Whole-body reduced-dose dual-energy CT with deep learning image reconstruction for detection of osteolytic lesions in multiple myeloma.
Authors
Affiliations (4)
Affiliations (4)
- Department of Radiology, The First Affiliated Hospital of Anhui Medical University, Research Center of Clinical Medical Imaging, Anhui Province Clinical Image Quality Control Center, No.218 Jixi Road, Hefei, Anhui, 230022, China.
- Department of Cardiology, The First Affiliated Hospital of Anhui Medical University, No. 218 Jixi Road, Hefei, Anhui, 230022, China.
- Department of Hematology, The First Affiliated Hospital of Anhui Medical University, No. 218 Jixi Road, Hefei, Anhui, 230022, China. Electronic address: [email protected].
- Department of Radiology, The First Affiliated Hospital of Anhui Medical University, Research Center of Clinical Medical Imaging, Anhui Province Clinical Image Quality Control Center, No.218 Jixi Road, Hefei, Anhui, 230022, China. Electronic address: [email protected].
Abstract
Deep learning image reconstruction (DLIR) has been incorporated into dual-energy CT (DECT) to improve image quality. However, its applications in reduced-dose DECT for evaluating multiple myeloma remain unclear. This study aimed to evaluate image quality and osteolytic lesion detectability of reduced-dose DECT with DLIR, compared with routine-dose single-energy CT (SECT) with adaptive statistical iterative reconstruction-Veo (ASIR-V). This prospective study enrolled consecutive participants with known or precursor multiple myeloma from July 2023 to October 2024. Each participant underwent whole-body non-contrast SECT (120-kVp images; ASIR-V 40% [AR40]), followed by DECT (74- and 50-keV virtual monochromatic images [VMIs], material decomposition [MD] images; high-strength DLIR [DH]). Image noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and bone-to-lesion CNR (BLR) were calculated. Three radiologists assessed osteolytic lesion detectability and image quality. Lesion detection performance was analyzed using mixed-effects logistic regression. A total of 43 participants (mean age, 68 ± 10 years; 31 men) with 121 osteolytic lesions were included. DECT reduced radiation dose by 53% versus SECT. Compared with AR40 120-kVp images, DH 74- and 50-keV VMIs exhibited lower and comparable image noise, respectively, and improved CNR, SNR, BLR, and overall image quality (p < 0.05). DH 50-keV VMIs and MD images demonstrated superior lesion conspicuity (p < 0.001), and DH MD images achieved higher sensitivity and specificity than AR40 120-kVp images (p < 0.05). Whole-body reduced-dose multiparametric DECT with DLIR provides improved image quality and osteolytic lesion detectability compared with routine-dose SECT, with a 53% reduction in radiation dose. Multiparametric DECT with DLIR provides a more dose-efficient imaging method for visualizing osteolytic lesions in multiple myeloma compared with standard 120-kVp protocols.