Absolute radionuclide concentration measurement using maximum-likelihood expectation-maximization iterative reconstruction, attenuation, and scatter correction

The aim of this study was to evaluate the accuracy with which radionuclide concentration could be measured after implementation of the channel ratio (CR) scatter correction method and incorporation of transmission attenuation coefficients into a maximum-likelihood expectation-maximization iterative reconstruction algorithm regularize using a multinomial prior. A water-filled thorax phantom containing a liver insert and a variable spleen volume was used to simulate different clinical situations. An uncollimated Co-57 sheet source was used to obtain attenuation matrices. All emission data were acquired in two 10% energy windows straddling the photopeak. Planar and SPECT sensitivities were determined. After scatter correction was performed data were first reconstructed using the measured attenuation matrices, and, second, using the good geometry attenuation coefficient for water. Radionuclide concentration with the attenuation matrix using 64 projections varied between 48.9/spl plusmn/3.1% (49.6/spl plusmn/3.1%) and 76.5/spl plusmn/3.0% (76.5/spl plusmn/3.2%) when 25 and (50) iterations were used. Similar results were obtained using 128 projections, and no statistical difference could be found (p<0.05). The inaccuracy of the results obtained with the implementation of the attenuation matrix from the transmission tomogram is due to the effective attenuation coefficients used in conjunction with the scatter compensation method. Results obtained with the attenuation coefficient of water varied between 70.1/spl plusmn/3.1% (70.8/spl plusmn/3.0%) and 103.2/spl plusmn/3.5% (103.3/spl plusmn/3.4%). The influence of volume and concentration is clearly demonstrated. Edge detection plays an important role in the accuracy of concentration calculations.