ScholarWorks@성균관대학교

초록

BACKGROUND: Accurately detecting amyloid-β (Aβ) accumulation is critical for identifying patients with early stage Alzheimer's disease (AD) who may benefit from Aβ-targeted therapies. While Aβ positron emission tomography (PET) and cerebrospinal fluid (CSF) biomarkers are reliable, their high cost and invasiveness limit widespread use. This study evaluated a two-stage diagnostic workflow integrating magnetic resonance imaging (MRI)-based brain atrophy assessments and plasma p -Tau217 testing for predicting Aβ PET positivity in patients with early-stage AD. METHOD: This prospective cohort study was conducted in a tertiary hospital setting, including participants diagnosed with mild cognitive impairment (MCI) or early Alzheimer-type dementia (ATD) from the K-ROAD cohort and ADNI cohort. A random forest classifier was developed to predict Aβ PET positivity using MRI-derived brain atrophy patterns and APOE ε4 status. Participants were stratified into low-, intermediate-, and high-risk groups for Aβ PET positivity based on the prediction from this classifier. Additional plasma p -Tau217 testing was then applied exclusively to individuals in the intermediate-risk group. Diagnostic performance was assessed using positive predictive value (PPV), and negative predictive value (NPV). RESULT: The study included 807 participants form K-ROAD (median age: 72.0 years, 58.7% female) and 230 from ADNI (median age: 70.9 years, 49.1% female). In the first stage, using the 95% sensitivity/specificity strategy, the NPV in the low-risk group was 94.7% (91.7-97.7%) in K-ROAD and 99.0% (97.0-100.0%) in ADNI. The PPV in the high-risk group was 97.6% (95.9-99.3%) in K-ROAD and 98.8% (96.5-100.0%) in ADNI. The intermediate-risk group comprised 33.3% in K-ROAD and 20.9% in ADNI. In the second stage, plasma p -Tau217 testing for intermediate-risk group achieved a PPV of 92.5% (88.7-96.3%) and an NPV of 83.1% (75.0-91.2%) in K-ROAD. In ADNI, the PPV was 90.0% (79.0-100.0%) and the NPV was 83.3% (66.1-100.0%). The overall accuracy of the workflow was 94.2% (92.6-95.8%) in K-ROAD and 96.5% (94.1-98.9) in ADNI. CONCLUSION: This study demonstrates that the two-stage diagnostic workflow improves diagnostic accuracy for predicting Aβ PET positivity in early-stage AD. Despite some limitations, this approach provides a cost-effective and scalable strategy for early detection and optimized healthcare resource allocation.