Abstract
655
Objectives: 18F-Flurpiridaz (FPZ) is a novel radiotracer for myocardial perfusion PET currently undergoing late stage clinical trials [1-3]. We have recently demonstrated the added incremental diagnostic value of stress myocardial blood flow (MBF) and reserve (MFR) using data from the first phase III trial of FPZ (301 trial, NCT01347710) with a 2-tissue-compartment (2TC) model developed in animal studies [4]. Since the appropriate FPZ kinetic model in humans has not yet been determined, we sought to evaluate the simpler 1-tissue-compartment (1TC) model for FPZ PET and its diagnostic accuracy compared to invasive coronary angiography (ICA).
Methods: We analyzed available dynamic PET data from the 301 trial, which were acquired using a standardized protocol at rest and during pharmacologic stress (15x10s, 5x30s, 5x60s, 1x300s). Residual activity in the second PET scan was corrected by simple subtraction. Frame-by-frame alignment was visually checked and misalignment due to patient motion was manually corrected. The arterial input function was derived from a region of interest centered at the mitral valve plane and used without correction. MBF was estimated with 1TC and 2TC models assuming no adjustment for flow-dependent extraction (MBF=K1), and irreversible binding for the 2TC model (k4=0). A geometric model for partial volume correction was assumed for both TC models. Left ventricular MBF and MFR were assessed globally and per vascular territory. Vessels with evidence of infarction on rest images were excluded (N=37). The incremental diagnostic value of kinetic parameter estimates was evaluated with multivariable logistic regression (stenosis threshold 50%) and continuous net reclassification improvement (NRI).
Results: Dynamic PET data were evaluable in 656 vessels of 231 patients. Median kinetic parameters per vessel for patients without obstructive CAD (N=131) are shown in Table 1. Kinetic parameters at stress and rest tended to be higher in women than men, and higher for 2TC compared with 1TC models, although 1TC MFR tended to be slightly higher than 2TC MFR. Stress MBF and MFR from both 1TC and 2TC models were significant independent predictors of obstructive CAD after adjustment for patient age, sex, BMI, pretest likelihood of CAD, and stress total perfusion deficit (TPD). Overall NRI and likelihood ratio χ2 were similar for 1TC and 2TC models (Table 2) but tended to be slightly higher for 2TC models for Stress MBF, and 1TC models for MFR. All other kinetic parameters (rest MBF, k2, and k3; stress k2 and k3) were not significant predictors of obstructive CAD.
Conclusions: Although kinetic parameter estimates from the 1TC model tended to be biased downward compared to the 2TC model, stress MBF and MFR for both models provided similar incremental diagnostic performance beyond patient characteristics and relative perfusion analysis. Further FPZ kinetic model validation compared with gold standard flow measurements in humans is warranted.
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.