The purpose of the study was to compare the diagnostic accuracy of fluorodeoxyglucose (FDG) images obtained with a dual-head coincidence gamma camera (DHC) with those obtained with a dedicated PET in a series of 26 patients.
Methods: Nineteen patients with known or suspected malignancies and 7 patients with neurological disorders underwent PET imaging after injection of approximately 10 mCi of FDG. Whole-body imaging was performed on 19 patients and brain imaging on 7 patients. DHC images were then acquired for 30 min over the region of interest using a dual-head gamma camera equipped with 3/8-in.-thick NaI(TI) crystals and parallel slit-hole collimators. The images were reconstructed in the normal mode, using photopeak/photopeak, photopeak/Compton and Compton/photopeak coincidence events.
Results: Although the spatial resolutions of PET with a dedicated PET scanner and of DHC are in the same range, the lesion detectability remains superior with PET (4 mm for PET versus 13.5 mm for DHC in phantom experiments) with a contrast ratio of 5:1. This is most probably attributable to the higher sensitivity of PET (2238 coincidences/min/microCi for PET versus 89 coincidences/min/microCi for DHC). The pattern of uptake and interpretation for brain imaging was similar on both PET and DHC images in all patients. In the 19 oncology patients, 38 lesions ranging from 0.7 to 5 cm were detected by PET. DHC imaging detected 28 (73%) of these lesions. Among the 10 lesions not seen with DHC, 5 were less than 1.2 cm, 2 were located centrally within the liver and suffered from marked attenuation effects and 3 were adjacent to regions with high physiological activity. The nondetectability of some lesions with DHC compared with PET can be explained by several factors: (a) start of imaging time (mean+/-SD: 73+/-16 min for PET versus 115+/-68 min for DHC, leading to FDG decay to 6.75 mCi for PET and 5.2 mCi for DHC); (b) limited efficiency of a 3/8-inch-thick Nal(TI) crystal to detect 18F photons; (c) suboptimal two-dimensional reconstruction algorithm; and (d) absence of soft-tissue attenuation correction for centrally located lesions.
Conclusion: FDG DHC imaging is a promising technique for oncological and brain imaging.