Predicting amyloid PET positivity using plasma p‐tau181 and other blood‐based biomarkers
Hyuk Sung Kwon, Eun Hye Lee, Hyung‐Ji Kim, So‐Hee Park, Hyun‐Hee Park, Jee Hyang Jeong, Seong‐Ho Koh, Seong Hye Choi, Jae‐Hong Lee
Abstract
Abstract Introduction This study aimed to determine the efficacy of combining plasma phosphorylated tau (p‐tau)181, amyloid beta (Aβ)42/Aβ40, neurofilament light (NfL), and apolipoprotein E ( APOE ) genotypes for detecting positive amyloid positron emission tomography (PET), which is little known in the Asian population, in two independent cohorts. Methods Biomarkers were measured using a single‐molecule array (Simoa) in a cohort study (Asan). All participants underwent amyloid PET. Significant changes in the area under the curve (AUC) and Akaike Information Criterion values were considered to determine the best model. The generalizability of this model was tested using another cohort (KBASE‐V). Results In the Asan cohort, after adjusting for age and sex, p‐tau181 (AUC = 0.854) or APOE ε4 status (AUC = 0.769) distinguished Aβ status with high accuracy. Combining them or adding NfL and Aβ42/40 improved model fitness. The best‐fit model included the plasma p‐tau181, APOE ε4, NfL and Aβ42/40. The models established from the Asan cohort were tested in the KBASE‐V cohort. Additionally, in the KBASE‐V cohort, these three biomarker models had similar AUC in cognitively unimpaired (AUC = 0.768) and mild cognitive impairment (MCI) (AUC = 0.997) participants. Conclusions Plasma p‐tau181 showed a high performance in determining Aβ‐PET positivity. Adding plasma NfL and APOE ε4 status improved the model fit without significant improvement in AUC.