[¹⁸F]MK-6240 Radioligand Delivery Indices as Surrogates of Cerebral Perfusion: Bias and Correlation Against [¹⁵O]Water

Abstract

[¹⁸F]MK-6240 PET (where MK-6240 is 6-(fluoro)-3-(1H-pyrrolo[2,3-c]pyridin-1-yl)isoquinolin-5-amine) is used to assess in vivo tau deposition across the Alzheimer disease (AD) spectrum. We aimed to quantify the associations and bias of early-frame [¹⁸F]MK-6240 PET as surrogates for cerebral perfusion against gold standard [¹⁵O]water PET and the potential impact of cerebral perfusion on [¹⁸F]MK-6240 tau quantification across aging and the AD spectrum. Methods: Fourteen cognitively normal (CN, 4 young CN and 10 old CN) and 3 AD participants underwent dynamic [¹⁸F]MK-6240 PET, with 9 undergoing arterial sampling. A subset (n = 11) underwent [¹⁵O]water PET. [¹⁸F]MK-6240 perfusion indices were estimated as radiotracer delivery indices K₁ (using 2-tissue-compartment models), and relative perfusion indices were estimated as R₁ (using compartmental and reference tissue models, cerebellar gray matter reference region) and early-frame SUV ratio (0–3 min). [¹⁵O]water K₁ and R₁ were estimated using 1-tissue-compartment models). [¹⁸F]MK-6240 tau burden was estimated using distribution volume ratio and SUV ratio at 90–110 min. Spearman correlations, linear mixed-effect models, and Bland–Altman analyses examined relationships between [¹⁸F]MK-6240 perfusion indices against [¹⁵O]water and between estimates of perfusion and tau burden in tau-relevant regions. The impact of partial-volume correction was examined. Results: Significant correlations were observed between [¹⁸F]MK-6240 K₁ and [¹⁵O]water K₁ (ρ = 0.57); However, [¹⁸F]MK-6240 K₁ underestimated [¹⁵O]water K₁ by up to 50%, with a strong negative proportional bias. Significant correlations were observed between [¹⁸F]MK-6240 relative perfusion and [¹⁵O]water R₁ (ρ > 0.84), with minimal bias. In 2 AD participants, significant correlations were observed between perfusion and [¹⁸F]MK-6240 retention. Applying partial-volume correction did not significantly impact the correlations or improve the underestimations in [¹⁸F]MK-6240 K₁. Conclusion: Using head-to-head [¹⁸F]MK-6240 and [¹⁵O]water data, we showed that [¹⁸F]MK-6240 exhibited a relatively low extraction fraction, leading to underestimation of cerebral perfusion. Our results provide further support for [¹⁸F]MK-6240 R₁ as a reliable estimate of relative cerebral perfusion, with strong associations and minimal bias compared with [¹⁵O]water. In addition, lower perfusion may be associated with higher [¹⁸F]MK-6240 retention in tau-relevant regions in AD. These findings further support the use of dynamic [¹⁸F]MK-6240 in dual-imaging assessments of tau burden and vascular health.