The increasing availability of Artificial Intelligence (AI) would benefit from larger imaging databases to allow for algorithm development. AI training requires similar images for input for a more precise delineation of pathology. In the real-world clinical setting, scanner and MRI sequence protocol variations present a major challenge for developing AI algorithms due to the non-uniformity of images. Therefore, harmonization and standardization of MRI scans must be implemented as a post-acquisition step for machine learning, predictive modeling, and outcome research.
A validated imaging biomarkers will allow for a more accurate diagnosis, disease activity assessment, prognosis, and treatment response. Standardization of novel biomarkers and their application to both cross-sectional and longitudinally dataset would lead to improved quantification and assessment of pathophysiological changes. After initial validation steps are achieved, their application to a large curated test database of images for measuring disease activity and progression will lead to reducing the “turnaround” time for drug development and cost burden to the healthcare system.
Validation of novel imaging biomarkers must be compared to widely accepted reference standards, such as tissue histopathology or previously substantiated markers, in appropriate disease model (clinical or basic science). Validation studies must establish accuracy, reliability and reproducibility of the proposed imaging metric (volumetric, tissue susceptibility, perfusion, diffusion, magnetization, relaxometry, etc.) for physicians to have dependability in the results as it is relevant to clinical decision making process.
There is an unmet need in neurology to develop improved techniques for examining pathology in an accurate and high throughput means, vis-à-vis machine learning. We hope that this special issue of Frontiers in Neurology will serve to bridge the gap between those who have expertise in basic MRI technology and those clinicians searching for better MRI biomarkers for optimal disease assessment, management, and outcome.
The scope of this topic is broad. We seek original research, reviews, mini reviews and editorial manuscripts to undergo peer review. Of particular interest is the application of harmonization to large, curated imaging archives and A.I. application in human disease. The subject matter of appropriate topics includes a wide range of subject matter papers from technical development and/or clinical application of:
-Image harmonization,
-Quantification,
-Validation,
in the management of a wide range of neurologic diseases.
The increasing availability of Artificial Intelligence (AI) would benefit from larger imaging databases to allow for algorithm development. AI training requires similar images for input for a more precise delineation of pathology. In the real-world clinical setting, scanner and MRI sequence protocol variations present a major challenge for developing AI algorithms due to the non-uniformity of images. Therefore, harmonization and standardization of MRI scans must be implemented as a post-acquisition step for machine learning, predictive modeling, and outcome research.
A validated imaging biomarkers will allow for a more accurate diagnosis, disease activity assessment, prognosis, and treatment response. Standardization of novel biomarkers and their application to both cross-sectional and longitudinally dataset would lead to improved quantification and assessment of pathophysiological changes. After initial validation steps are achieved, their application to a large curated test database of images for measuring disease activity and progression will lead to reducing the “turnaround” time for drug development and cost burden to the healthcare system.
Validation of novel imaging biomarkers must be compared to widely accepted reference standards, such as tissue histopathology or previously substantiated markers, in appropriate disease model (clinical or basic science). Validation studies must establish accuracy, reliability and reproducibility of the proposed imaging metric (volumetric, tissue susceptibility, perfusion, diffusion, magnetization, relaxometry, etc.) for physicians to have dependability in the results as it is relevant to clinical decision making process.
There is an unmet need in neurology to develop improved techniques for examining pathology in an accurate and high throughput means, vis-à-vis machine learning. We hope that this special issue of Frontiers in Neurology will serve to bridge the gap between those who have expertise in basic MRI technology and those clinicians searching for better MRI biomarkers for optimal disease assessment, management, and outcome.
The scope of this topic is broad. We seek original research, reviews, mini reviews and editorial manuscripts to undergo peer review. Of particular interest is the application of harmonization to large, curated imaging archives and A.I. application in human disease. The subject matter of appropriate topics includes a wide range of subject matter papers from technical development and/or clinical application of:
-Image harmonization,
-Quantification,
-Validation,
in the management of a wide range of neurologic diseases.