Operator-less processing of myocardial perfusion SPECT studies.

Abstract

We have developed a completely automated algorithm to generate reoriented tomographic images from projections in myocardial perfusion SPECT. The algorithm consists of three software modules. The first module determines reconstruction limits for the projection dataset using two-dimensional feature extraction techniques. The second module reconstructs the projection images into transaxial images using standard filtered backprojection. The third module reorients the transaxial images into short-axis images. The algorithm was validated on 350 rest 201Tl and 350 stress 99mTc-sestamibi studies acquired on a single-detector (178 studies), a 90 degrees dual-detector (230 studies) or a triple-detector camera (292 studies). The complete processing sequence was successful in 93.6% of the studies (166/178 + 216/230 + 273/292). As for the individual modules, myocardial boundaries were correctly determined in 96.3% of the studies (171/178 + 222/230 + 281/292), while reorientation was successful in 97.2% of the studies (166/171 + 216/222 + 273/281). No significant difference in success rates for 201Tl versus 99mTc-sestamibi images was found. Our automated approach to myocardial perfusion SPECT processing is highly successful, intrinsically reproducible and can produce time and cost savings while improving accuracy in a clinical or research environment.