Manual segmentation qualification platform for the EADC-ADNI harmonized protocol for hippocampal segmentation project

Area of Science:

• Neuroimaging
• Biomarker Development
• Alzheimer's Disease Research

Background:

• Hippocampal volumetry is crucial for Alzheimer's disease (AD) biomarker development.
• Standardized segmentation methods are required for reliable hippocampal volume measurements across different laboratories.
• The European Alzheimer's Disease Consortium and Alzheimer's Disease Neuroimaging Initiative (EADC-ADNI) Harmonized Protocol (HarP) provides a standardized method.

Purpose of the Study:

• To present a platform for training and qualifying tracers for manual hippocampal segmentation.
• To demonstrate that the platform's training process enhances compliance with the HarP.
• To establish a standardized method for evaluating tracer performance in hippocampal volumetry.

Main Methods:

• Thirteen tracers performed manual hippocampal segmentation on MRI scans.
• Segmentations were evaluated against benchmark images using volume, Jaccard similarity index, and spatial distance.
• Three training phases with visual feedback were followed by a qualification phase on new MRI scans.

Main Results:

• Statistical analysis showed a significant increase in Jaccard overlap (P < .001) between training phases for all tracers.
• The training process demonstrably improved tracer compliance with the HarP.
• Defined qualification thresholds were successfully met by all tracers.

Conclusions:

• The developed platform provides a robust infrastructure for standardized training and evaluation of tracers.
• This standardization is essential for the reliable use of hippocampal volumetry as an AD biomarker.
• The platform supports the integration of hippocampal volumetry in clinical and research settings for AD diagnosis and monitoring.