Early prediction of Alzheimer’s disease using artificial intelligence and cortical features on T1WI sequences

• DOI: 10.3389/fneur.2025.1552940
• Publication Date: 2025-03-12T13:35:51Z
• Authors: Rong Zeng; Beisheng Yang; Faqi Wu; Huan Liu; Xiaojia Wu; Lin Tang; Rao Song; Qingqing Zheng; Xia Wang; Dajing Guo
• Keywords: Brain morphometry; Concordance

Background Accurately predicting the progression of mild cognitive impairment (MCI) to Alzheimer’s disease (AD) is a challenging task, which crucial for helping develop personalized treatment plans improve prognosis. Purpose To new technology early prediction AD using artificial intelligence and cortical features on MRI. Methods A total 162 MCI patients were included from Disease Neuroimaging Initiative (ADNI) database. By 3D-MPRAGE sequence, T1W images each patient acquired. All randomly divided into training set ( n = 112) validation 50) at ratio 7:3. Morphological cerebral cortex extracted with FreeSurfer software. Network gray matter GRETNA toolbox. The network, morphology, network-clinical, morphology-clinical, morphology-network morphology-network-clinical models developed by multivariate Cox proportional hazard model. performance model was assessed concordance index (C-index). Results In group, C-indexes 0.834, 0.926, 0.915, 0.949, 0.928, 0.951, respectively. those in group 0.765, 0.784, 0.849, 0.877, 0.884, 0.880, performed best. multi-predictor nomogram high accuracy individual (C-index 0.951) established. Conclusion occurrence could be accurately predicted our nomogram. This help doctors make decisions clinical practice, showed important significance.