253 poster Effects of geometric distortion in pelvic MRI on radiotherapy treatment planning of prostate cancer patients

$94

Posters

ues than distributions of values. Also, the relative importance of the modelling parameters determining the. slope of the TCP curve is different if distributions of values rather than single values are used. Therefore, distributions of values are needed for realistic TCP modelling. 250

poster

Optimisation of 3D conformal radiotherapy for prostate cancer: a radiobiological analysis P. Diez 1, P.J. Hoskin 2, P.J. Bownes 1, P.J. Ostler 3, E.G.A. Aird t 1Mount Vernon Cancer Centre, Clinical Physics, Middlesex, United Kingdom 2Mount Vernon Cancer Centre, Marie Curie Research Wing, Middlesex, United Kingdom 3Mount Vernon Cancer Centre, Oncology, Middlesex, United Kingdom External beam radiotherapy (EBRT) is commonly used for the management of adenocarcinoma of the prostate. A typical treatment, such as that proposed in the RT-01 UK prostate trial protocol, would involve delivering 64 Gy in 32 fractions to the PTV, plus a 10 Gy boost in 5 fractions to the GTV, all in 2 Gy per fraction. A similar dose escalation may also be achieved using high dose-rate brachytherapy (HDR-BT). This study aims to determine whether HDR-BT is indeed an optimisation tool in radiotherapy of prostate cancer by comparing predicted local control rates and rectal toxicity for both dose-escalating techniques. Ten random intermediate4o-high risk patients have been planned for different treatment schedules: RT-01 (high-dose arm) [as described above]; EBRT alone [55 Gy in 20f] and EBRT+HDR-BT [35.75 Gy in 13f + 17 Gy in 2f] as delivered at MVCC; HDRBT alone according to protocols used at the William Beaumont, Michigan, [38 Gy in 4f] and in Osaka, Japan [48Gy in 8f or 54Gy in 9f]. Local control is defined by predicted tumour control probabilities (TCP) and normal tissue complication probabilities are calculated using rectum as the critical normal tissue. Different cd~ ratios for prostate cancer have been considered. Boosting of the target volume using HDR-BT has significant advantages for tumour control if the ed[3 ratio for prostate cancer is low, comparable to that of late-reacting tissue, without compromising rectal morbidity. HDR-BT as monotherapy does not show a clear benefit for TCP over the combined modality but both schedules using HDR are superior to EBRT alone if cd~ is low. An assessment of the impact of dose optimisation with external beam using IMRT is under way. 251

poster

Treatment precision and treatment effect in conformal radiotherapy of urinary bladder cancer - preliminary results L.P. Muren 1,2, R. Ekerold2, R. Smaaland2, O. Dahl 1 1University of Bergen, Section of Oncology, Institute of Medicine, Medical Faculty, Bergen, Norway 2Haukeland University Hospital, Dept of Oncology and Medical Physics, Bergen, Norway Purpose: Geometrical uncertainties may compromise the treatment precision of conformal radiotherapy (CRT) and thereby cause local treatment failures. This study aimed to correlate treatment precision to treatment effect in radical CRT of urinary bladder cancer (UBC). Materials and methods: Treatment precision and treatment effect data for 20 UBC patients were analysed. Patients were prescribed a dose of 60-64 Gy, and treated with a four-field CRT technique using CTV-ITV margins of 1.5-2.0 cm. Target coverage measures were derived from treatment planning data supplemented with the actual treatment Iocalisation of the bladder from co-registration of weekly repeat CT scans. The outcome of the treatment in terms of local control was quantified from the routine follow-up cystoscopies, commencing 3 months post-treatment. Results: In 105 of the 128 repeat scans (82%), at least 99% of the CTV received at least 95% of the prescribed dose (CTV_95% > 99%). 15 of the 20 patients (75%) had adequate CTV coverage, defined as a CTV_95% > 99% in all except one of the repeat scans. Local control was achieved in 11 of the 17 patients (65%) that could be evaluated (12 months median followup). In the 11 cases with local control, the CTV_95% > 99% in 64 of 73 repeat scans (88%), compared to in 31 of 39 repeat scans (79%) in the six cases with local failure (p=0.38). Nine of the 11 cases with local control had adequate CTV coverage (81%), compared to four of the six patients with local failure (67%) (p=0.92). There were also no statistically significant differences between patients with local control and patients with local failure in the average CTV coverage measures, including minimum CTV dose and the average CTV fraction receiving less than 90% or 95% of the dose. Con¢lusi0n$: Local control was achieved in 65% of patients with UBC prescribed 60-64 Gy to the bladder with wide margins. Patients with local con-

trol tended to have slightly higher CTV coverage rates than patients with local failures, but the differences were not of statistical significance. Future studies will investigate whether the inclusion of information about the exact tumour Iocalisation and patient positioning data derived from repeat electronic portal imaging will improve the correlation between target coverage and local control.

I M A G I N G FOR RT 252

poster

A novel, rigorous approach to dosimetric characterisation of CCD-based electronic portal imaging devices J.C.J. de Beet E.M. Franken, B.J.M. Heijmen Erasmus MC, Daniel den Hoed Oncology Center, Rotterdam, The Netherlands Introduction: The excellent long term response reproducibility of CCD-camera based electronic portal imaging devices (CEPIDs) renders them highly suited for accurate portal dosimetry. Five of our Theraview-NT CEPIDs are equipped with cooled (-20C) CCDs. By suppressing the deleterious effect of radiation damage, cooling has boosted the CCD-lifetime by years and improved signal-to-noise ratios (SNR) markedly over previous generation CEPIDs. However, precise CEPID-based dosimetry requires knowledge of the low-amplitude but spatially extended tails of the point-spread functions (PSFs), mainly stemming from optical cross-talk between the mirror and the fluorescent screen (FS). We present a method to measure these PSFs directly, without making any assumptions on their properties. Methods and Materials: Position dependent, two-dimensional PSFs were extracted from CEPID images (EPI) as follows. Using the linac collimators, EPI were obtained for a series of small fields (side < 3 cm) positioned on a grid covering the FS. For high SNR, the CEPID software allowed for videoframe summation up to large exposures. The small fields thus imaged effectively constitute input pencil-beams. A transmission EPI, obtained with collimators closed, was subtracted. The resultant EPI directly measures the PSF for a given field position. This procedure was executed for a low and an intermediate elbow CEPID (distance between FS and mirror of 5 respectively 13 cm). In addition, the optical propagation in the CEPID was modelled to predict the PSFs. Results: For both elbows, PSF-tails could be confidently measured using 2000 MU per EPI, even for tail magnitudes of 0.1% of the PSF-peak. The PSFs were asymmetrical, dependent on position on the FS, and were quite different for the low and intermediate elbow. These features were in good agreement with the model predictions. Hence, a small number of PSF measurements (-9) in combination with our model was sufficient for a complete PSF characterisation of both CEPID types. Conclusions: The high SNR offered by present CEPIDs, together with a basic optical cross-talk model, allowed for a complete experimental determination of PSFs using a few calibration measurements per CEPID. We are now implementing robust and fast deconvolution schemes to translate EPI into portal dose images for position dependent PSFs. Hence, a general framework for precise portal dosimetry with CEPIDs, even in case of highly asymmetrical IMRT fields, is established. 253

poster

Effects of geometric distortion in pelvic MRI on radiotherapy treatment planning of prostate cancer patients B. Petersch 1, J. Bogner 1, A. Fransson 2, T. Lorang 3, R. P~tter 1 1University of Vienna, Department of Radiotherapy and Radiobiology, Vienna, Austria 2Karofinska Hospital, Department of Hospital Physics, Stockholm, Sweden 3University of Vienna, Institut for Medizinische Computerwissenschaften, Vienna, Austria Purpose: To evaluate the impact of two different methods of geometric distortion correction of MR images from a Siemens Magnetom Open Viva 0,2T resistive MR unit on the process of external beam radiotherapy treatment planning (RTTP) for prostate cancer. Material and methods: A method for correction of system related AND object induced distortions and one for correction of purely system related distortions have been evaluated. The latter used information extracted from MR images of a 3D phantom specifically designed for geometric distortion evaluation. An active shim procedure was performed prior to all phantom and patient scans For each of five patients five standard treatment plans were compared using uncorrected and corrected MR images alone (density=water) and CT images alone. Finally internal anatomical landmarks were

Posters

S95

used for image registration between MR images (corrected and uncorrected) and CT images to evaluate the impact of distortion correction on the image registration process. Summary: Maximum distortions of 28mm (mean 2,2mm) were found within the FOV in frequency encode direction. Maximum distortions could be reduced by a factor of two (mean: factor four) by our phantom measurement based technique. Distortion patterns were found to be stable and reproducible over several weeks with this MR unit. For 4/5 patients, relative doses at the normalization point as calculated on the distortion corrected MR images only (all tissues taken water equivalent) were all within 1% of the corresponding value from the standard CT-based plan (actual Hounsfield units). The largest differences in isocentric dose found in one case were 3,1% MR uncorrected vs.CT and 2,6% MR corrected vs. CT. Typical sites of internal anatomical landmarks chosen for image registration show distortions up to 3ram. Conclusions: Object induced distortions are negligible at such low field strengths compared to system related distortions. Treatment plans for prostate cancer do not seem to differ significantly from "standard" plans calculated on CT images when calculated on distortion corrected MR images, even if all tissues are assigned the electron density of water. Distortion correction of MR images can theoretically improve the starting point for image registration of MR and CT images. 254

poster

Correlation between pelvic patient movement derived from electronic portal images and external focus-skin distance measurements D.G. Lewis 1, K. Ryan2 I Velindre Hospital, Medical Physics, Cardiff, United Kingdom 2Velindre Hospital, Radiotherapy, Cardiff, United Kingdom Many radiotherapy centres traditionally perform machine monitor unit (MU) adjustment to treatment fields based on external focus-skin-distance (FSD) measurements, averaged over the first few treatment fractions. This provides a simple method of adapting doses delivered by individual fields to changes in external patient contour (and in some cases position) between CT scanning, simulation and treatment. The advent of electronic portal imaging devices (EPIDs) has allowed more information on patient position, based on bony landmarks, to be obtained routinely. The purpose of this study was to determine whether or how these two different sets of information can be reconciled when developing patient positioning correction protocols locally. We have therefore investigated the degree of correlation that exists between these two sets of data, based on measurements performed on a group of 39 patients undergoing pelvic radiotherapy. Anterior field images were acquired using an EPID and field placement error (FPE) relative to a reference simulator image was determined in each case using the image matching software available with the EPID. Average FPEs were calculated from images obtained over the first three treatment fractions. FSD measurements were also averaged over the first three fractions for the right and left lateral treatment fields to derive information on displacement of external contours relative to those used in the CT-based treatment plan. The FSD-based displacement data were graphed against the EPID-derived FPEs. The two datasets were found to be only weakly correlated (correlation coefficient = 0.3 approximately) indicating that if patient position correction protocols based on FPEs derived from EPIDs are to be introduced, careful consideration of the validity and value of the more traditional FSDbased correction methods will need to be made. 255

poster

Evaluation of the response of radiotherapy to squamous cell carcinoma of the head and neck using 18FDG-PET S.S. Shin 1, S. Lee 1, J.S. Kim2, J.S. Ryu2, E.K. Choi 1, J.H. Kim 1, S.D. Ahn 1, B.Y. Yi 1, S.Y. Kim 3, S.Y. Nam3 l Asan Medical Center, Radiation Oncology, Seoul, Korea 2Asan Medical Center, Nuclear Medicine, Seoul, Korea 3Asan Medical Center, Otorhinolaryngology, Seoul, Korea Purpose: To evaluate the efficacy of positron emission tomography with 2[F-18]fluoro-2-deoxy-D-glucose in discrimination of response in the squamous cell carcinoma of the head neck patients who treated with definitive radiotherapy. Methods and Materials: Twenty4our patients who underwent FDG-PET scan before and after radiotherapy for no disseminated head and neck carcinoma at the Asan Medical Center between August 2001 and September 2002 were evaluate by prospective analysis. First FDG-PET scan per-

formed before radiotherapy within 1 month, and second FDG-PET scan performed 1 month after radiotherapy. FDG-PET images were analyzed by standard uptake value (SUV). Follow-up period was more than 6 months. Results_: The pretreatment SUV was 3.4-14.0 (median: 6.5) and posttreatment SUV was ground level-6.7 (median:2.3). The overall sensitivity and specicity of FDG-PET to evaluate residual tumors in the squamous cell carcinoma of the head and neck were 94% and 94%. Conclusions: FDG-PET is effective in evaluation of radiation response in the squamous cell carcinoma of the head and neck. We think that the timing of one month after finished radiotherapy FDG-PET scan was not too fast to evaluation of radiation response. 256

poster

Quality assurance of imaging for stereotactic treatments J. Novofny, J.J. Novotny Medical Physics Department, Na Homolce Hospital, Prague, Czech Republic The stereotactic treatment planning for Leksell gamma knife or linear accelerator has two main aspects: the first one reflects geometrical accuracy of target localization; the second one takes into account a dosimetric accuracy of dose distribution and absorbed dose calculation. High precision of dose delivery to target volume can be assured if both aspects are carefully controlled. The geometric accuracy depends on target imaging and precision of stereotactic instruments, which is about 1 mm or less. Computerized tomography, magnetic resonance or digital angiography (DSA) is used for stereotactic imaging, which is performed with localization frame. All imaging methods and sequences must be checked for distortions in order to assure a good quality of localization. A special set of phantoms has been prepared and will be presented. AI! imaging modes and sequences used for investigations were checked accordingly using comparison of estimated coordinates from TPS with real coordinates obtained from mechanical readings. CT images is within experimental error free of distortions, but for MRI and DSA distortions are noticeable depending on the type of scanner and sequences used. Some examples of results obtained for common sequences are documented in tab I. Table I: distortions observed during stereotactic imaging Imaging/sequence Equipment Mean deviation Meandeviation (ram) (ram) axial/lateral coronal/frontal DSA Toshiba 1.2 1.3 T1 3D MR 1TExpert 0.3 0.9 T1 3D MR 1.5 T Symphony 0.3 1.3 T2 MR 1 T Expert 0.6 0.5 T2 MR 1.5 T Symphony 1.0 1.3 Flow MR 1 T Expert 0.2 0.9 Flow MR 1.5 T Symphony 0.7 1.0 Maximum geometrical precision is achieved for 1T scanner and common scanning sequences (see Tab. I). Sophisticated sequences performed on the 1.5 T scanncer produce errors which are larger than those given in Tab I. These may cause problems during some precise stereotactic applications (electrode implantation of stimulator, coagulation, and LKG treatments). The distortions for different sequences will be discussed in detail and possible consequences for treatment planning outlined. 257

poster

Synchronizing portal images and A/D measurements B.A. Brand, J.-J. Sonke, M.B. van Herk The Netherlands Cancer Institute, Radiotherapy, Amsterdam, The Netherlands Introduction: When recording both portal images and external signals (MU count, diode, signal indicating respiration, heartbeat or other patient motion), it is important to synchronize image acquisition and external signal acquisition. Materials and methods: The equipment consists of an amorphous silicon flat panel x-ray detector (Heimann RID1680 AF5/Elekta iView GT) and an 8 channel AID converter which also has simultaneous digital input channel (Data Translation DT310). The detector's frame grabber (Eltec) and the A/D converter are mounted in the same computer running windows NT 4. The A/D converter software allows the addition of date- and time stamps to each sample. The frame grabber software, however, only has a callback to add a date- and time stamp some time after an image is acquired. It is undocumented exactly when this callback occurs. The frame grabber has a separate digital output which gives end-of-line and end-of-frame signals. As a benchmark test to find out which image corresponds to which part of the recorded A/D or digital input signal, images of an electromagnet with a