MYOCARDIAL VIABILITY STUDIES USING FLUORINE-18-FDG SPECT - A COMPARISON WITH FLUORINE-18-FDG PET

Multidetector SPECT systems equipped with a high-energy, or 511-keV co llimator, have been proposed to offer a less expensive alternative to PET in myocardial viability studies with [F-18]FDG. The objectives of this investigation included: (a) measuring the physical imaging charac teristics of SPECT systems equipped with either a high-energy general- purpose collimator (HE), or the dedicated 511-keV collimator (UH), whe n imaging 511-keV photons, and comparing them with conventional FDG PE T; and (b) directly and quantitatively comparing the diagnostic accura cy of SPECT, with either an UH or HE collimator, to that of PET in myo cardial viability studies using F-18-FDG. Methods: Physical imaging ch aracteristics of SPECT and PET were measured and compared. Both SPECT and PET studies were performed in two groups of 18 patients each, with Group I using HE SPECT and Group II using UH SPECT. Myocardial perfus ion studies were also performed using Rb-82 PET at rest and during dip yridamole stress to identify areas of persistent hypoperfusion. For ea ch myocardial region with a persistent perfusion defect, a perfusion-m etabolism match or mismatch pattern was established independently, bas ed on the results of F-18-FDG SPECT as well as PET, Results: PET is su perior to SPECT in all physical imaging characteristics, particularly in sensitivity and contrast resolution. PET had a sensitivity 40-80 ti mes higher than that of SPECT, and its contrast resolution was 40-100% better than SPECT. Between FDG-SPECT using an HE collimator and that using a 511-keV collimator, the latter showed marked reduction in sept al penetration (from 56% to 38%), improvement in spatial resolution (f rom 17 mm to 11 mm FWHM) as well as contrast resolution (from 34% to 4 5%), while suffering reduced system sensitivity (from 75 to 34 cpm/mu Ci). Patient studies demonstrated that although FDG-SPECT, using a HE or UH collimator, provided concordant viability information as FDG PET in a large majority of myocardial segments with persistent perfusion defects (88% and 90%, respectively), there is an excellent statistical agreement (kappa = 0.736) between SPECT with UH collimator and PET, w hile the agreement between SPECT using HE collimator and PET are moder ate (kappa = 0.413). Conclusion: Despite its markedly inferior physica l imaging characteristics compared with PET, SPECT with the dedicated 511-keV collimator offers a low-cost, practical alternative to PET in studying myocardial viability using [F-18]FDG. SPECT systems with a hi gh-energy, general-purpose collimator, on the other hand, are inadequa te in such studies.