Lesion detection with single-photon emission computed tomography (SPECT) compared with conventional imaging.

Abstract

We have evaluated analytically and experimentally the effectiveness of both conventional nuclear medicine imaging and single-photon emission computed tomography (SPECT) imaging to detect small photon-deficient areas (approximately the size of the system's resolution) with a relatively uniform background. The experimental model is based on the Tc-99m sulfur colloid study of the liver. The experimental data were obtained from a liver phantom containing two small photon deficient areas, nominally 1 and 1.5 cm in diameter. The liver phantom was placed in a water-filled Alderson body phantom and scanned with the cold defects located both centrally and peripherally. Lesion image contrast for both conventional and SPECT imaging is proportional to the lesion uptake ratio and is degraded by the system's finite spatial resolution and Compton-scattered photons. However, for conventional imaging the contrast is significantly degraded by the effect of radionuclide superposition (as modified by attenuation), while for SPECT imaging the contrast is essentially independent of these effects. This results in a significant increase in lesion-to-background contrast with SPECT as compared with conventional imaging. The measured SPECT image contrasts for the 1- and 1.5-cm areas of low uptake averaged more than five times the measured image contrasts for the conventional system.