The Effects of Respiratory Motion on CT Based Attenuation Correction of Positron Emission Tomography Data: Liver and Heart Studies

Current clinical FDG imaging with the NIH's PET/CT machine (and in fact with most commercial PET/CT machines) requires a CT scan be acquired prior to the PET scan. This CT scan is used for attenuation correction of the PET data and to permit fusion of anatomical and metabolic image data. One difficulty with use of the CT scan for attenuation correction ('CTAC') is that the CT scan is rapid compared to the breathing cycle. Each CT slice captures the lungs at one (usually arbitrary) phase of the respiratory cycle. The PET data on the other hand take many minutes to acquire, and so average the motion effects of the entire respiratory cycle together. This 'freezing' of the respiratory cycle by the CT, and blurring of the respiratory cycle during PET produces a mis-match between the PET and CT data. This mis-match can produce improper attenuation correction during PET reconstruction, especially at soft tissue/lung interfaces. The existence of this phenomenon has been reported in the literature \[1-7\], especially at the dome of the liver, but it has not been thoroughly quantified. Little or no data is available to indicate the quantitative errors the effect produces in cardiac imaging\[8\]. This despite the fact that the myocardium is only about 10mm thick, and it has been reported that the heart moves on average 9mm (and as much as 14mm) during a single average respiratory cycle\[9\]. The purpose of this protocol is to determine the degree to which respiratory motion may influence quantitative PET imaging, especially at the lung/liver interface and free wall of myocardium. By characterizing the magnitude of the effect, we hope to gain knowledge about when correction for the effect is and is not necessary. It is hoped that the information gained will also allow us to suggest potential methods to perform such corrections. To gain the above information we will modify the CT acquisition protocol for subjects already scheduled to undergo a whole body FDG PET/CT scan. In summary, rather than acquiring the usual single full radiation exposure CT (taking approx 20-30 seconds), we will acquire 3 CT scans (each taking approximately 20-30 seconds), each at 1/3 the usual radiation exposure. Each CT will be acquired at a different phase of the respiratory cycle (as opposed to current acquisition, which is only at end-expiration). The subsequent PET data will be acquired as usual, but will be processed and analyzed separately with each of the CT scans and also with a summed CT scan. Note that subjects agreeing to participate in this protocol will receive no additional radiation exposure. Participation in the protocol will extend the current overall scan time (typically greater than 45 minutes) a very small amount (about 40-60 additional seconds for imaging, and about 60 additional seconds for patient set up, etc).