Objective: Analyze the differences in the detection results of Alzheimer’s Disease (AD) serum biomarkers and the expression of APOE and SLCO1B1 genotypes, then construct an AD diagnostic model integrating serum biomarkers with APOE and SLCO1B1 genotypes to enhance the sensitivity and specificity of AD screening.

Methods: Serum samples from 127 participants (41 AD, 39 cognitive impairments, 47 controls) were analyzed for Neurofilament light chain (NF), Aβ42/Aβ40 ratio, p-TAU231, and p-TAU217 via ELISA, combined with APOE/SLCO1B1 genotyping. A nomogram model was constructed using logistic regression and validated by ROC-AUC and Hosmer-Lemeshow tests.

Results: The AD group showed elevated NF (p < 0.01) and p-TAU231 (p < 0.05), reduced Aβ42/Aβ40 (p < 0.05). APOE ∑4 homozygotes exhibited the highest p-TAU181 (p < 0.001), while SLCO1B1*1b/*1b correlated with increased Aβ42 (p < 0.05). The biomarker model achieved AUC = 0.8365, improving to 0.8832 after genetic integration (calibration p > 0.05). Conclusion: This study pioneers an AD diagnostic model combining serum p-TAU231, NF, and SLCO1B1 polymorphisms, surpassing single-marker limitations (AUC > 0.88). It provides a high-accuracy tool for non-invasive screening and genetic stratification, demonstrating substantial clinical potential.