Geometric accuracy evaluation of the new VERO stereotactic body radiation therapy system Tom Depuydt*, Olivier C. L. Haas**, Dirk Verellen*, Stephan Erbel***, Mark De Ridder*, Guy Storme* * Radiotherapy department, University Hospital UZ Brussel, Belgium (Tel: +32 2 474 90 89, email: tom.depuydt@uzbrussel.be) **Control Theory and Applications Centre, Coventry University, UK (Tel: +44 24 7688 7658, o.haas@coventry.ac.uk) *** BrainLAB AG, FeldKirchen,,DE Abstract: Real-time tracking of moving tumors is one of today’s challenges in radiation therapy. This work investigates the tracking performance of VERO, a novel treatment device with gimbaled linear accelerator, especially designed for four dimensional image guided radiotherapy. It is found that the significant impact of organ motion on dose distribution can be overcome by combining a polynomial predictor with a prediction horizon of 50ms with the VERO tracking system. Tracking errors can be reduced from 1.7mm to 0.6mm for realistic patient signals as well as sine wave from 5 to 30 bpm. Keywords: radiotherapy, tracking, measurements, position accuracy, medical robot 1. INTRODUCTION In radiation therapy, intra-fraction motion results in significant geometric and dosimetric uncertainties in the therapeutic dose delivery treating thoracic and abdominal tumors. To accommodate these issues and to assure adequate dosimetric coverage of the tumor, geometric safety margins can be introduced which incorporate the entire motion. However, the use of large margins, where healthy tissues are present, prevents dose escalation as the dose to the healthy tissue would be a limiting factor. Tumour dose escalation has been shown to be beneficial in terms of outcome (Wulf et al, 2005; McGarry et al, 2005; Komaki et al, 2005) when it can be combined with a decrease in excessive irradiation of surrounding normal tissue and critical structures to minimize complications (Kwa et al, 1998, Hernando et al, 2001). Beam modulation devices can help deliver complex radiation fields that can accurately conform to the PTV whilst at the same time sparing critical structures. The drawback of such treatments known as IMRT and to a lesser extent conformal radiotherapy is the presence of dose distribution with large gradient between the organs at risk and the tumor. If the tumor moves it could receive a significantly lower dose than planned and the organ at risk a significantly higher dose. Such event could negate the expected benefit of dose shaping and modulation. This paper focuses on one method to address organ motion referred to as real time tumor tracking (RTTT). RTTT is believed to be able to bring together the goals of dose escalation and surrounding tissue sparing for moving tumors. Most of the current RTTT solutions presented exploit existing radiation therapy equipment in new ways to adapt the tumor motion to the therapeutic beam or adapt the beam to the tumor motion. A first approach is respiratory gating (Verellen et al, 2006) where the therapeutic beam is switched on only when the moving tumor is adequately covered and switched off otherwise. A drawback here is that the duty cycle usually is no more than 20 to 30 % which leads to very long treatment times. In an attempt to increase the duty cycle approaches like active breathing control (ABC) Wong et al, 1999) or breath-holding techniques Mah et al, 2000) have been used. However, patients affected by tumors in lungs and abdomen, where intra-fraction motion is most significant, may experience difficulties even with normal breathing, hence both breath-holding and ABC have limited applicability and may cause discomfort to the patients. The most technically challenging method with potentially the best result is to accommodate tumor motion during irradiation with a real-time tumor tracking radiation delivery system. The principle is that the tumor should stay in the same relative position with respect to the beam. This can be accomplished by continuous adaptation of the patient/tumor position using the patient support system (PSS) with respect to a static beam or by real-time adaptation of the treatment beam itself to the tumor position. The technical feasibility of using a PSS to compensate for measurable tumour motion with respect to the treatment beam has been demonstrated (DeSouza et al , 2006, Skworcow et al, 2007, Wilbert J. et al, 2008). Whilst some studies seems to indicate that moving patient does not lead to any discomfort and can even be relaxing (Wilbert, 2008), the influence of counteracting part or all the motion on patient/tumor position behavior during this continuous repositioning requires further investigation. Indeed compensation of the motion in one direction may lead to a motion in another direction. Other approaches with a fixed patient position were presented where the dynamic behavior of the PSS were taken over by dynamic operation of the multi-leaf collimator (DMLC) (Keall et al, 2006, Sawant