Abstract Book

S959

ESTRO 37

retrospectively calculated using an in-house developed software by shifting the target point within the planning CT based on the motion information and beam status recorded with the Calypso system. To assess the accuracy of the dose calculation algorithms of both TPS and PG, a PinPoint® chamber measurement was used for normalization. Results Both PG measurement and 4D dose calculation algorithm were able to reproduce the dose fall-off in the direction of dominant motion. A representative dose profile is shown in Fig.1 (a). However, the calculated dose showed a dose overestimation of up to 0.4Gy. Analysis of the motion trajectory, however, showed small deviations between the motion trajectory derived from the Calypso TM -system and that acquired directly from the motion robot. Deviations were especially large for regions of maximum elongation, where the container position was underestimated with a mean deviation of −1.07 ± 0.38mm. To analyze this effect in more detail, the 4D dose calculation was repeated using the robot-based rather than the Calypso TM -system-based container trajectory. As a result, the shape of the profiles in the dose gradient showed a better agreement with the measured data (see Figure 1 (b)) with a high 3D-gamma passing rate of 91.8%. Conclusion The 4D dose calculation algorithm in combination with the PG is very accurate to calculate dynamic target irradiation without the application of motion- compensation techniques. It should be emphasized that the PG was able to detect small deviations being caused by the Calypso TM . Therefore, the PG can be considered as a valuable and robust tool for the purpose of workflow verification in adaptive radiation therapy. EP-1787 Use of two in vivo monitoring devices in the breast irradiation: an anthropomorphic phantom study C. Arilli 1 , Y. Wandael 2 , M. Casati 1 , L. Marrazzo 1 , C. Galeotti 2 , I. Meattini 2 , P. Bonomo 2 , G. Simontacchi 2 , S. Pallotta 3 , C. Talamonti 3 1 Azienda Ospedaliera Universitaria Careggi, Medical Physics Unit, Florence, Italy 2 Azienda Ospedalieria Universitaria Careggi, Radiotherapy Unit, Florence, Italy 3 University of Florence, Biomedical Experimental and Clinical Science "Mario Serio", Florence, Italy Purpose or Objective Two systems for in vivo dosimetry, the IQM detector (Integral Quality Monitoring, iRT Systems GmbH, Koblenz, Germany) and the SoftDiso software (Best Medical Italy Srl), were tested to evaluate their sensitivity in detecting some delivery errors which can occur in standard 3DCRT external breast irradiation. Material and Methods The irradiations were performed with a Precise Elekta linac (Crawley, UK) equipped with silicon amorphous portal imaging (EPID). IQM detector is designed to be mounted below the MLC during treatment irradiation. The IQM signal (dependent of the radiation fluence) is compared with a reference signal. SoftDiso is a software which reconstructs the dose distribution at the isocenter

plan from EPID images acquired during treatment. The measured dose (Diso) is compared with the calculated dose (DTPS) by R parameter defined as Diso/DTPS. A suitable phantom was made to simulate the female anatomy by modifying the ALDERSON RANDO with silicon breast prosthesis (figure 1a). 3DCRT plan of left breast was created with Pinnacle3 Professional TPS (Philips Medical Systems, Madison, WI) (figure 1b) and modified to simulate common errors: MU were perturbed to mimic an output error (adding 2-3-5-10MU); jaws positions were perturbed mimicking a calibration jaw bank error (opening and closing one jaw of 2-3-5-7mm). Both devices were employed during the irradiation of all plans. The deviations of IQM signal and R parameter from the original values were evaluated.

Results The short term reproducibility of R parameter of SoftDiso system was checked by repeating twenty times the original plan with the phantom located on the couch during irradiation. The reproducibility on terms of standard deviation/R mean value was resulted 0.6%. Same tests were performed using IQM [R&O, Volume 119,S215-S216]. Figure 2 shows the deviations of IQM signal and R value from the original for all delivery errors induced. The IQM signal and the R values linearly increase with the increment of MU (R 2 =0.99 for IQM and R 2 = 0.97 for SoftDiso) (Fig2a). The IQM signal is linearly correlated with the position of jaws too (R 2 =0.97 for jaws open and R 2 = 0.93 for jaws close), while R parameter is less sensitive to jaws positioning variations (Fig2b-2c).

Conclusion Towards safer radiotherapy, centres should have protocols in place for in vivo dosimetry. The two systems exhibit good performances for in vivo monitoring of the external beams breast irradiation. The IQM device is able to detect small delivery errors in terms of output and field size of beams. The SoftDiso software shows a good capacity to detect small output errors while is less sensitive to jaws positioning variations. On the other hand SoftDiso has the potential of detecting set-up errors, which cannot be detected with IQM. The concurrent use of the two tested systems allow for a check of the correct functioning of all components in the radiotherapy chain, including the treatment planning, the delivery system and the patient positioning and thus play an important role in meeting the needs of modern and upcoming radiotherapy QA.