An edge-aware salient context fusion and refinement network for hippocampal segmentation in MR images and its diagnostic value for mild cognitive impairment - Report - MDSpire

A Context Fusion and Refinement Network with Edge Awareness for Hippocampal Segmentation in MRI and Its Diagnostic Importance for Mild Cognitive Impairment

  • By

  • Limin Liu

  • Xiaolong Chen

  • Qiqun Zeng

  • Shili Zhou

  • Xia Zhang

  • July 17, 2026

Share

Clinical Report: Context Fusion and Refinement Network for Hippocampal Segmentation

Overview

The ESCFR-Net demonstrates significant improvements in hippocampal segmentation accuracy, achieving a Dice coefficient of 0.9004. Clinical analysis indicates that hippocampal volumes are significantly smaller in patients with mild cognitive impairment compared to healthy controls.

Background

Accurate hippocampal volume assessment is crucial for early diagnosis and monitoring of Alzheimer's disease (AD). The challenges of manual segmentation, including time consumption and variability, are noted.

Data Highlights

MetricValue
Dice Coefficient0.9004
AUC for distinguishing HCs from MCI0.927
Sensitivity90.11%
Specificity83.52%

Key Findings

  • ESCFR-Net outperforms state-of-the-art methods in hippocampal segmentation.
  • Bilateral hippocampal volumes in MCI patients are significantly smaller than in healthy controls (p < 0.001).
  • Total hippocampal volume has an AUC of 0.927 for distinguishing MCI from HCs.
  • Sensitivity and specificity of the total hippocampal volume in distinguishing MCI are 90.11% and 83.52%, respectively.
  • The proposed network utilizes multiple advanced modules to enhance segmentation accuracy.

Clinical Implications

The ESCFR-Net provides a tool for automated hippocampal segmentation.

Conclusion

The study presents advancements in automated hippocampal segmentation.

Related Resources & Content

  1. npj Digital Medicine, 2025 -- An Efficient CVTC Framework for Precise MRI Evaluation and Lesion Marking in Alzheimer’s Disease
  2. Frontiers in Neurology, 2026 -- HDFT-MViT: A Progressive Core-Enhanced Mix Framework for Alzheimer's Disease Classification using MRI images
  3. Frontiers in Medicine, 2026 -- An intelligent gradient-guided hybrid inpainting framework for brain MRI reconstruction and Alzheimer's disease classification in connected healthcare systems
  4. Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup - PubMed, 2025
  5. Frontiers in Neurology — Disrupted basal forebrain-cortical connectivity in amnestic mild cognitive impairment: unveiling circuit-level links to cognitive decline
  6. Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup - PubMed
  7. Dementia and Movement Disorders
  8. FDA approves treatment for adults with Alzheimer’s disease | FDA

Original Source(s)

Related Content