In perspective of the proposed proton and heavy ion (up to oxygen nuclei) therapy facility at Heidelberg [1] and on the basis of the positive impact of positron emission tomography to quality assurance of carbon ion therapy at GSI Darmstadt [2], we investigated the potential of PET for proton therapy monitoring [3]. We simulated by means of the FLUKA Monte Carlo code the b⁺-activation of a 9×9×30 cm³ PMMA target (C₅H₈O₂, r = 1.19 g/cm³) irradiated by monoenergetic pencil-like proton beams (FWHM   = 10 mm) in the energy range of interest for therapy (70 - 200 MeV). The selected target contains the most abundant elements of the human body and is very suitable for experimental investigations. The simulated yield of b⁺-emitters of main relevance for the PET monitoring, namely ¹¹C, ¹⁵O and ¹⁰C, and the outcoming activity in 5 minutes (as a typical irradiation time) were found one order of magnitude lower than those induced by carbon ion irradiation at the same range and number of projectiles. But keeping into account the @ 20 times higher fluence of protons required in order to deliver the same physical dose and, furthermore, the lower relative biological effectiveness of protons in comparison to carbon ions, the activation induced by protons is expected to be at least twice as intense than for carbon ions. However, the spatial correlation between the positron emitter distribution and dose is poorer for protons, since they cannot experience the projectile fragmentation reaction leading to the sharp activity peak close to the dose maximum in the carbon ion case (Fig. 1, right). Nevertheless, the range and Bragg peak position of protons are still correlated to the distal edge of the b⁺-emitter depth distribution (Fig. 1, left), depending on the energy threshold of the fragmentation reactions and on the O/C ratio of the irradiated target. Therefore, an important check of particle range and dose localisation seems to be possible for proton irradiation, too.
In practical clinical cases a strategy could be to compare the measured activity with a simulated pattern based on the treatment plan, as already done in the carbon therapy case [2]. Moreover, an in-beam PET scanner is required in order to exploit the quantitative gain in the activity signal, dominated by the contribution of ¹⁵O (T_1/2 = 121.8 s) according to our estimations in typical mean irradiation times. On the basis of these simulated results and of their agreement with a previous set of off-line PET measurements [4], sensitive only to the activity contribution coming from ¹¹C, a new experiment has been performed with the FZR in-beam PET scanner at GSI Darmstadt in December 2000. The data analysis is currently in progress.

Fig. 1 Simulated spatial depth distribution of b⁺-emitters produced by 10⁶ proton and carbon ion projectiles. The dashed-dotted line displays the dose profile. The factor of about 20 between the amount of dose delivered by the same number of primary particles can be seen.

References

[1] K. D. Gross and M. Pavlovic (eds), Proposal for a dedicated ion beam
     facility for cancer therapy (GSI Darmstadt, 1998) pp 43-6
[2] W. Enghardt et al., Strahlenther. Onkol. 175 (1999) 33
[3] K. Parodi and W. Enghardt, Phys. Med. Biol. 45 (2000) N151
[4] U. Oelfke et al., Phys. Med. Biol. 41 (1996) 177

 IKH 06/26/01 © K. Parodi