Search

We propose and test a novel approach for direct parametric image reconstruction of dynamic PET data. We present a theoretical description of the problem of PET direct parametric maps estimation as an inference problem, from a probabilistic point of view, and we derive a simple iterative algorithm, based on the Iterated Conditional Mode (ICM) framework, which exploits the simplicity of a two-step optimization and the efficiency of an analytic method for estimating kinetic parameters from a nonlinear compartmental model. The resulting method is general enough to be flexible to an arbitrary choice of the kinetic model, and unlike many other solutions, it is capable to deal with nonlinear compartmental models without the need for linearization. We tested its performance on a two-tissue compartment model, including an analytical solution to the kinetic parameters evaluation, based on an auxiliary parameter set, with the aim of reducing computation errors and approximations. The new method is tested on simulated and clinical data. Simulation analysis led to the conclusion that the proposed algorithm gives a good estimation of the kinetic parameters in any noise condition. Furthermore, the application of the proposed method to clinical data gave promising results for further studies.

Aim The aim of this study was to evaluate the possible impact of stress-induced left ventricular (LV) diastolic dysfunction at cadmium–zinc–telluride (CZT) imaging, on the detection of significant coronary artery disease (CAD). Methods and results Four hundred and twenty-five consecutive patients underwent myocardial perfusion imaging at rest and after stress with a low-dose CZT protocol and the evaluation of coronary anatomy by invasive or computed coronary angiography. The summed difference score (SDS) was calculated in every patient. Left ventricular ejection fraction and peak filling rate (PFR) at baseline and after stress were derived from gated CZT images and the ‘% stress-to-rest’ PFR difference, as an indicator of stress-induced diastolic dysfunction, determined. In the study population, the mean SDS was 5 ± 4, while mean stress PFR and rest PFR were 2.5 ± 0.8 end-diastolic volumes (EDV)/s and 2.5 ± 0.7 EDV/s, respectively. There was a strict correlation between the presence and extent of CAD and both myocardial SDS and ‘% stress-to-rest’ PFR ( P < 0.001 for both). Interestingly, while myocardial SDS and ‘% stress-to-rest’ PFR were significantly correlated ( P < 0.001), they resulted independent predictors of the presence of significant CAD ( P < 0.001 and P < 0.032, respectively). Of note, at receiving operating characteristic analysis, a ‘% stress-to-rest’ PFR ≤3 showed 71% sensitivity in unmasking the presence of significant coronary luminal narrowings. Conclusion The present study shows that the assessment of stress-induced diastolic dysfunction with an ultrafast scintigraphic protocol can improve the accuracy in detection of significant ischaemic heart disease.