Comparison of dedicated breast positron emission tomography and whole-body positron emission tomography/computed tomography images: a common phantom study

Abstract

Objective High-resolution dedicated breast positron emission tomography (dbPET) can visualize breast cancer more clearly than whole-body PET/computed tomography (CT). In Japan, the combined use of dbPET and whole-body PET/CT is necessary in indications for health insurance. Although several clinical studies have compared both devices, a physical evaluation by the phantom test has not been reported. The aim of this study was to compare the ability of ring-shaped dbPET and whole-body PET/CT using a common phantom with reference to the Japanese guideline for the oncology 18 F-fluorodeoxyglucose (FDG)-PET/CT data acquisition protocol. Methods A cylindrical breast phantom with four spheres of different diameters (16, 10, 7.5, and 5\xa0mm) filled an FDG solution at sphere-to-background radioactivity ratios (SBRs) of 2:1, 4:1, and 8:1 was prepared. Images were then acquired by whole-body PET/CT and subsequently by dbPET. The reconstructed images were visually evaluated and the coefficient of variation and uniformity of the background (CV background and SD ΔSUVmean ), percentages of contrast and background variability (% Q H,5mm and % N 5mm ), and their ratio (% Q H,5mm / N 5mm ), and relative recovery coefficient were compared with the standards defined in the protocol for whole-body PET/CT. Results The parameters were calculated at an SBR of 8:1, which was the only SBR in which a 5-mm sphere was visible on both devices. The standards were defined as\u2009<\u200910% for CV background ,\u2009≤\u20090.025 for SD ΔSUVmean ,\u2009<\u20095.6% for % N 5mm ,\u2009>\u20092.8 for % Q H,5mm / N 5mm , and\u2009>\u20090.38 for the relative recovery coefficient of the smallest sphere (10\xa0mm in diameter) in the protocol for whole-body PET/CT (the % Q H,5mm was not determined for that protocol); the respective values were 6.14%, 0.024, 4.55%, 3.66, and 0.33 for dbPET and 2.21%, 0.021, 3.11%, 1.72, and 0.18 for PET/CT. The Q H,5mm was 16.67% for dbPET and 5.34% for PET/CT. The human images also showed higher lesion-to-background contrast on dbPET than on PET/CT despite the noisier background observed with dbPET. Conclusion The common phantom study showed that the background was noisier and that the contrast was much higher in the dbPET image than in the PET/CT image. The acquisition protocol and standards for dbPET will need to be different from those used for whole-body PET/CT.