.png){kind=icon}
.png){kind=icon}
Return
Deep learning-derived arterial input function for dynamic brain PET
.png){kind=icon}
OA
AI
DOI：10.1016/j.neuroimage.2025.121609.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Developed DLIF, a novel deep learning framework for fully non-invasive estimation of arterial input functions in dynamic PET imaging. • Employs Vision Transformer and learned basis functions for continuous, close-form, subject-specific AIF estimation. • Demonstrated superior accuracy over traditional image-derived and decomposition-based methods, achieving state-of-the-art kinetic modeling performance. • Introduced a novel evaluation metric designed specifically for quantifying the accuracy of estimated AIFs relevant for downstream kinetic modeling analyses, such as Logan graphical analysis. • DLIF offers potential for broad clinical and research applications, significantly advancing dynamic PET neuroimaging workflows.
Keywords:
Arterial input function
Dynamic PET
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.5
Papers:
1.3K
Citations:
10.6W