The effect of finite spatial resolution on the measurement of cardiac phantom wall thickness in single photon emission computerized tomographic imaging.

OBJECTIVE To devise and validate a method of estimating accurately myocardial wall thickness from cardiac images acquired with single photon emission tomography (SPECT). MATERIALS AND METHODS We simulated the imaging process by convolving a spatial resolution function, experimentally determined for a clinical SPECT system, with rectangular profiles mimicking the myocardial walls for different thicknesses and separations. Wall thicknesses were estimated by fitting the resulting profiles to a linear superposition of two offset Gaussians. The method was validated by extensive computer simulation and by testing on real phantom images. RESULTS Accurate estimates of the wall thickness of SPECT phantoms were obtained when the estimated thickness (using fitting to Gaussians) was deconvolved with the point spread function (PSF) of the imaging system using a look-up table. CONCLUSIONS This method is a novel method of estimating wall thickness from cardiac images. It is particularly useful for small separations (e.g. at end systole and/or towards the apex of the heart) of walls that are narrow compared to the PSF of the imaging system.