Card. NH3 (de Grado)

This 1-tissue compartment model has been developed by DeGrado et al. [30] for cardiac PET studies using ¹³NH₃ ammonia bolus injection.

It includes a linear metabolite correction and describes the exchange of the NH₃ between blood and the myocardium by the following differential equation

The factor mCorr represents the metabolite correction factor. A value of 0.077 [1/min] has been found for mCorr in humans [30].

Additionally, the model incorporates a cardiac dual spillover correction by the operational equation

where
V_lv = spill-over fraction of the blood activity in the left ventricle C_lv(t),
V_rv = spill-over fraction of the blood activity in the right ventricle C_rv(t) .

DeGrado et al. recommend to only use the first 4 minutes of data after injection of the tracer to reduce the effects of metabolite buildup and washout.

Implementation Notes:

• The right ventricle curve is only used for spillover correction of septal TACs. It must be loaded as the Total blood curve.
• The left ventricle curve serves two purposes: (1) multiplied by the metabolite correction (1-mCorr*t) it serves as the input curve, (2) it is used for spillover correction of all myocardial TACs. It must be loaded as the Input curve.
• The spill-over fraction from the right ventricle V_rv is automatically fixed to zero if the string "Sep" is not contained in the name a region. The assumption is that such a TAC is not from septal tissue and should thus be modeled with spill-over from the left ventricle only. The reason for this automatism is usage of the model in the PCARD tool.
• To set V_rv = 0 in all regional models proceed as follows: in one region, set V_rv =0 and disable the fit checkbox; fit the region; configure the button below Copy to all regions to Model and Par. and activate it.
• To set mCorr to a different value in all regions (eg. for animal studies) proceed as described for V_rv above.

Abstract [30]

"BACKGROUND: Although several modeling strategies have been developed and validated for quantification of myocardial blood flow (MBF) from ¹³N- labeled ammonia positron emission tomographic data, a comparison of noise characteristics of the various techniques in serial studies is lacking. METHODS AND RESULTS: Dynamic ¹³N-labeled ammonia positron emission tomographic imaging was performed at baseline and after pharmacologic stress in (1) single studies of four dogs with concomitant measurement of microsphere blood flow and (2) initial and follow-up studies of eight normal volunteers. Data were obtained from short-axis images for the blood pool and myocardial regions corresponding to the three arterial vascular territories. Indexes of MBF were obtained by four distinct techniques: (1) University of California, Los Angeles, two-compartment model, (2) Michigan two-compartment model, and (3) a one- compartment model with variable blood volume term. Coronary flow reserve (CFR) was measured as the ratio of stress/rest MBF. The estimated standard deviation of the measurement error for the relative change between studies of rest and stress MBF and CFR was determined for each technique. Estimates of MBF from all techniques showed good correlation with microsphere blood flow (r = 0.95 to 0.96) in canine myocardium. In human studies, similar mean estimates of MBF were found with all techniques. Techniques 1 and 3 showed the smallest interstudy variability in MBF and CFR. The estimated standard deviations for these techniques were approximately 20%, 30%, and 27% for rest MBF, stress MBF, and CFR, respectively. CONCLUSION: Noninvasive quantification of MBF and CFR from dynamic ¹³N-labeled ammonia positron emission tomography is most reproducible with technique 1 or 3. The ability to account for differences in myocardial partial volume gives preference to technique 3. However, substantial interstudy variability in regional MBF remains, suggesting the importance of procedural factors or real temporal fluctuations in MBF."