Premium
Sci—Thur PM: Imaging — 01: Position‐sensitive noise characteristics in multi‐pinhole cardiac SPECT imaging
Author(s) -
CuddyWalsh SG,
Wells RG
Publication year - 2014
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4894974
Subject(s) - imaging phantom , pinhole (optics) , noise (video) , voxel , attenuation , sensitivity (control systems) , standard deviation , physics , image resolution , single photon emission computed tomography , myocardial perfusion imaging , optics , field of view , position (finance) , image noise , signal to noise ratio (imaging) , nuclear medicine , artificial intelligence , computer science , medicine , mathematics , radiology , perfusion , statistics , finance , economics , image (mathematics) , engineering , electronic engineering
Myocardial perfusion imaging (MPI) with Single Photon Emission Computed Tomography (SPECT) is invaluable in the diagnosis and management of heart disease. It provides essential information on myocardial blood flow and ischemia. Multi‐pinhole dedicated cardiac‐SPECT cameras offer improved count sensitivity, and spatial and energy resolutions over parallel‐hole camera designs however variable sensitivity across the field‐of‐view (FOV) can lead to position‐dependent noise variations. Since MPI evaluates differences in the signal‐to‐noise ratio, noise variations in the camera could significantly impact the sensitivity of the test for ischemia. We evaluated the noise characteristics of GE Healthcare's Discovery NM530c camera with a goal of optimizing the accuracy of our patient assessment and thereby improving outcomes. Theoretical sensitivity maps of the camera FOV, including attenuation effects, were estimated analytically based on the distance and angle between the spatial position of a given voxel and each pinhole. The standard deviation in counts, σ was inferred for each voxel position from the square root of the sensitivity mapped at that position. Noise was measured experimentally from repeated (N=16) acquisitions of a uniform spherical Tc‐99m‐water phantom. The mean (μ) and standard deviation (σ) were calculated for each voxel position in the reconstructed FOV. Noise increased ∼2.1× across a 12 cm sphere. A correlation of 0.53 is seen when experimental noise is compared with theory suggesting that ∼53% of the noise is attributed to the combined effects of attenuation and the multi‐pinhole geometry. Further investigations are warranted to determine the clinical impact of the position‐dependent noise variation.