Curtis Caldwell PhD, Presenter: Nothing to Disclose

Kathy Mah MS, Abstract Co-Author: Nothing to Disclose

Parminder Basran PhD, Abstract Co-Author: Nothing to Disclose

C.C.,K.M.: While the current work was not supported by any company, both Curtis Caldwell and Kathy Mah have research funding for other projects through Philips Medical Systems.

Purpose/Objective: PET/CT imaging promises to significantly impact the practice of radiation therapy simulation. Our objectives were to evaluate a new PET/CT scanner in terms of requirements for use as a radiation therapy simulator. Materials/Methods: A new PET/dual-slice CT scanner (Philips GEMINI PET/CT) was acquired solely for radiation oncology simulation and associated research. Acceptance testing was undertaken with priority in evaluating specifications required as a radiation therapy simulator. Testing was divided into the evaluation of i) CT imaging and associated mechanical components; ii) PET-imaging; iii) alignment and display of PET and CT datasets; and (iv) integration with the Philips Pinnacle Planning system. Performance of the CT-simulator sub-system was evaluated with tests described by the AAPM Task Group Reports 2 and 66, plus additional tests to examine multi-slice capabilities and to ensure accurate geometric simulation of treatment units. Performance of the PET sub-system was evaluated with NEMA 2001 standards plus additional tests of spatial integrity and linearity. Alignment and display of combined datasets were tested using geometric phantoms scanned under both CT and PET modes. Results: The conventional performance characteristics of both CT and PET imaging systems were more than acceptable. Using typical CT-simulation protocols requiring helical scanning with large FOVs, uniformity was better than 1%, high contrast resolution was typically 7 lp/cm, low contrast resolution was 5 mm @ 1.0% and 9.0 mm @ 0.5%, and slice sensitivity profile width was ≤ 0.5 mm for all collimations ≥1 mm. There was an expected increase in radiation profile thickness for the multi-slice CT images due to beam divergence as well as signature image artifacts from multi-slice helical scanning. For PET images, for a point 10 cm off-axis, transverse radial and transverse tangential resolution was 5.3 and 5.6 mm (FWHM), respectively. The peak true count rate was 100. kcps at 54.8 kBq/ml and the peak noise equivalent count rate was 39.0 kcps at 37.4 kBq/ml. The average measured system scatter fraction was 41.6 %. Near axis, the system sensitivity was 2.0 kcps/MBq and 3.4 kcps/MBq, respectively, for trues and trues plus scatter. Integration of the co-registered modalities was spatially accurate to within a PET voxel. The same multi-modality imaging viewing and co-registration software is available on both the scanner console and the Pinnacle planning system, making for ease of integration. While image quality in both modalities was excellent, there were mechanical challenges in using the Gemini as a radiotherapy simulator due to greater demands for geometrical accuracy when simulating therapy than is required in a PET/CT unit for diagnosis. Two examples where modifications were required to permit accurate PET/CT simulation were noted. (1) There was a systematic couch tilt of approximately 1o. This angle corresponds to a vertical rise of 2.5 cm in a scan length of 150 cm in the cranio-caudal direction. This was initially dealt with an in-house designed modification to the patient pallet. Philips subsequently supplied a newly designed patient pallet. (2) A second finding was a 0.5o in-plane rotation in reconstructed CT images. Although insignificant for diagnostic purposes, this in-plane rotation can lead to a 3 mm discrepancy between true horizontal points over a 30 cm separation, resulting, for example, in a systematic error in patient lateral set-ups. This problem was solved by physical rotation of the CT gantry. Conclusions: While care must be taken in the acceptance of any PET/CT device for radiation therapy simulation purposes, the GEMINI device proved capable of meeting standards for simulation, albeit after some modification. Since November 2003, the PET/CT-simulator has been in clinical use solely as a radiation therapy simulator. Its performance has met the geometrical tolerances required for radiation therapy while achieving excellent patient image quality.

Caldwell, C, Mah, K, Basran, P, Evaluation of a Combined Positron Emission Tomography (PET)/Computed Tomography (CT) Scanner for Radiation Therapy Simulation.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4418017.html