November 6, 2021Journal of Magnetic Resonance ImagingOpen Access

Improving Sensitivity of Arterial Spin Labeling Perfusion MRI in Alzheimer's Disease Using Transfer Learning of Deep Learning‐Based ASL Denoising

View Full Paper

Ask AI

Bookmark

Why the study?

Does a deep learning-based ASL MRI denoising method improve image quality and sensitivity for detecting hypoperfusion in Alzheimer's disease compared to non-DL methods?

Comparison

Deep learning-based ASL denoising transferred or fine-tuned vs existing non-DL method

Design

Retrospective study

Authors

LZLei Zhang DXDanfeng Xie YLYiran Li

Discussion

Member takes

Overview

Supports DL transfer learning for ASL denoising in AD; leaves open prospective validation before clinical adoption.

Structured PICO

Does a deep learning-based ASL MRI denoising method improve image quality and sensitivity for detecting hypoperfusion in Alzheimer's disease compared to non-DL methods?

Population

428 subjects (189 females) from three cohorts, including young healthy adults, normal controls (NC), and Alzheimer's disease (AD) patients.

Intervention

Deep learning-based ASL MRI denoising method (DLASL) trained on young healthy adults' PCASL data and transferred (DTF) or fine-tuned (DLASLFT) to patient data.

Comparator

Existing non-deep learning method (NonDL).

Outcome

Cerebral blood flow (CBF) image quality (contrast-to-noise ratio, radiologic score) and CBF map sensitivity for detecting hypoperfusion (peak t-value and suprathreshold cluster size).surrogate

Deep learning-based denoising models trained on healthy subjects can be successfully transferred to patient data to improve ASL MRI image quality and sensitivity for detecting Alzheimer's disease-related hypoperfusion.

Cite This Study

Zhang et al. (2021) studied this question.

synapsesocial.com/papers/6a7cccdfe458e52fd728a014 https://doi.org/10.1002/jmri.27984

View Full Paper

Ask AI

Bookmark

November 6, 2021Journal of Magnetic Resonance ImagingOpen Access

Improving Sensitivity of Arterial Spin Labeling Perfusion MRI in Alzheimer's Disease Using Transfer Learning of Deep Learning‐Based ASL Denoising

View Full Paper

Ask AI

Bookmark

Why the study?

Does a deep learning-based ASL MRI denoising method improve image quality and sensitivity for detecting hypoperfusion in Alzheimer's disease compared to non-DL methods?

Comparison

Deep learning-based ASL denoising transferred or fine-tuned vs existing non-DL method

Design

Retrospective study

Authors

LZLei Zhang DXDanfeng Xie YLYiran Li

Discussion

Member takes

Overview

Supports DL transfer learning for ASL denoising in AD; leaves open prospective validation before clinical adoption.

Structured PICO

Does a deep learning-based ASL MRI denoising method improve image quality and sensitivity for detecting hypoperfusion in Alzheimer's disease compared to non-DL methods?

Population

428 subjects (189 females) from three cohorts, including young healthy adults, normal controls (NC), and Alzheimer's disease (AD) patients.

Intervention

Deep learning-based ASL MRI denoising method (DLASL) trained on young healthy adults' PCASL data and transferred (DTF) or fine-tuned (DLASLFT) to patient data.

Comparator

Existing non-deep learning method (NonDL).

Outcome

Cerebral blood flow (CBF) image quality (contrast-to-noise ratio, radiologic score) and CBF map sensitivity for detecting hypoperfusion (peak t-value and suprathreshold cluster size).surrogate

Cite This Study

Zhang et al. (2021) studied this question.

synapsesocial.com/papers/6a7cccdfe458e52fd728a014 https://doi.org/10.1002/jmri.27984

View Full Paper

Ask AI

Bookmark