PT  - JOURNAL ARTICLE
AU  - Tracy Brown
AU  - Martin Lodge
AU  - Laura Dow
AU  - Corina Voicu
AU  - Frank Bengel
TI  - Timing of the pre-scan delay for static <sup>82</sup>Rb myocardial perfusion PET: Insights from dynamic/list-mode analysis using a LYSO PET-CT system
DP  - 2007 May 01
TA  - Journal of Nuclear Medicine
PG  - 223P--223P
VI  - 48
IP  - supplement 2
4099  - http://jnm.snmjournals.org/content/48/supplement_2/223P.2.short
4100  - http://jnm.snmjournals.org/content/48/supplement_2/223P.2.full
SO  - J Nucl Med2007 May 01; 48
AB  - 1064 Objectives: Timing of the delay between tracer injection and start of static/gated PET acquisition is challenging in 82Rb myocardial perfusion studies. Blood pool activity may limit image quality in case of a short delay, while count statistics decrease due to ultra-short tracer half-lives in long delays. We revisited this issue using a novel PET-CT system with fast LYSO crystals and list mode capability. Methods: Rest/dipyridamole 82Rb PET datasets were acquired in 2D (8 min list mode starting at time of infusion of 45-50 mCi of 82Rb) using a GE Discovery Rx scanner with LYSO PET and 64-slice CT. Four patients who had normal perfusion studies after being referred for exclusion of CAD were included in this analysis. List mode data were reformatted to static images using pre-scan delays of 60-, 75-, 90-, and 120-sec, and ROI analysis of myocardium, blood pool (BPLV), and lungs was performed. Dynamic imaging sequences of 32 frames were created and the time between start and arrival of activity in BPLV, and decrease of BPLV below myocardial activity was recorded from time-activity curves. Results: Heart/blood ratio was highest in 90-sec delay data sets for rest (3.16±2.30), and in 120-sec delay data sets for stress (4.23±1.16; p=0.08 vs. rest). Heart/lung ratio was highest in the 120-sec delay data sets for both rest and stress (7.49±0.76 and 10.94±2.87 at rest and stress, respectively). Noise, as assessed by the coefficient of variation of counts within the myocardium, was highest in the 120-sec delay data sets in both rest and stress images (CV=12.6% at rest and 18.9% at stress). Analysis of dynamic data revealed a stable delay of 25-30 sec between start of list mode acquisition and bolus arrival in the heart, but there was significant variability in the time until clear divergence of blood pool curve below myocardial curve (range 60-110 sec). Conclusions: While myocardium-to-background ratios increase with increasing time of pre-scan delay for static 82Rb cardiac images, intra-myocardial noise also increases. An often recommended 90-sec delay is feasible in novel LYSO systems, but the observed variation in tracer kinetics suggests that an algorithm for individual definition of delay based on list mode analysis of kinetics may be desirable.