Page 363 - Biomedical Engineering and Design Handbook Volume 2, Applications
P. 363

NUCLEAR MEDICINE IMAGING INSTRUMENTATION  341

                          attenuation, while those that traverse the thickest portion of the patient can have path lengths of
                          greater than 50 cm in large patients, leading to reductions in the coincidence detection rate of more
                          than a factor of 100. This causes a nonlinear relationship between the detected counts and the source
                          activity. Thus, reconstruction of the acquired data without attenuation correction leads to the char-
                          acteristic artifacts, including enhanced count density at the skin and in the lung fields and decreased
                          count density in the central portions of the patient (Fig. 11.16b).
                            The amount of attenuation depends only on the trajectory of the annihilation photons through the
                          patient, not on the actual location of the source. Measurement of the transmission factors can be per-
                          formed using radionuclide sources that revolve about the patient using the PET tomograph as a crude
                                                                                           68
                          CT scanner. 11,30  Radionuclide sources that have been used for that application include  Ge and
                          137 Cs. However, collecting the transmission information this way is slow (1 to 3 minutes per bed
                          position) and produces low-resolution, noisy corrections. These problems have been largely solved
                          by using CT tomographic images from the combined PET CT imaging systems (Fig. 11.17).
                            The CT scanner provides whole body transmission data in less than 1 minute. In addition to the
                          reduction in transmission time, the level of noise in the CT images is much less than that of the
                          radionuclide transmission data. Because the average energy of the x-rays used in the CT scan is 60 to
                          80 keV, the CT attenuation coefficients have to be mapped to the appropriate values at 511 keV.
                          Although this is a nonlinear process, the mapping has been successfully implemented. 31  However,
                          there are several potential problems that occur when using CT data for attenuation correction.


                                  A                                  B






































                                 FIGURE 11.17  (a) PET CT scanners from General Electric, Philips, and Siemens. (b) An
                                      18
                                 example  F FDG whole body PET CT clinical study.
   358   359   360   361   362   363   364   365   366   367   368