R. de Jong

OC-0075: Simple spatula improves the geometrical accuracy of a cranial mask for brain tumor radiotherapy

Purpose or Objective The demands on accuracy in radiotherapy have increased especially in stereotactic treatments using IMRT or VMAT techniques where margins to the PTV are as small as 1 or 2 mm, but also for non-stereotactic treatments. Commercial systems providing better geometrical accuracy in patient positioning than conventional standard three-point masks, entail high costs. Although translations can be corrected with a couch shift, rotations cannot be corrected with a standard treatment couch and need to be prevented as much as possible. Therefore we aim to investigate whether the use of a dental fixation created with an inexpensive and simple wooden spatula, will improve accuracy in patient positioning. Material and Methods In 40 patients receiving non-stereotactic cranial radiotherapy, 144 conebeam CTs (CBCT) were acquired prior to treatment. Twenty patients had a standard threepoint thermoplastic mask with a standard base (MacroMedics®, Waddinxveen, The Netherlands); next 20 patients had an identical mask and base, but with the addition of a dental fixation moulded by a wooden spatula, to create an extra point of fixation between the teeth rows. Patients were asked to bite gently on the wooden spatula during moulding of the mask to create an indentation in the mask for dental fixation. After cooling and hardening of the mask, the wooden spatula is removed. During the acquisition of the planning CT and all treatment fractions patients are instructed to bite gently on the indentation. All CBCTs were registered on bony anatomy of the skull. For patients with an online correction protocol, all data were included. For patients with extended NAL correction protocol, only the data of the first ‘NAL’-phase were included. Thereby, the position inaccuracy was calculated on position errors before a position correction was applied. Individual systematic (Ʃ) errors were calculated and analyzed with Levene test. Individual random errors (σ) were calculated and analyzed with the Mann-Whitney test. Results The table summarizes the group setup errors for both fixation systems. Most errors are smaller when using the three-point mask with dental fixation created with a simple wooden spatula compared to the three-point mask alone. Geometrical accuracy shows significant improvement in the systematic and random error for the rotation over the X axis and the random error for rotation over the Y axis. Conclusion Adding a dental fixation point to a standard three-point cranial mask by a simple wooden specula improves geometrical accuracy, particularly by reducing rotational errors. This may be of clinical importance, since rotational errors cannot be corrected by a standard treatment couch. Although the absolute errors are already small for the standard three-point mask, but given the small effort and the low additional costs of a simple wooden spatula, we decided to accept the mask with dental fixation as our standard for non-stereotactic brain tumor radiotherapy.

EP-1658: The need for quality assurance of the image guidance process in radiotherapy

Purpose/Objective: This work evaluated the inter-observer variability in CBCT registration for prostate patients treated with IMRT (Intensity Modulated Radiation Therapy). We also demonstrated the importance of daily CBCT registration for this treatment. Materials and Methods: In the first study VMAT plans were produced for 2 prostate cancer patients and a voluntary shift of the isocenter was introduced in the anterior/posterior direction. A total of 22 different VMAT plans were evaluated with 13 plans for the first patient and 9 plans for the second. In the second study, 12 technologists each registered 22 CBCT for 22 different prostate patients. Medical doctors have done the same registrations that was considered as the Gold Standard reference. Registration was done using bony anatomy and adjustments made on by the user on evaluation of the rectum and prostate. Shifts in all three axis were documented for a total of 286 CBCT. Two statistical methods were used to analyse the results. The first was the 95 percentile to calculate the minimum threshold under which the users found similar values. The second was the ANOVA test, followed by the Post-Hoc/Bonferroni test. These tests were used to find differences in inter-observer registration variability and determine whether any individual users performed registrations which differed significantly from those of the other users. Results: The dosimetric study showed that a shift of 5 mm in the posterior direction was enough to deliver a higher dose than acceptable to the rectum in both cases.A different threshold was found by shifting anteriorly, ranging from 5 mm for patient B to 12 mm for patient A. On the other hand, the statistical analysis of the registration study showed that using the 95 percentile, threshold values were demonstrated of 2.1, 3.5 and 7.3 mm in the left/right, target/gun and anterior/posterior respectively. The Anova test showed a low p-value in the target/gun axis but using the Post-Hoc analysis there were no significant differences between the technologists and the medical doctors. Conclusions: This study showed the importance of a daily CBCT/CT registration in prostate radiation treatment. The different studies also showed that partial delegation of the prostate registration fto the technologists is feasible under some security thresholds. However, regular training and evaluation should be done by the medical doctors and physicists. EP-1658 The need for quality assurance of the image guidance process in radiotherapy M. Frank, M. Kamphuis, R. De Jong, A. Bel, M. Hulshof, N. Van Wieringen Academic Medical Center (AMC)/ University of Amsterdam, Radiotherapie, Amsterdam, The Netherlands

Clinical experience with IGRT using cone-beam CT

Cone-beam computed tomography integrated with a linear accelerator is one of the more promising ways of bringing image guidance into radiotherapy. We will describe the clinical introduction of such a system in our hospital and a number of image guided protocols we have developed over the past three years.