医学
舟状骨骨折
置信区间
核医学
图像质量
放射科
断层摄影术
射线照相术
计算机科学
内科学
图像(数学)
人工智能
作者
Nina Kämmerling,Erik Tesselaar,Ronald Booij,Lotta Fornander,Anders Persson,Simon Farnebo
标识
DOI:10.1016/j.ejrad.2024.111383
摘要
Abstract
Purpose
Scaphoid fractures in patients and assessment of healing using PCD-CT have, as far as we know, not yet been studied. Therefore, the aim was to compare photon counting Detector CT (PCD-CT) with energy integrating detector CT (EID-CT) in terms of fracture visibility and evaluation of fracture healing. Method
Eight patients with scaphoid fracture were examined with EID-CT and PCD-CT within the first week post-trauma, and with additional scans at 4, 6 and 8 weeks. Our clinical protocol for wrist examination with EID-CT was used (CTDIvol 3.1 ± 0.1 mGy, UHR kernel Ur77. For PCD-CT matched radiation dose, reconstruction kernel Br89. Quantitative analyses of noise, CNR, trabecular and cortical sharpness, and bone volume fraction were conducted. Five radiologists evaluated the images for fracture visibility, fracture gap consolidation and image quality, and rated their confidence in the diagnosis. Results
The trabecular and cortical sharpness were superior in images obtained with PCD-CT compared with EID-CT. A successive reduction in trabecular bone volume fraction during the immobilized periods was found with both systems. Despite higher noise and lower CNR with PCD-CT, radiologists rated the image quality of PCD-CT as superior. The visibility of the fracture line within 1-week post-trauma was rated higher with PCD-CT as was diagnostic confidence, but the subsequent assessments of fracture gap consolidation during healing process and the confidence in diagnosis were found equivalent between both systems. Conclusion
PCD-CT offers superior visibility of bone microstructure compared with EID-CT. The evaluation of fracture healing and confidence in diagnosis were rated equally with both systems, but the radiologists found primary fracture visibility and overall image quality superior with PCD-CT.
科研通智能强力驱动
Strongly Powered by AbleSci AI