doi:10.1016/j.meddos.2008.08.005 THE PERFORMANCE OF MULTILEAF COLLIMATORS EVALUATED BY THE STRIPE TEST MARIA SASTRE-PADRO, M.SC., CHRISTOFFER LERVÅG, M.SC., KARSTEN EILERTSEN, M.SC., and EIRIK MALINEN,PH.D. Department of Medical Physics, Division of Cancer Medicine and Radiotherapy, Rikshospitalet University Hospital, Oslo, Norway (Received 3 June 2008; accepted 21 August 2008) Abstract—The performance of 3 multileaf collimator (MLC) systems (Varian Medical Systems, Elekta, and Siemens Medical Solutions) mounted on 7 different radiotherapy linear accelerators was investigated by a stripe test. The stripe test consisted of 8 adjacent multileaf segments of 2.5  40 cm 2 , enclosed by all leaf pairs. With 6-MV photons, the segments were used to irradiate Agfa CR films. The optical density profile of the irradiated film in the travel direction of the MLC was used to estimate the short- and long-term leaf positioning reproducibility. The short-term reproducibility was found by analyzing 6 consecutive stripe tests. The long-term reproducibility was obtained by performing 3 to 5 stripe tests over 2 months. The short-term reproducibility was mainly within 0.3 mm for all systems. For the long-term reproducibility, the Varian and Elekta MLCs were within 0.4 to 0.5 mm, while the Siemens MLC showed a wider distribution, with values up to 1 mm for some leaf pairs. The inferior long-term reproducibility of the Siemens MLCs was mainly due to a decrease of the segment size with time. In conclusion, the stripe test is a useful method for evaluating MLC performance. Furthermore, the long-term reproducibility varied among the MLC systems investigated. © 2009 American Association of Medical Dosimetrists. Key Words: MLC, Stripe test, IMRT. INTRODUCTION Intensity modulated radiotherapy (IMRT) is usually fa- cilitated by multileaf collimators (MLCs). For segmented IMRT, the treatment beams comprise several MLC seg- ments, each treating a portion of the target. The mechan- ical performance of the MLC is particularly important for IMRT, as leaf positioning errors may affect the target dose distribution and thus the treatment outcome. 1 The performance of the MLC is related to its design. Many different MLC systems have previously been characterized. 2–7 However, as a standard MLC commissioning method is not available, a comparison between different systems is not straightforward. A dosimetric comparison between 3 different types of MLCs was done by Huq et al., 8 and showed advan- tages and disadvantages of each system according to their particular design features. To our knowledge, no comparison has been made between the performance of different MLC systems. A stripe test consists of several adjacent MLC seg- ments that are used to irradiate a film. The joints between the adjacent segments, as appearing on the film, are sensitive to leaf positioning errors. For a given leaf pair, the dose variations at the joints originate from the addi- tion of the 2 segment profiles. Features that affect the profiles are, among others, the position of the MLC in the head of the linear accelerator and whether the leaf tip is rounded or flat. The stripe test is appropriate to quickly observe positioning errors of individual leaves and de- cide whether or not a new MLC calibration should be performed. Some authors have previously used the stripe test to visually check the leaf positioning accuracy. 9,10 Furthermore, a quantitative stripe test for estimating in- dividual leaf positioning with subsequent MLC calibra- tion has been developed by Sastre-Padro et al. 11 In the current study, such a test is used to compare the mechan- ical performance of Varian, Elekta, and Siemens MLCs. METHODS AND MATERIALS The MLCs vendors included in the study were Var- ian Medical Systems (Palo Alto, CA), Elekta (Crawley, UK), and Siemens Medical Solutions (Erlangen, Germany). All MLCs included have 10-mm pitch at isocenter distance. The Varian MLCs investigated have 40 leaf pairs with rounded leaf tips that move in a linear path that remains on a plane perpendicular to the central axis of the beam. The distance from the leaf tip to the isocenter is 46.5 cm. The Varian MLCs were a Mark-2 model (Varian 1) and a Millenium model (Varian 2). The Elekta MLCs have 40 leaf pairs with rounded leaf tips that move like the Varian MLC. In this case, the MLC replaces the upper jaws, and a backup collimator is placed underneath to reduce the trans- Preliminary results of this paper were presented as a poster at ESTRO 26, Barcelona, Spain, September 2007. Reprint requests to: Maria Sastre-Padro, M.Sc., Department of Medical Physics, Division of Cancer Medicine and Radiotherapy, Rikshospitalet University Hospital, Montebello, 0310 Oslo, Norway. E-mail: Maria.Sastre@radiumhospitalet.no Medical Dosimetry, Vol. 34, No. 3, pp. 202-206, 2009 Copyright © 2009 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/09/$–see front matter 202