Effect of 176Lu intrinsic radioactivity on dual head PET system imaging and data acquisition, simulation, and experimental measurements.

PURPOSE: In this work, the authors aim for the estimation of the effect of (176)Lu intrinsic radioactivity on the performance of a dual head PET system dedicated to small animal imaging.

METHODS: A prototype camera has been used as a reference system in order to validate two GATE simulation models, which were used for the investigation. The first model includes the (176)Lu intrinsic radioactivity, while the second does not. The two models have been designed in order to provide similar sensitivities, in terms of count rate performance and scatter fraction, in the linear range of activities. In addition, the model with the (176)Lu intrinsic radioactivity, has been validated in terms of background count rate. Different acquisition schemes have been examined in order to determine the optimum conditions to minimize the (176)Lu effects, while maintaining a high trues count rate. In addition, the effect on the image quality, in terms of spatial resolution, signal-to-noise ratio, and minimum detectable activity, was investigated.

RESULTS: Both models are in good agreement with the measured data. While, the presence of the (176)Lu altered the dead time of the model, it also affected the singles, trues, and randoms count rates. The noise equivalent count rate curves of the two models indicate that for low activities, the lack of (176)Lu radioactivity leads to better noise properties due to the underestimation of the randoms. Signal-to-noise ratio measurement on coincidence images confirm the aforementioned claim, since the model without the (176)Lu provides better less noisy images. Furthermore, the spatial resolution and the minimum detectable activity are overestimated.

CONCLUSIONS: It has been proven that the lack of the (176)Lu intrinsic radioactivity has an impact on the design of the simulation model's dead time. Even if there is an alignment with experimental results still the noise properties, for a wide range of activities, are overestimated. In addition, for low activities, better image quality, is presented, mainly due to the lack of the (176)Lu photons which blur the images. Furthermore, it was shown that the use of a high hardware threshold allows the application of a wide energy window, which improves the statistics without a significant compromise on the quality of the detected photons, mainly, due to the early rejection of the low energy photons and the low scatter factor.