU)], 37% at the LPA (31.16±16.16 vs. 49.54±15.99 HU), and 40% at the RPA (29.99±15.96 vs. 49.94±16.48 HU) (all P<0.001). Correspondingly, DLIR-H yielded higher CNR values (46.88±21.33 vs. 24.40±10.41 at PA; 39.16±18.72 vs. 22.59±9.52 at LPA; 39.17±15.20 vs. 22.12±8.12 at RPA) and higher SNR values (50.21±21.95 vs. 26.17±10.71 at PA; 32.88±14.27 vs. 24.18±9.84 at LPA; 41.96±15.89 vs. 23.71±8.47 at RPA) (all P<0.001). Subjectively, DLIR-H achieved the highest median scores (5.0) for noise, spatial resolution, noise texture, and overall image quality, significantly outperforming both ASIR-V 50% and 70% (P<0.001).
CONCLUSIONS: The DLIR-H algorithm significantly enhances the image quality of 40 keV VMI images under low-dose DE-CTPA scanning protocols. It outperforms DLIR-M, ASIR-V 50%, and ASIR-V 70%, making it a promising tool for improving image quality in CTPA, particularly in clinical settings where minimizing radiation dose and contrast agent volume is essential.
METHODS: This prospective study enrolled patients who underwent DE-CTPA between January and April 2025. Using a Revolution CT scanner, 40 keV VMI were reconstructed with four distinct algorithms: ASIR-V 50%, ASIR-V 70%, Deep learning image reconstruction with medium setting (DLIR-M), and deep learning image reconstruction with high setting (DLIR-H). Iodixanol (350 mgI/mL) was administered at a dose of 0.4 mL/kg. The image quality was assessed through both objective measures [image noise, contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR)] and subjective evaluation via a Likert scale. Statistical analysis was conducted using SPSS 27.0, employing analysis of variance (ANOVA) for normally distributed data and the Kruskal-Wallis test for non-normally distributed data.
RESULTS: A total of 75 patients with clinical suspicion of pulmonary embolism were included in the study. The mean effective dose (ED) was 3.76±1.02 mSv, with a mean CT volume dose index (CTDIvol) of 6.13±1.69 mGy and a mean dose-length product (DLP) of 221.12±59.85 mGy·cm. The mean contrast agent volume was 26.0±5.0 mL. Statistical analysis of image quality revealed significant differences between the four groups in terms of image noise, CNR, and SNR, measured at the levels of the main pulmonary artery, left pulmonary artery, and right pulmonary artery (P<0.001). Post-hoc analysis demonstrated that the DLIR-H algorithm provided the highest image quality, significantly reducing noise while enhancing CNR and SNR relative to both ASIR-V and DLIR-M (P<0.001). Compared with ASIR-V 50%, DLIR-H reduced image noise by 45% at the PA [24.25±16.18 vs. 44.49±18.18 Hounsfield unit (HU)], 37% at the LPA (31.16±16.16 vs. 49.54±15.99 HU), and 40% at the RPA (29.99±15.96 vs. 49.94±16.48 HU) (all P<0.001). Correspondingly, DLIR-H yielded higher CNR values (46.88±21.33 vs. 24.40±10.41 at PA; 39.16±18.72 vs. 22.59±9.52 at LPA; 39.17±15.20 vs. 22.12±8.12 at RPA) and higher SNR values (50.21±21.95 vs. 26.17±10.71 at PA; 32.88±14.27 vs. 24.18±9.84 at LPA; 41.96±15.89 vs. 23.71±8.47 at RPA) (all P<0.001). Subjectively, DLIR-H achieved the highest median scores (5.0) for noise, spatial resolution, noise texture, and overall image quality, significantly outperforming both ASIR-V 50% and 70% (P<0.001).
CONCLUSIONS: The DLIR-H algorithm significantly enhances the image quality of 40 keV VMI images under low-dose DE-CTPA scanning protocols. It outperforms DLIR-M, ASIR-V 50%, and ASIR-V 70%, making it a promising tool for improving image quality in CTPA, particularly in clinical settings where minimizing radiation dose and contrast agent volume is essential.