We demonstrate the reconstruction of a 3D, time-varying bolus of radiotracer from first-pass data obtained at the dynamic SPECT imager, FASTSPECT, built by the University of Arizona. The object imaged is a CardioWest Total Artificial Heart. The bolus is entirely contained in one ventricle and its associated inlet and outlet tracts. The model for the radiotracer distribution is a time-varying closed surface parameterized by 162 vertices that are connected to make 960 triangles, with uniform intensity of radiotracer inside. The total curvature of the surface is minimized through the use of a weighted prior in the Bayesian framework. MAP estimates for the vertices, interior intensity and background scatter are produced for diastolic and systolic frames, the only two frames analyzed. The strength of the prior is determined by finding the corner of the L-curve. The results indicate that qualitatively pleasing results are possible even with as few as 1780 counts per time frame (total after summing over all 24 detectors). Quantitative results will require correcting certain undesirable features of the reconstruction due to inappropriate assumptions in the model, e.g. inhomogeneities in the radiotracer distribution and smoothness of the surface at the tract/ventricle join.
Keywords: limited-angle tomography, Bayesian reconstruction, Bayes Inference Engine (BIE), curvature prior, deformable model, single photon emission computed tomography (SPECT), FASTSPECT, heart imaging
Get full paper (pdf, 320 KB)
Return to publication list
Send e-mail to author at kmh@hansonhub.com