TY - JOUR
T1 - Measurement of myocardial blood flow with positron emission tomography
T2 - Correction for count spillover and partial volume effects
AU - Herrero, Pilar
AU - Markham, Joanne
AU - Myears, Donald W.
AU - Weinheimer, Carla J.
AU - Bergmann, Steven R.
N1 - Funding Information:
Funded in part by NHLBI grant HL 17646, SCOR in Ischemic Heart Disease and NIH Division of Research Resources Grant RR 01380, A Resource for Biomedical Computing. We thank Becky Parrack for preparation of the typescript. Address correspondence to Steven R. Bergmann, M.D., Ph.D., Cardiovascular Division, Washington University School of Medicine, 660 South Euclid Avenue, Box 8086, St. Louis, Missouri 63110.
PY - 1988
Y1 - 1988
N2 - We previously demonstrated that regional myocardial blood flow can be obtained with H215O using a one-compartment model when the input function and tissue tracer concentration are measured directly. Positron emission tomography (PET) is a quantitative imaging technique that permits assessment of the concentration of positron-emitting radiotracers within the body. However, because of the limited spatial resolution of the current generation of tomographs, measurement of tracer concentration in the heart with PET are subject to errors due to partial volume effects (underestimation of tracer concentration when imaging an object small with respect to the resolution of the tomograph) and count spillover (contamination of activity from one region to an adjacent one). Correction for these errors are based on the assumptions that the point spread function (the reconstructed response to a point of activity) of the imaging device in one dimension is Gaussian and that the activity in the blood pool and tissue are uniformly distributed. After experimental validation that the point spread function approximates a normal distribution, analytical profiles were obtained by convolving a cylinder (representing blood pool) and an annulus (representing tissue) of uniform activity with the Gaussian point spread function. These profiles were compared with PET reconstructions of a heart phantom. The correlation was excellent. Integration of the profiles were then used to derive regional recovery coefficients and spillover fractions. Subsequently, after bolus administration of H215O in eight dogs, myocardial blood flow was estimated non-invasively with PET using values of the input function and tissue radiotracer concentration corrected with the algorithm developed. Results correlated closely with flow values obtained with radiolabeled microspheres. The data suggest that quantitative estimates of radiotracer concentration can be obtained with PET when appropriate corrections for count recovery and spillover are employed, and that regional myocardial blood flow can be quantitated non-invasively.
AB - We previously demonstrated that regional myocardial blood flow can be obtained with H215O using a one-compartment model when the input function and tissue tracer concentration are measured directly. Positron emission tomography (PET) is a quantitative imaging technique that permits assessment of the concentration of positron-emitting radiotracers within the body. However, because of the limited spatial resolution of the current generation of tomographs, measurement of tracer concentration in the heart with PET are subject to errors due to partial volume effects (underestimation of tracer concentration when imaging an object small with respect to the resolution of the tomograph) and count spillover (contamination of activity from one region to an adjacent one). Correction for these errors are based on the assumptions that the point spread function (the reconstructed response to a point of activity) of the imaging device in one dimension is Gaussian and that the activity in the blood pool and tissue are uniformly distributed. After experimental validation that the point spread function approximates a normal distribution, analytical profiles were obtained by convolving a cylinder (representing blood pool) and an annulus (representing tissue) of uniform activity with the Gaussian point spread function. These profiles were compared with PET reconstructions of a heart phantom. The correlation was excellent. Integration of the profiles were then used to derive regional recovery coefficients and spillover fractions. Subsequently, after bolus administration of H215O in eight dogs, myocardial blood flow was estimated non-invasively with PET using values of the input function and tissue radiotracer concentration corrected with the algorithm developed. Results correlated closely with flow values obtained with radiolabeled microspheres. The data suggest that quantitative estimates of radiotracer concentration can be obtained with PET when appropriate corrections for count recovery and spillover are employed, and that regional myocardial blood flow can be quantitated non-invasively.
KW - Coronary blood flow
KW - non-invasive imaging
KW - oxygen-15 water
UR - http://www.scopus.com/inward/record.url?scp=0000615384&partnerID=8YFLogxK
U2 - 10.1016/0895-7177(88)90605-X
DO - 10.1016/0895-7177(88)90605-X
M3 - Article
AN - SCOPUS:0000615384
SN - 0895-7177
VL - 11
SP - 807
EP - 812
JO - Mathematical and Computer Modelling
JF - Mathematical and Computer Modelling
IS - C
ER -