Skip to Content

Validation of Automated Hippocampal Volumetrics Using Neuroreader® ​

Ahdidan, Raji, DeYoe, Mathis, Noe, Rimestad, Kjeldsen, Mosegaard, Becker, Lopez
Journal of Alzheimer’s Disease, vol. 49, no. 3, pp. 723-732
October 15, 2015


Overview

Accurately measuring hippocampal volume is critical for diagnosing and tracking multiple neurological disorders—including Alzheimer’s disease (AD), mesial temporal sclerosis, depression, schizophrenia, PTSD, and traumatic brain injury. Historically, the gold standard has been manual segmentation, a labor-intensive process that is impractical for clinical use.

This peer-reviewed study validates the performance of Neuroreader®, an FDA-cleared volumetric MRI software, demonstrating its strong accuracy, speed, and clinical utility in automated hippocampal segmentation.


Purpose of the Study

To evaluate how accurately Neuroreader® measures hippocampal volume compared with expert manual segmentation using the ADNI (Alzheimer’s Disease Neuroimaging Initiative) dataset.


Study Population & Methods

  • 99 subjects from the ADNI initiative
  • MRI scans from both 1.5T (n=59) and 3T (n=40) scanners
  • Neuroreader® performed fully automated hippocampal segmentation using a multi-atlas approach
  • Accuracy was evaluated using the Dice Similarity Coefficient (DSC)
  • Manual segmentation performed by ADNI “master tracers” served as the gold standard


Key Findings

High Accuracy Compared to Manual Segmentation

  • Mean DSC ≈ 0.87 for both left and right hippocampus
  • Maximum DSC reached 0.91, indicating near-perfect agreement
  • Performance consistent across 1.5T and 3T scanners
  • Neuroreader® segmentation accuracy remained strong across:
    • Normal controls
    • Mild cognitive impairment (MCI)
    • Alzheimer’s disease (AD) patients

Clinical Significance

  • Neuroreader® enables sensitive detection of subtle hippocampal atrophy, particularly important for:
    • Early Alzheimer’s disease
    • Mild traumatic brain injury
    • PTSD
    • Depression
    • Temporal lobe epilepsy
  • Automated analysis takes 3–7 minutes, vs. ~30 minutes per scan for manual segmentation
  • Enables routine clinical use, improving workflow efficiency


Why It Matters

Hippocampal atrophy is one of the earliest imaging markers of Alzheimer’s disease and contributes to the diagnostic picture in multiple neuropsychiatric conditions. Traditional visual assessment frequently fails to detect early changes—sometimes with sensitivity as low as 27%.

This study demonstrates that Neuroreader® provides:

  • Objective, reproducible volumetrics
  • High agreement with the gold standard
  • Scalable use in real clinical environments
  • Reliable performance even on older (1.5T) MRI systems

The findings support Neuroreader® as a viable tool for advancing precision imaging in neurodegenerative and neuropsychiatric diagnostics.

Read the Full ​​​​Article