Optimizing Lumipulse plasma pTau217/Aβ42 ratio cutpoints in an Asian cohort with high cerebrovascular disease burden

The generalizability of FDA-cleared Lumipulse plasma pTau217 / Aβ42 ratio cutpoints to Asian populations remains uncertain, especially in those with high burden of concomitant cerebrovascular disease (CeVD). We evaluated pTau217, Aβ42 and their ratio in 205 Singapore-based participants with corresponding amyloid neuroimaging. In-house cutpoints were derived using a three-range strategy targeting 95% sensitivity and specificity. Plasma pTau217/Aβ42 ratio outperformed pTau217 alone (area under the curve, 0.951 versus 0.925; p=0.004), with comparable performance across CeVD severities. Plasma ratio cutpoints achieved 98% negative predictive value, 88% positive predictive value, and 95% accuracy, classifying 10% as intermediate-risk. Using FDA-cleared cutpoints for our cohort retained 97% sensitivity and negative predictive value but yielded lower specificity (86%), positive predictive value (77%) and accuracy (85%), with 29% classified as intermediate-risk. Our findings support the need for population-specific validation of cutpoints from FDA-cleared blood biomarkers of brain amyloid pathology.