Shaofan Jiang Siyu Yang Kaiji Deng Rifeng Jiang Yunjing Xue ^*

• Fujian Medical University Union Hospital, Fuzhou, China

Objective: This study aimed to develop and validate machine learning models (MLMs) to diagnose Alzheimer's disease (AD) using cortical complexity indicated by fractal dimension (FD).: 296 participants with normal cognitive function (NC) and 182 with AD from the AD Neuroimaging Initiative database were randomly divided into training and internal validation cohorts. Then, FDs, demographic characteristics, baseline global cognitive function scales (MOCA, FAQ, GDS, NPI), phospho-tau (P-tau 181), Amyloidβ-42/40, Apolipoprotein E (APOE) and polygenic hazard score (PHS) were collected and established multiple MLMs. Receiver operating characteristic curves were used to evaluate model performance. Participants from our institution (n = 66, 33 with NC and 33 with AD) served as external validation cohorts to validate the MLMs models. Decision curve analysis was used to estimate the models' clinical values.Results: The FDs from 30 out of 69 regions showed significant alteration. All MLMs models were conducted based on the 30 significantly different FDs. FD model had good accuracy in predicting AD in three cohorts (AUC=0.842, 0.808, 0.803). There were no statistically significant differences in AUC values between the FD model and the other combined models in the training and internal validation cohorts except MOCA+FD and FAQ+FD models. AmongMLMs, the MOCA+FD model showed the best predictive efficiency in three cohorts (AUC=0.951, 0.931, 0.955) and had the highest clinical net benefit.The FD model showed favorable diagnostic performance for AD. Among MLMs, the MOCA+FD model can predict AD with the highest efficiency and could be used as a noninvasive method to diagnose AD.