PT - JOURNAL ARTICLE AU - Austin Chan AU - Jonathon Nye AU - John Votaw AU - Tracy Faber TI - Sensitivity of absolute flow calculations to variations in heart size, external activity, noise, and lesion location and severity: A simulation study DP - 2011 May 01 TA - Journal of Nuclear Medicine PG - 2079--2079 VI - 52 IP - supplement 1 4099 - http://jnm.snmjournals.org/content/52/supplement_1/2079.short 4100 - http://jnm.snmjournals.org/content/52/supplement_1/2079.full SO - J Nucl Med2011 May 01; 52 AB - 2079 Objectives Various physiological and physical factors affect the time activity curves (TACs) used to compute absolute blood flow in dynamic PET. This study used the realistic NCAT simulation to model these effects and evaluate their impact upon a flow measurement algorithm. Methods Dynamic NCAT simulations were created using known RV, LV, and myocardial tissue curves generated from known kinetic parameters. Heart sizes of 170, 110, and 75g were modeled. Lesions were placed in the lateral, septal, inferior, and apical walls. Liver and lung activities 10 and 20% of myocardium intensity were used. Defect severities were 40%, 60%, and 80% of maximum myocardial activity. Noise levels were estimated from patient studies, and normal, .5xnormal, and 2xnormal levels were simulated. Absolute flow was measured from the resulting 324 simulated 5min dynamic PET studies by automatically segmenting a summation of the last 2 minutes of the study into LV, RV, and 17 myocardial regions of interest (ROIs). Factor analysis was used to create pure tissue TACs for each of the 17 segments, and an optimization algorithm was used to find the best 3 compartmental model parameters for each. Segmental absolute flow values in the known normal and abnormal segments were then used in ROC analysis to determine whether any of these variables affected diagnostic ability. Results Neither heart size nor noise level affected accuracy of the flow algorithm. The least severe lesion was less detectable than the two more severe lesions. Higher liver activity, but not lung activity, was associated with lower accuracy. Finally, apical lesions were significantly less detectable than lesions in other locations; however, this may be associated with the apical lesion being approximately half the size of the other lesions rather than its specific location. Conclusions This absolute flow algorithm, like all perfusion quantification programs, is less accurate with smaller and less severe lesions. However, it is not affected by heart size or noise