Annual Report

2. Effects of External End-Shields for Positron Emission Tomography

Tomoyuki Hasegawa and Hideo Murayama

Keywords: radiation shields, Monte Carlo simulation, positron emission tomography, nuclear medicine

Radiation from outside the field-of-view of a PET system operating in the 3D data collection mode increases the single event count rates of the system, which produces an increase in the accidental count rate. These accidental coincidences can be measured and subtracted from the image data, however the process adds statistical noise to the data and causes deterioration in the quality of the images. The effectiveness of an end-shield depends on its diameter, thickness and composition, and the distribution of the sources of radiation.

The system employed in this work was an ECAT EXACT HR+ (CTI/Siemens produced), operating in the 3D data collection mode with the standard data collection parameters (span 9 and maximum ring difference 21). Each end-shield was a circular slab of lead mounted on a steel plate for strength and rigidity. The sides employed were 10 or 20 mm thick lead, 3.2 or 6.4 mm thick steel, 300 or 350 mm inner diameter, and 550 mm outer diameter. The end-shield was mounted on the gantry just outside the first detector. The standard uniform cylinder phantom (20 cm inner diameter and 18.5 cm inner length) containing 18F dissolved in water was placed 34 cm from the center of the axial field-of-view of the scanner. By using the Monte Carlo code based on GEANT 3.21, the single event and accidental count rates were obtained. 100 million events were simulated for each condition under study. It was confirmed that the simulated sensitivity and scatter fraction agreed with phantom measurement to be better than 5 %.

The measured accidental count rates versus the Monte Carlo results for each configuration are plotted in fig. 2. The lightest end-shield (t10-d350) reduced the accidental count rate to about 3 % of the count rate without end-shields (no e-sh). The rate was slightly better with the thicker end-shield of the same diameter (t20-d350). A somewhat larger drop in count rate was noted with the shield with the smaller inner diameter (t10-d300). These results indicate that 10 mm of shielding is adequate and that larger reductions in background are caused by blocking line-of-sight access of the radiation to the detector. Physical measurements and Monte Carlo simulation results confirm that even relatively thin lead shielding can dramatically reduce the count rates for activity originating outside of field-of-view of the PET system. Users should be aware of the possibility that scatter in the shielding or support material can lead to significant background rates.

Hasegawa, T., Michel, C., Kawashima, K., Murayama, H., Nakajima, T., Matsuura, H. and Wada, Y. : IEEE Trans. Nucl. Sci., 47, 1099-1103, 2000.

Fig.2. Measured random count rates in comparison with Monte Carlo calculations.

[Back to index]