ESTRO 36 Abstract Book

S885 ESTRO 36 2017 _______________________________________________________________________________________________

The number of patients eligible for inclusion was 18 (3 was rejected because of missing (2) or faint (1) clips). All patients received daily IGRT with CBCT using XVI from Elekta. CBCT’s for the first ten fractions were included resulting in 180 CBCT. Two matches were performed for each patient: First, a chest wall-match were performed, where CBCT images were registered automatically to the planning CT using a grey value translational match of the thorax wall. Secondly, a boost specific match was performed, where the surgical clips were manually registered to their position on the planning CT. Results The systematic (Σ) and random errors (σ) between the thorax wall and the clips based boost matches are seen in the table. The CTV-PTV margins were calculated based on the systematic and random errors using the Van Herk margin recipe. These numbers are found both using no re- matching and a re-matching if there were more than 5 mm differences between the two matches. In the figure the percentage of rematches needed for a given threshold of allowed difference between the thorax wall and boost match is shown in a). The boost CTV to PTV margin required to account for the difference between chest wall and boost match for a given threshold is shown in b). The grey shaded area shows the 90 percent confidence interval obtained by bootstrapping.

scanning and re-matching will be required in only 5% of the fractions. EP-1653 PolymarkTM fiducial markers migration in Prostate Image Guided Radiation Therapy using CBCT images C. Camacho 1 , I. Valduvieco 1 , J. Sáez 1 , A. Herreros 1 , J. García-Miguel 1 , E. Agustí 1 , C. Castro 1 1 Hospital Clinic i Provincial, Radiotherapy, Barcelona, Spain Purpose or Objective Polymer-based fiducial markers, FM (Polymark TM ) location was analyzed to test the idea that there is no intraseed migration within the prostate, which is fundamental for patient set-up good quality overthe entire course of radiotherapy treatment (RT). Material and Methods Six prostate cancer patients with transperineal placement of 3mm-long 1mm-diameter polymer-based fiducials underwent 3mm slice thickness plan CT scan on day 14 after markers implantation, which is consider as a safe waiting time according to the literature. All patients were managed with the same IGRT protocol: before each daily treatment, two planar KV images were acquired with the OBI 1.4 system (Varian Medical Systems) at 45º and 315º. A manual marker match between the KV images and the planning CT DRRs was performed and automatically transfered to the treatment couch position to correct the patient position in the three translational directions (rotations were not taken into account). Weekly, after patient re-position and just before session delivery, a CBCT scan is acquired, that is used to assess rectum and bladder filling (slice thickness between 1mm and 3mm) These CBCT images, as being acquired in patient corrected position, have been used to evaluate the FM locations at different times during the course of treatment. A total of 37 CBCT images have been analysed to reconstruct the FM 3D coordinates. The displacement of each FM was calculated relative to its reference position on the reference planning CT, and also shift of the middlepoint of each 3 FM set. The distance between markers in each set at the time of planning CT and during specific evaluated treatment have also been computed. Results The average marker migration observed is 0.68±0.51 mm (range between 0 – 3.90 mm). This observation seems independent of the marker position inside the prostate, but not of the spatial coordinate: the antero-posterior direction presents the largest FM average displacement. Although the average migration observed is low, there are cases among the six patients where the migration observed an specific day was greater than 2mm. This observation may be directly related to the degree of prostate desplacement caused by the influence of the rectum and bladder, and also with the posible pelvic rotation in the moment of daily RT (not corrected with the 2D DRR vs KV image comparison). Changes in distance between pairs of FM in each set have been, on average, 0.12±0.11 mm (range between 0.02 – The low average FM migration observed is expectable, according to the waiting time between marker implantation and the planning CT scan procedure. A futher investigation should be done in order to reduce this waiting time. The fact of having observed cases among all patient with displacement greater than 2 mm should be taken into account in the CTV-PTV margins: an adequate expansión of margins might compensates for this set-up uncertainty. 4.38 mm). Conclusion

Conclusion The presented protocol can reduce the required CTV to PTV margin for the boost region by re-scanning and re- matching the boost region only for patients where the two regions differ by more than a set threshold. Results are presented that can be used for selecting a threshold with the corresponding required CTV-PTV margin. If e.g. a threshold of 5 mm is used, the required CTV-PTV margin can be reduced from approximately 8mm to 5mm and re-