ESTRO 38 Abstract book

S1025 ESTRO 38

MRCAT CT. Therefore, rectum cancer cases are now being tested with improved MRCAT software which is able to generate pseudo-CT images with infinite number of grey levels.

combination of 3D-CRT and a hydrogel spacer may become a treatment option for prostate RT. EP-1887 Dosimetric and volumetric evaluation of MR- only planning for radiotherapy of rectal cancer S. Vieira 1 , O. Pares 1 , N. Loução 2 , J. Stroom 1 , I. Santiago 3 , C. Greco 1 , C. Matos 3 1 Fundação Champalimaud, Radiotherapy, Lisboa, Portugal ; 2 Philips Iberia, Radiology, Lisbon, Portugal ; 3 Fundação Champalimaud, Radiology, Lisboa, Portugal Purpose or Objective MR delineations in radiotherapy are generally more accurate than on CT but for dose calculations CT data are mandatory. However, software has become available that transforms MR data into pseudo-CT data. In this abstract we study the pros and cons of switching to MR-only planning for radiotherapy of rectal cancer. Material and Methods Preliminary data of five patients with locally advanced rectal cancer were available. Radiotherapy VMAT plans (Varian) were developed using our clinical simulation protocol that includes the use of FDG PET/CT. Magnetic Resonance for Calculating Attenuation (MRCAT, Philips®) images with five different grey values were generated from 3D T1-weighted mDIXON images (outside the clinical protocol). We used 2 echoes, 1x1x2.5 mm spatial resolution, TR/TE1/TE2 1.64/3.8/5.7(ms), and scan time 3min16s. Two aspects of the new procedure were studied: Part 1: We compared dose calculation on the MRCAT and on the planning CT. MRCAT was matched with the planning CT and the treatment plan with tumor and OAR delineations was copied to the MRCAT. Then the dose was recalculated on the MRCAT (figure 1a/b). Part 2: A radiation oncologist and radiologist, both specialists in rectal cancer, together delineated MRI-only target volumes and OARs (figure 1c/d). New radiotherapy plans were then generated and compared with the existing clinical plans.

Conclusion Dosimetrically even low-resolution MRCAT images are suitable for planning rectum cancer patients. Using MRI only for target delineation reduced the CTV size considerably. When higher resolution MRCAT images become available for better online matching, planning CTs may be discarded completely for rectum radiotherapy. EP-1888 Validation of automated planning with RapidPlan for prostate bed VMAT radiotherapy B.L. Rekstad 1 , P. Lønne 1 1 Oslo University Hospital, Department of Medical Physics, Oslo, Norway Purpose or Objective RapidPlan is a module for automated planning (AP) in the Varian Eclipse treatment planning system based on DVH- estimation from previous treatment plans. We wanted to explore and validate the module for prostate bed VMAT, prescribed with only one target volume dose level and rectum, bladder, anal canal, femoral heads and penile bulb as organs at risk (OARs). The aim of the study was to investigate possible use of AP and to what extent AP is comparable to or better than manual treatment planning regarding target volume coverage, dose to OARs. Material and Methods A model was built in Eclipse 13.6 based on 70 clinical prostate bed VMAT plans. Treatment was given with 6 MV on a Varian Clinac 2300iX equipped with Millennium 120 MLC. Either one (n=55) or two (n=15) full arcs were used. Prescribed dose to the prostate bed CTV was 66 – 70 Gy delivered in 33 -35 fractions. The model was tuned by slightly varying the normal tissue objective (NTO) function parameters and also adding a maximum dose constraint to rectum, bladder and anal canal. A validation of the model was then performed by applying the model to 30 other prostate bed cases and comparing with manual clinical plans. Isocenter was the same in clinical and AP plans. All AP plans were made with one arc. With AP the model was applied and run through optimization once without further interaction. Results Target volume doses were comparable for manual and AP plans. Mean CTV D 98% was 98.3 % and 98.5 % and mean PTV D 98% was 95.8 % and 95.3 % for manual and AP plans respectively. This equal target volume coverage is also evident from the mean DVHs for CTV and PTV for the 66 Gy cases (n=23) (Figure 1). For OARs mean DVHs show lower AP doses for rectum, anal canal, femoral heads and penile bulb in medium dose regions, and slightly higher AP doses for bladder (Figure 1). Also, the rectal dose tends to be higher in the high dose region for some of the AP plans. This should be explored more before clinical use. For

Results Part 1: Regarding the dose comparison, doses in target volumes and critical organs calculated on MRCAT differed only slightly from clinical doses: on average mean doses in target and OARs <1%, with standard deviations of 0.4-2.2% (figure 2). Deviations could be explained more by differences in patient anatomy between planning CT and MRCAT than by differences in Hounsfield units. Part2: MRI-only pelvic PTV volumes were on average 14% (±19%, 1SD) smaller than clinical volumes. Just one case actually presented a larger MRI-only target volume due to better visualisation of the anterior mesorectal fascia. MRI- only boost PTV volumes were on average 40% (±5%, 1SD) smaller. Dose coverage of tumor volumes was similar in both MRI-based and clinical plans. A disadvantage of the MRCAT-based planning was that online matching with CBCT on the linac appeared hardly feasible due to the limited number of grey values in the