ABSTRACT
Objective To investigate the brain age differences between Alzheimer's disease(AD)and mild cognitive impairment(MCI)patients,and further explore the correlations between brain age gap(BAG)and clinical features.Methods The clinical data and radiologic findings of 132 probable AD and AD-derived MCI patients diagnosed at Beijing Tiantan Hospital,Capital Medical University from December 2018 to July 2021 were retrospectively analyzed.According to the diagnostic criteria for AD and MCI,the patients were assigned into AD and MCI groups.In addition,156 volunteers without neurological diseases and other severe diseases were recruited as the control group.The general data,Montreal cognitive assessment(MoCA)score,and mini-mental state examination(MMSE)score were compared among the three groups.The deep learning-based brain age prediction model was employed to calculate the BAGs of the three groups.Spearman correlation analysis was conducted to explore the correlations between BAG and clinical features.Results The 132 patients included 106 patients in the AD group and 26 patients in the MCI group.The MoCA and MMSE scores followed an ascending trend of AD group<MCI group<control group(all P<0.001).The predicted brain age and BAG in the AD group were higher than those in the MCI group(P=0.040,P=0.003)and control group(P=0.001,P<0.001).There was no significant difference in predicted brain age or BAG between MCI and control groups(P=0.352,P=0.224).BAG was negatively correlated with MoCA score(r=-0.341,P<0.001)and MMSE score(r=-0.324,P=0.001)in the AD group.Conclusion BAG can be used as an imaging biomarker to evaluate the degree of brain structural variation and the severity of brain injury in the patients with cognitive impairment.