On the Modelling of Facility-Specific PET Imaging for Proton Treatment Verification: Experimental Validation and Inter-Facility Comparison
J Bauer1,2*, M Hildebrandt3, D Unholtz1,2,4, C Kurz1,2, K Parodi1,2,3, (1) Heidelberg Ion-Beam Therapy Centre, Heidelberg, Germany, (2) Radiation Oncology, Department of Radiology, Heidelberg University Hospital, Germany, (3) Ludwig-Maximilians-University Munich, Garching, Germany, (4) now with Leica Microsystems CMS GmbH, Mannheim, GermanyTH-C-144-12 Thursday 10:30AM - 12:30PM Room: 144
We report on an experiment-driven approach to validate the Monte-Carlo (MC) modelling for Positron-Emission-Tomography (PET)-based post-therapeutic proton treatment verification. The presented strategy was pursued to overcome (1) uncertainties in the available nuclear cross-section data and (2) facility-specific uncertainties due to the irradiation technique and due to the imaging characteristics of the installed PET device. Furthermore, we show a comparison of production rates of the most relevant PET radionuclides, estimated from measurements at two treatment facilities using different PET-imaging and data processing concepts.
At our hospital-based facility, several homogeneous phantoms were irradiated with mono-energetic proton beams. The induced PET signal was detected with a full-ring PET/CT-scanner shortly after beam delivery. Separation of different radionuclide contributions to the measured signal was achieved by analysing the activity decay in dynamically reconstructed time-resolved PET images. The resulting spatially resolved activity maps were used to tune the modelling of the most relevant radionuclide production processes. These measurements were compared to data acquired at another facility with a proto-type dual head PET camera, employing a different data processing concept.
The MC modelling was adjusted to accurately reproduce the measurements at our facility. The activity signal level could be predicted within a few percent and we found a range agreement of better than 0.6 mm, which constitutes a crucial property for the clinical purpose of beam range verification. The inter-facility comparison of production rates estimated from independent PET measurements show a reasonable overall agreement. However systematic discrepancies are observed, which suggest an impact of the different imaging and data processing concepts.
The results of the inter-facility comparison confirm the importance of a facility-specific experimental validation of the MC modelling. The presented experimental strategy can also be applied at other facilities to minimize uncertainties in the activity prediction for PET-based proton treatment verification.
Funding Support, Disclosures, and Conflict of Interest: The presented work was funded by the German Federal Ministry for Research and Education (BMBF) under the grant agreement number 01IB08002F.