Segmentation errors and intertest reliability in automated and manually traced hippocampal volumes

Area of Science:

• Neuroimaging
• Epilepsy Research
• Medical Image Analysis

Background:

• Accurate hippocampal volumetry is crucial for diagnosing and monitoring epilepsy.
• Manual tracing is time-consuming and prone to inter-rater variability.
• Automated segmentation offers potential for increased efficiency and consistency.

Purpose of the Study:

• To compare the accuracy, repeatability, and clinical acceptability of automated (NeuroQuant, FreeSurfer) versus manual hippocampal segmentation.
• To evaluate performance across a range of imaging abnormalities in epilepsy patients.

Main Methods:

• Manual and automated (NeuroQuant, FreeSurfer) hippocampal segmentation were performed on 3T MPRAGE MRI scans from 49 epilepsy patients.
• Reproducibility was assessed through repeated manual tracings and automated measurements.
• Performance was evaluated in cases with anatomical variations and segmentation errors.

Main Results:

• Automated methods (NeuroQuant, FreeSurfer) demonstrated superior volume reproducibility compared to manual tracing.
• Both automated methods showed lower asymmetry index reproducibility than manual tracing.
• Automated segmentation failed in cases with significant anatomical deformations and produced clinically significant errors in a subset of patients.

Conclusions:

• Automated hippocampal segmentation offers improved reproducibility over manual tracing in epilepsy.
• Clinical application requires careful validation due to potential failures and significant segmentation errors in complex cases.