Related Experiment Video
Updated: Jul 25, 2025
Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
Published on: December 15, 2023
Excellent Interrater Reliability for Manual Segmentation of the Medial Perirhinal Cortex
Nicolas A Henzen1,2, Julia Reinhardt3,4,5, Maria Blatow6
1University Department of Geriatric Medicine FELIX PLATTER, 4055 Basel, Switzerland.
Abstract:
Objective: Evaluation of interrater reliability for manual segmentation of brain structures that are affected first by neurofibrillary tau pathology in Alzheimer's disease. Method: Medial perirhinal cortex, lateral perirhinal cortex, and entorhinal cortex were manually segmented by two raters on structural magnetic resonance images of 44 adults (20 men; mean age = 69.2 ± 10.4 years). Intraclass correlation coefficients (ICC) of cortical thickness and volumes were calculated. Results: Very high ICC values of manual segmentation for the cortical thickness of all regions (0.953-0.986) and consistently lower ICC values for volume estimates of the medial and lateral perirhinal cortex (0.705-0.874). Conclusions: The applied manual segmentation protocol allows different raters to achieve remarkably similar cortical thickness estimates for regions of the parahippocampal gyrus. In addition, the results suggest a preference for cortical thickness over volume as a reliable measure of atrophy, especially for regions affected by collateral sulcus variability (i.e., medial and lateral perirhinal cortex). The results provide a basis for future automated segmentation and collection of normative data.
Insights
This study found that measuring cortical thickness is more reliable than volume for assessing brain changes in Alzheimer's disease. Manual segmentation of specific brain regions showed high consistency for thickness measurements.
Area of Science:
- Neuroimaging
- Neurology
- Alzheimer's Disease Research
Background:
- Alzheimer's disease (AD) is characterized by neurofibrillary tau pathology affecting specific brain regions.
- Accurate measurement of brain structure changes is crucial for understanding AD progression and developing treatments.
- Interrater reliability in manual segmentation is essential for consistent and reproducible neuroimaging studies.
Purpose of the Study:
- To evaluate the interrater reliability of manual segmentation for key brain structures in Alzheimer's disease.
- To compare the reliability of cortical thickness versus volumetric measurements for these regions.
- To establish a reliable protocol for future automated segmentation and normative data collection.
Main Methods:
- Manual segmentation of medial perirhinal cortex, lateral perirhinal cortex, and entorhinal cortex by two independent raters.
- Structural magnetic resonance imaging (MRI) data from 44 adult participants were used.
- Intraclass correlation coefficients (ICC) were calculated for cortical thickness and volume estimates.
Main Results:
- Very high ICC values (0.953-0.986) were observed for cortical thickness measurements across all segmented regions.
- Consistently lower ICC values (0.705-0.874) were found for volumetric estimates, particularly for medial and lateral perirhinal cortex.
- The manual segmentation protocol demonstrated high reliability for cortical thickness estimation.
Conclusions:
- Manual segmentation protocols can yield highly consistent cortical thickness estimates for parahippocampal gyrus regions.
- Cortical thickness is a more reliable measure of atrophy than volume, especially in regions with anatomical variability like the medial and lateral perirhinal cortex.
- These findings support the use of cortical thickness for future automated segmentation and the development of normative datasets in AD research.
