Wednesday, 18 September 2002
Proffered papers THERAPY IMAGING 37
Megavoltage CT for image-guided radiotherapy K. Ruchala 1, G. Ofivera 1,2, L. Forrest 3, M. Mehta 4, J. Kapatoes 1, J. Welsh4, T. Mackie 1,2,4 lTomoTherapy Inc, Middleton, Wisconsin, U.S.A. 2University of Wisconsin, Madison, Medical Physics, Madison, WI, U.S.A. 3University of Wisconsin, Madison, Veterinary Medicine, Madison, Wl, U.S.A. 4University of Wisconsin, Madison, Human Oncology, Madison, WI, U.S.A. Purpose: The development of on-board imaging systems can enable improved patient positioning and verification for radiotherapy. This capability is particularly important in the context of IMRT, as the ability to deliver precise treatments is ideally complemented by the ability to detect anatomical changes. To this end, an on-board megavoltage CT (MVCT) imaging system was developed for image-guided radiotherapy. Materials and Methods: A 738 channel xenon detector was incorporated into a helical tomotherapy system The megavoltage linear accelerator was used as the X-ray source, but when in MVCT mode the linac tuning is adjusted to reduce dose. The MVCT FOV is 40 cm, which is established by the field width at the axis with all MLC leaves open. Helical slices were collected in which 2 slices were collected per 10 sac gantry period, although periods as low as 4 sac are possible. Performance phantoms were scanned to determine system characteristics suchas contrast, resolution, linearity, and dose. Two pet clogs being treated for spontaneously arising nasopharyngeal sarcoma were also imaged on this system. Images collected in these studies were also used to test automatic registration techniques that align on-board images and planning images. Results: Phantom scans showed that clinically useful images can be collected with doses of 1 to 3 cGy. With a 2 cGy scan, contrasts of 3% can be clearly seen. Air-holes can be resolved at sizes of 1.2 mm for a 512 x 512 reconstruction matrix covering the 40 cm FOV. In the veterinary MVCT images, the tumor can be identified filling half of the nasal cavity and causing bony destruction, Tests of automatic image registration techniques demonstrate that MVCT images can be registered to kVCT planning images to further assist with patient repositioning and verification, Conclusions: MVCT has been found to be a fast and efficient method for collecting on-board patient images in the context of radiotherapy verification. It can provide tomographic images with contrasts of 3% and resolulions of 1.2 mm with a dose of 2 cGy. Developments in detector efficiency, use of multi-row detectors, and increased gantry speeds can further improve performance, 38
The CCD-camera based electronic portal imaging device (EPID) revisited; can a-Si flat panel EPIDs do better ? J.C.J. de Boer, J.C. Bamhoorn, B.J.M. Heiirnen Erasmus MC - Danie/den Hoed Cancer Center, C/inica/physics, Rotterdarn, The Netherlands Background: There are several reasons to actively pursue further development of CCD-camera based EPIDs (CCD-EPIDs). Setup corrections derived from CCD-EPID images yield remaining systematic errors of 1-1.5 mm (1SD) for various treatment sites. Furthermore, CCD-EPIDs are extremely stable and have proven successful in accurate transit dosimetry and desimetrical QA of IMRT fields produced with dynamic multileaf collimation. With modern CCD-cameras excellent image quality can be obtained. The main drawback of CCD-EPIDs is the gradual degradation of image quality ('hot pixels') due to exposure to ionizing radiation. To mainrain a high image quality over a period of years, we have recently tested a modern Peltier-cooled CCD. Our main clinical interest lies in imaging of implanted radio-opaque markers at low exposures.
Materials and methods: The cooled 2/3" CCD consists of 1280x1024 pixels and has a quantum efficiency of 48% at 550 nm. The Peltier-cooling reduces the CCD temperature to 250 K. The camera was installed in a TheraView EPID assembly, and coupled to the Theraview NT software using digital video signal transfer through optical fiber. The quality of the EPID-images was assessed using the QC-3 phantom, and by imaging various kinds of radio-opaque markers in the Alderson phantom (gold: diametar (D) =0.9, 1.4 and 1.6 mm, length (L)=5 mm; platinum: D=0.9 ram, L=3 mm; titanium: D=0.8, L=8 mm and D=1.0, L=12 mm). Results: The resolution of the system, expressed by the f50-value, is 0.36 Ip/mm. After months of clinical operation the CCD-camera did not yet exhibit hot pixels, and the pixel value distribution of the dark current remained very narrow. Consequently, no degradation of image quality has yet been observed. This result contrasts strongly to our observations for non-cooled CCDs at beam energies > 10 MV. For 6 and 23 MV beams (AP and lateral), even the smallest markers could be clearly distinguished with an exposure of only 2-4 MU. Discussion and Conclusions: The cooled CCD-EPID allows for a high SNR at low exposures, good spatial resolution and a strongly improved lifetime. Small markers are well visible in images of only a few MU. Therefore, the system is suited for on-line setup corrections based on implanted marker displacements and we expect that a-Si EPIDs will not add value for this application. Presently, we are coupling an a-Si device to our CCD-EPID software to allow for a direct performance comparison focussed on clinical EPID use. 39
The use of electronic portal imaging to verify the patient setup and delivery of intensity modulated radiotherapy A.L. Fieldincl 1, P.M. Evans 1, C.H. Clark2 1Institute of Cancer Research, Joint Department of Physics, Sutton, United Kingdom 2Royal Marsden Hospital, Joint Department of Physics, Fulham Road, United Kingdom The precise shape of the 3-D dose distributions created by intensity modulated radiotherapy (IMRT) means that the verification of patient position and set-up is crucial to the outcome of the treatment. In this paper we investigate if recent developments in electronic portal imaging technology can be used to obtain patient position information during the delivery of tMRT to the thyroid of cancer patients. Electronic portal images of the intensity modulated treatment beam delivered using the dynamic multi-leaf collimator technique were acquired. The images were formed by measuring a series of frames throughout the delivery of the beams. The frames were then summed to produce an integrated portal image of the delivered beam. Methods for calibrating the integrated image were investigated with the aim of removing the intensity modulations of the beam. The simplest calibration method involved delivering the intensity modulated beam to a thickness of PMMA blocks similar to the thickness of the patient. An integrated portal image of the delivery to the blocks was acquired. The patient image was then simply divided by the calibration image. The data presented was measured using a Varian amorphous sillcon flat panel imager. The methods were first tested using a contrast phantom before images were acquired of IMRT treatments delivered to the thyroid of cancer patients at the Royal Marsden Hospital. We present results of applying the calibration method applied to images acquired during treatment of a thyroid patient. The calibration successfully removes the intensity modulations of the beam so that the skeletal anatomy of the upper thorax and neck region can be seen in the calibrated image. This new method could provide a complementary tool to existing position verification methods and it has the advantage that it is completely passive, requiring no further dose to the patient using only the treatment fields. The treatment time integrated image can also be used for dosimetric verification purposes.