Investigation of the feasibility of dynamic cardiac imaging with a dual detector gamma camera (DDGC)

Objectives Dynamic changes in tracer uptake in the whole heart may be useful as early predictors of cardiovascular disease or heart failure. The goal of this study was to demonstrate the feasibility of measuring changes in volumes-of-interest (VOI) activity using a DDGC. Previously we have developed the QPlanar method for estimating the activity organs or other VOI from planar projections. The method uses maximum-likelihood estimation techniques, and accurate models of the image formation to estimate the total activity in a set of 3D VOIs. In this study we adapted this methodology to extract dynamic information, in the form of the time activity curve for the myocardium and blood pool, from a conventional SPECT acquisition. In this study we used simulations to investigate the feasibility of estimating the time-activity curve (TAC) for the myocardium and blood pool.

Methods In this study we modeled uptake in the heart based on literature[1] TAC of Tc-99m teboroxime. We modeled patient anatomy using the 3D XCAT phantom. Simulated data were generated using an analytic simulator that models attenuation, scatter and the collimator-detector response. We modeled acquisition using a two-camera system in a right angle configuration. We modeled a 12 minute acquisition with data acquired at 120 views over 360° using continuous rotation. We studied the effects of noise by estimating TACs for 30 noise realizations. To study the effects of nonuniform myocardial uptake, we modeled a 100% severity perfusion defect in 38% of the myocardium.

Results TACs obtained of the heart and blood pool had an average (avg.) bias of 21% and 18% and avg. standard deviation of 26% and 25%, respectively. With the defect present the method underestimated the cardiac TAC by an avg. of 45.13%. We expect bias to be less severe with defect present in less than 25% of myocardium along with less than 100% severity.

Conclusions The proposed method allows relatively simply extraction of dynamic information about the myocardium and blood pool with a conventional slow acquisition protocol