AbstractBackgroundAlzheimer’s disease (AD) is an age‐related neurodegenerative disease pathologically characterized by beta‐amyloid plaque accumulation and tau‐protein neurofibrillary tangles in the brain. Recent research has sought to identify additional potential biomarkers for earlier detection and treatment of the disease, including one such protein named cerebrospinal fluid neurofilament light protein (CSF NfL). Despite CSF NfL’s essential contributions to forming the neuronal cytoskeleton and supporting neuroaxonal function, the roles of CSF NfL in longitudinal AD progression and AD‐related neurodegeneration remain to be elucidated. This study examined the relationships between baseline CSF NfL levels and cortical gray matter atrophy along the spectrum of AD on a longitudinal scale over the span of two years.MethodWe included 179 study subjects (mean age = 74.21; 90 Females) with 60 cognitively normal (CN) individuals, 63 patients with late mild cognitive impairment (LMCI), and 56 AD patients from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Baseline CSF NfL measures and demographic data were downloaded from the ADNI study. Structural magnetic resonance imaging (MRI) scans from three time points (baseline, year 1, year 2) were processed and analyzed using FreeSurfer longitudinal analysis pipelines to compute percent changes in gray matter thickness for Year 1 (Y1) and Year 2 (Y2) compared with baseline, respectively.ResultBaseline CSF NfL was positively associated with greater cortical thinning across all study subjects, which were more pronounced in Y2. This relationship between baseline CSF NfL and cortical thinning remained the same in the AD diagnosis group. For the LMCI diagnosis group, there was no clear trend between baseline CSF NfL measures and percent changes in cortical thickness. Interestingly, a positive association between baseline CSF NfL value and percent changes in cortical thickness was observed in the CN diagnosis group at Y1, especially in the left hemisphere.ConclusionCSF NfL may contribute to the disease progression potentially through cortical atrophy and has an ability to predict cortical thinning over the short period of time. Longitudinal follow‐up analyses are warranted to incorporate its trajectory into the A/T/N classification system.