The confounding effects of microvascular physiology on the uptake and diagnostic accuracy of [¹⁸F]fluoroethyl-tyrosine positron emission tomography in gliomas.

Abstract

To investigate the physiological influence of tissue perfusion and permeability on O-(2-[¹⁸F]-fluoroethyl)-L-tyrosine ([¹⁸F]FET) uptake in gliomas by simultaneous MRI dynamic contrast enhanced (DCE) perfusion imaging and the impact of vascular contributions on diagnostic accuracy. Dynamic [¹⁸F]FET PET/MRI scans from 50 patients with glioma WHO grade 2-4 were analysed. In each patient multiple subvolumes were delineated to account for intra-patient heterogeneity. Associations of 20-40 min. [¹⁸F]FET tumour-to-background ratio (TBR_med) with median blood volume (CBV), blood flow (F), and permeability (Ki) within 134 lesion subvolumes were investigated in linear mixed models. Also, associations with initial area under curve (iAUC₁₂₀), time to peak (TTP) and slope (Slope_20 - 40) from time activity curve were investigated. The influence of adjusting TBR_med for microvascular physiological parameters on the diagnostic accuracy for tumour recurrences was assessed in an independent dataset (n = 61 lesions from 48 patients). CBV, F and Ki individually accounted for 42%, 36% and 26%, respectively, of the overall variance in TBR_med. DCE metrics combined accounted for 49% of the total variance and for 65% of the regional variance in TBR_med. All DCE metrics were associated positively with iAUC₁₂₀ and negatively with both Slope_20 - 40 and TTP (p < 0.001 for all). Adjusting TBR_med for microvascular effects lowered TBR by 24% on average and reduced the diagnostic accuracy (ROC AUC) from 0.90 to 0.77 in the test dataset. Microvascular properties may contribute to a considerable fraction of the clinical [¹⁸F]FET measures in patients with gliomas, and contribute positively to diagnostic performance of [¹⁸F]FET PET imaging.