[Development of a new method of high-speed rotation multiplied projection single photon emission computed tomography of the triple-headed type: lung MAA SPECT based on phantom study].

PURPOSE A chest phantom study was conducted to evaluate the image quality of newly developed high-speed rotation multiplied projection-single photon emission computed tomography (HSRMP-SPECT) images. MATERIALS AND METHODS HSRMP-SPECT images of a chest phantom consisting of a simulated lung structure filled with 5000 ml of water containing 185 MBq Tc-99m-pertechnetate, and several small 11 mm simulated lung nodules of glass balls and one large 35 mm simulated lung nodule of a plastic sphere filled with water were obtained using a triple-headed SPECT system. During image acquisition, this phantom was regularly moving in the head-to-caudal direction with a range of 12 mm at a frequency of 15 cycles/min to simulate respiratory motion, and 360 degrees projection data of this moving phantom was acquired with an image acquisition time of 20 sec, which was repeated 10 times. To eliminate the setting time between projection and acquisition of multiple temporal samples of data, each detector was continuously rotated in the clockwise direction for 20 sec around a 120-degree arc. On the perspective SPECT images reconstructed from various numbers of the 20-sec projection data, the perfusion heterogeneity of the simulated lungs and perfusion defect clarity of the simulated nodules were assessed by the coefficient of variation (CV) of pixel counts and the defect-to-lung radioactivity ratios, respectively. The results were compared with those on conventional SPECT images of the moving phantom obtained with a data acquisition time of 8 min, and SPECT images of the standing phantom obtained with the same data acquisition time. RESULTS The average CV value of 0.28+/-0.01 on the SPECT image reconstructed from 5 projection data sets was not significantly different from that of 0.27+/-0.01 on the SPECT image reconstructed from 10 projection data sets (p<0.05). The perfusion defect contrast of the simulated nodules obtained from 5 projection data was significantly higher than that on conventional SPECT images (0.50 vs. 0.73) . CONCLUSIONS The present phantom study indicated that HSRMP-SPECT could be a useful technique for quickly obtaining high-quality SPECT images of a moving subject, thereby improving perfusion defect clarity in comparison with the conventional technique. This technique may have potential utility for obtaining high-quality breath-hold SPECT images of the chest in clinical practice.