An approach for attenuation corrected 3D internal radiation dosimetry

The key to individual patient dose calculation is a good estimate of isotope kinetics in organs of observed metabolism. Kinetics can be measured by organ specific time activity curves (TACs), derived from a series of planar whole body szintigrams, representing temporal evolution of radioactive uptake. Overlaps in projections and photon attenuation deteriorate directly measured counts in regions of interest (ROIs). In this work, corrections of TACs, based on three-dimensional (3D) SPECT and CT scans, are developed. An automated volume of interest (VOI) segmentation strategy combines local thresholds and topographic form-features from SPECT data with high-resolution CT morphology. VOIs are registered to whole body scans, allowing the differentiation of overlapping organs in projections. Photon attenuation is corrected based on CT scans, transformed to isotope specific linear attenuation coefficients. A simulation study is performed, based on real patient morphology, comparing native and uniformly attenuation corrected data, to non-uniform 3D attenuation correction. Its influence on effective patient dose is studied, and finally the new algorithms are applied to measured patient data. Attenuation correction, based on 3D data achieves substantially better results. Automated processing avoids time-consuming manual segmentation, giving raise to application in a clinical workflow.