Anatomically Refined Entorhinal Cortex Segmentation Improves MRI-Based Early Diagnosis of Alzheimer's Disease

Yongha Gi¹Sangyoon Park¹Hyungjin Lim¹Jeongwon Lee¹A Hyun Jung¹Seol-Hee Baek²Jong Hyun Kim³Byung-Jo Kim^2*Myonggeun Yoon^1*

• ¹Korea University, Seoul, Republic of Korea
• ²Korea University Medicine, Seongbuk-gu, Republic of Korea
• ³FieldCure Co., Seoul, Republic of Korea

The entorhinal cortex (EC) is one of the earliest cortical regions affected in Alzheimer's disease (AD) and serves as a key target for magnetic resonance imaging (MRI) biomarkers. However, conventional segmentation pipelines based on the Desikan–Killiany atlas do not clearly distinguish the EC from the adjacent perirhinal cortex, leading to mixed labels and reduced diagnostic sensitivity. To address this issue, we developed an anatomically refined EC segmentation framework that integrates expert-guided anatomical correction with deep learning. Specifically, FreeSurfer-derived EC labels were refined by selectively removing anterior perirhinal extensions and other anatomically inconsistent regions that are functionally distinct from the EC, and the resulting expert-corrected labels were used to train a two-stage no-new-Net (nnU-Net) model on Alzheimer's Disease Neuroimaging Initiative 1 (ADNI1) MRI data. This design preserves anatomically valid EC boundaries while enabling large-scale, consistent delineation across individuals and scanners. The refined segmentation showed stronger group-level differentiation among cognitively normal, mild cognitive impairment, and AD groups. When incorporated into volumetric and classification analyses, it provided more specific imaging biomarkers of early neurodegeneration and improved discrimination between diagnostic stages. External validation confirmed that the framework generalizes reliably across datasets. These findings demonstrate that anatomically precise and expert-informed EC delineation improves the sensitivity of MRI-based biomarkers for early AD diagnosis. The proposed framework offers a practical and reproducible approach for studying subtle cortical changes that precede overt clinical symptoms.