Over the last 3 years in-beam PET has proven its capability of quality assurance in carbon ion tumour therapy at the pilot project of cancer therapy at the Gesellschaft für Schwerionenforschung (GSI) in Darmstadt, Germany [1]. It is planned to integrate a new positron camera into the proposed dedicated hospital-based facility for ion beam therapy in Heidelberg, Germany [2]. We have investigated the possibility of using LSO as the scintillation material for the next generation of in-beam PET cameras.
We address the issue of background coincidences arising from the natural radioactivity of LSO. Natural lutetium contains the radioactive isotope ¹⁷⁶Lu with an abundance of 2.59 %, causing a background activity of 280 Bq per cm³ of LSO. ¹⁷⁶Lu undergoes b^--decay feeding excited levels of ¹⁷⁶Hf at 597 keV and 998 keV with probabilities of 99.66 % and 0.34 %, respectively. These levels are de-excited via a prompt g-ray cascade of 307, 202 and 88 keV, which is topped by a 401 keV transition in the case of the 998 keV level. This results in undesired true coincidences when the electron and low-energy g-rays deposit enough energy in the crystal, where the b^--decay takes place, and the g-rays of higher energy escape and deposit their energy in an opposite crystal. Because the true count rate measured with the BGO-based dual-head positron camera BASTEI during the cancer treatment averages at only 100 coincidences/s we studied the influence of the LSO background on realistic in-beam PET images.
Therefore, we have estimated the background coincidences expected for the detector geometry of BASTEI assuming LSO as detector material. The estimation takes into account the energy distribution of the electron resulting from the b^--decay, together with all g-ray combinations that lead to a background coincidence. This has been done for energy windows of 250 - 850 keV, 350 - 650 keV and 400 - 600 keV, two g-ray combinations that are possible within the energy windows and an energy resolution of 15 % FWHM. Furthermore, the attenuation of the escaping 307 or 401 keV g-rays due to the presence of the patient has been taken into account.
The analysis [3] shows that the major component of the background of true coincidences caused by ¹⁷⁶Lu originates from the escape of the 307 keV g-ray and subsequent detection by another detector of the positron camera. If, however, an energy window between 350 - 650 keV or narrower is applied these background coincidences are effectively rejected and an influence on the reconstructed b⁺-activity distribution can be avoided (Fig. 1).

Fig. 1 The left Figure shows a typical b⁺-activity distribution as it was measured by the BGO-based positron camera BASTEI. The image on the right shows the same measured b⁺-activity distribution but with background coincidences added to the original PET scan as they would be expected for an LSO-based BASTEI assuming an energy window of 350 - 650 keV.

¹ Gesellschaft für Schwerionenforschung Darmstadt

References

[1] W. Enghardt et al, Nucl. Physics A 654 (1999) 1047c
[2] K.D. Gross, M. Pavlovic (eds.), Proposal for a dedicated ion beam facility for cancer therapy, 1998
[3] K. Lauckner et al, An LSO-based scanner for in-beam PET:
     A feasibility study, IEEE NSS/MIC Conf. Rec., 2000

IKH 06/26/01 © K. Lauckner