ESTRO 36 Abstract Book

S906 ESTRO 36 _______________________________________________________________________________________________

1 Consorcio Hospitalario Provincial de Castellón, Oncología Radioterápica, Castellón, Spain 2 Consorcio Hospitalario Provincial de Castellón, Radiofísica y protección radiológica, Castellón, Spain Purpose or Objective To evaluate if planned doses for prostate and rectum are equal to the doses which are actually delivered and to determine adaptation points for the accumulated dose. Material and Methods Twenty four patients with intermediate and high-risk prostate cancer who were going to be treated with image guided radiotherapy were enrolled. A plan-CT (pCT) and nine treatment kilovoltage conebeam-CT (kvCBCT) scans were acquired prospectively during the first three weeks of a prostate lGRT treatment (a total of 240 CTs). A rectal emptying preparation and a full bladder protocol were used. For each patient, a deformable image registration (DIR) from the pCT to each of the nine kvCBCT was performed with RayStation treatment planning system. All registers were revised and recontoured by a Radiation Oncologist, establishing regions of interest (ROIs) for a second DIR with control of such ROIs. For every patient, a hypofractionated VMAT schedule (15 x 3.82 Gy) was planned and correlated with their kvCBCT images, being able to determine the accumulated and total doses that would have been actually delivered. Since the pCT day, a nutritional evaluation control with anthropometric and biochemical parameters was performed for each of the 24 patients.

generate a library of plans for bladder treatments with a combined target of the total bladder and the GTV. Material and Methods Two CT scans were acquired and registered (empty/full bladder). The bladder CTVs and GTVs were delineated on both CTs. An in-house developed script was developed to calculate intermediate CTVs and GTVs based on the empty and full bladder delineations. The script, which utilizes a Robust Point Matching (RPM) algorithm (Osorio, 2012), yields a deformation vector field that can transform the target structure to the reference structure. The algorithm can be tuned with the following parameters: stiffness, density of points, number of iterations and the final 'temperature”. To create intermediates, the deformation can be applied partially, e.g., to create a structure in the middle of the two input structures, a 50% deformation would be applied. Dividing the maximum spacing required between consecutive intermediate plans by the maximum distance between reference and target structure, will give the excitation percentages required to get to equidistant intermediate structures. Bladder CTV and GTV need to be handled by separate RPM processes because the required parameters are very different due to large discrepancy of deformation and size. The number of plans is set by the maximum distance between full and empty CTVs. Therefore first the intermediate structures for CTV are created and then the same excitation percentages are applied to GTV. Results Figure 1 shows an example of a generated library of plans for CTVs and GTVs.To evaluate the results we create a structure with 100% deformation, which should coincide with the target structure. Using the default stiffness parameter for 10 patients (1000 for CTV and 250 for GTV) we found a success rate of 60%. By tuning the stiffness parameter, intermediate structures were created successfully for the remaining cases. On average it takes 3.50 and 2 minutes for the CTVs and GTVs to be created, respectively.

Results A significant difference between planned and delivered D98 CTV 57 (p=0.026) and D2 CTV 57 (p=0.005) was observed; however, the average D98 CTV 57 delivered was higher than the prescription dose. Despite not having observed a significant difference in V36.5 of the planned and delivered to the rectum, the delivered doses to 50% of the rectum exceeded the planned constraints in 37.5% of the patients. A significant rectum volume variation was observed during the first week of treatment. An accumulated delivered dose to 50% of rectum > 1194 cGy in fraction five was a significant predictor for exceeding

Conclusion We have developed a robust, quick and straightforward method to generate a library of plans for a combined bladder CTV and GTV using delineations of full and empty bladder CTs. The method is able to generate plans at every cm from full bladder. EP-1666 Adaptive radiotherapy in prostate cancer: when and why? R. Muelas 1 , R. García 2 , L. Vidueira 2 , J. Bonaque 2 , A. Conde 1 , C. Ferrer 1