Abstract
93
Objectives: Static whole body (SWB) imaging in PET suffers from image blurring and underestimation of activity concentration due to respiratory motion. Phase based gating (PBG) and amplitude based optimal gating (HD_Chest) have been shown to reduce these effects, but at the cost of increased image noise due to their use of a fraction of the acquired data. One approach to reduce image noise is to make use of all the acquired data by employing a mass preserved optical flow (MPOF) algorithm along with an elastic motion correction (EMC) reconstruction. MPOF determines motion vectors between a reference gate and the remainder of the acquired PET data, and incorporates these motion vectors into EMC reconstruction (EMC_MPOF). EMC_MPOF uses all of the acquired counts for the acquisition, rather than a fraction, decreasing the image noise encountered with PBG and HD_Chest. This work evaluates the impact of PBG, HD_Chest, and EMC_MPOF in comparison to SWB on image quality and quantitation of lung and liver lesions.
Methods: 28 patients were imaged on a 4 ring Siemens mCT with continuous bed motion. A table speed of 0.5 mm/s was used over the areas affected by motion, and a respiratory waveform was acquired with the Anzai belt. Four image reconstructions were performed: SWB, PBG with 8 bins, HD_Chest, and EMC_MPOF. All reconstructions were performed with 2 iterations, 21 subsets, TOF, PSF, and a 5mm Gaussian filter. Measurements of lesion SUVmax, contrast to noise ratio (CNR) and background noise were made for 41 lesions located in the lung or liver. Background noise was determined from the STDev of a 3 cm diameter ROI in healthy lung and liver tissue. The percent change of the average values were made with respect to SWB reconstructions. For the PBG reconstructions, the average value (PBGav) of the eight bins and the bin with the maximum value (PBGmax) were compared respectively to SWB.
Results: EMC_MPOF, HD_Chest, and PBG, all resulted in increased lesion SUVmax compared to SWB. While PBGmax had the highest SUVmax value due to image noise, HD_Chest and EMC_MPOF had similar but higher values (by 20%) than SWB due to a reduction in motion blur. Increases in background noise in the lung and liver was the lowest for EMC_MPOF (4 & 14%) followed by HD_Chest (38 & 56%) and PBG (>100%) due to including more of the acquired PET data when compared to SWB. CNR % change in the lung and liver lesions showed a similar trend with the highest value for EMC_MPOF (20 & 8%) followed by HD_Chest (-20 & -50%) and PBG (-41 & -67%). The quantitative results are tabulated below.
Conclusion: While HD-Chest and PBG reduce motion blur and increase SUVmax, the use of all the acquired PET data in EMC_MPOF has the added advantage of also reducing background noise thereby improving image quality. Research Support: SIEMENS research grant