ESTRO 2020 Abstract book

S940 ESTRO 2020

delivered dose using radiotherapy planning (RTP) CT and cone beam CT (CBCT) acquired in 10 patients during cervix EBRT. Target (CTV) and OARs, were manually segmented on 90 datasets and ART strategies modelled were: 1. ITV+ standard margins based on RTPCT and diagnostic MRI (clinically delivered plan) 2. CTV+ standard margins based on radiotherapy planning CT (RTPCT-only) 3. Partial uterine irradiation: CTV defined from GTV on RTP MRI+15 mm (Small CTV) 4. Plan of the day, using CTV defined on full, mid and empty bladder volume RTP imaging (POD) Composite target volume defined from CBCTs acquired days 1-3 Daily online re-planning using CTV defined on CBCT acquired at treatment +3 mm PTV margin Daily online re-planning using CTV defined on CBCT acquired at treatment +5 mm PTV margin Daily online re-planning defined on CBCT acquired at treatment using partial uterine irradiation MRI-CTV +3 mm PTV margin CBCT electron density was established by bulk density assignment and dose for each plan was calculated on the weekly CBCT. Computed deformable image registration was used to deform the CBCT and its dose to the RTPCT. All plans were quality assured for standard target and OAR planning objectives. The simulated delivered dose to the target and OARs was summated and compared for different treatment strategies using Friedman’s test with Dunn’s test for multiple comparisons. Results Underdosing of CTV occurred in one patient for the RTPCT only, D1-3 and partial uterine irradiation non-adaptive plans, where significant inferior CTV motion occurred due to bladder and rectal volume change. Compared to the clinically delivered plan, ART resulted in OAR sparing, particularly at higher doses, with online re- planning providing maximal OAR sparing. Table 1 and figure 1 illustrate increasing OAR sparing observed with increasing sophistication of adaptation. Largest reductions were seen with online 3 mm PTV margin strategy with standard and small CTV. Compared to the clinically delivered plan, POD resulted in 44% and 70% reduction in rectum and bladder V4500 cGy. There were statistically significant differences in delivered V4000 cGy to rectum and bladder for all online strategies with re-planning, when compared to the clinically delivered plan. 5. 6. 7. 8.

(c) deformable radiobiological accumulation In (a) image sets were merged initially using rigid registration. In (b) and (c) DIR was applied to deform anatomical information from a previous image set to the current image set. These DIR based transformation matrices were then used to deform the corresponding dose matrix (dose warping) from previous to current anatomy to allow deformable dose accumulation. In (a) and (b), physical dose accumulation was performed, assuming all contributing dose plans were prescribed as the standard 2Gy per fraction. In (c) correction for non- standard dose fractionation, was achieved using radiobiological calculation of individual dose plans to create its EQD2-based DVHs prior to dose accumulation. See figure 1. The Dice Similarity Coefficient (DSC) was used as a parameter for assessing the result of DIR on multiple patient studies as recommended by AAPM TG132 1 . 1 AAPM Task Group 132: Brock KK, et al. Med Phys. 2017 Jul;44(7)

Results Addition of dose by combining multiple treatment plans for retreatment patients was best achieved using method (c) deformable radiobiological accumulation. Quantitative DIR assessment in terms of DSC values showed that DIR performed on rigid structures such as the spinal cord (0.8) and mandible (0.8) were clinically acceptable as per AAPM TG132. OARs affected by motion such as the heart (0.82) and normal lung tissue (0.97) also showed acceptable DIR. The bowel (0.4) and bladder (0.35) were classified as complex OARs to perform DIR. Correction for hypofractionated dose prescriptions in examined patients showed clinically significant differences when compared with corresponding physical dose accumulation methods. Conclusion This study suggests that combining DIR and radiobiological dose recalculations for overall dose accumulation provides more representative dose information when compared to standard rigid registration based analysis. DIR and radiobiological dose in treatment plan evaluation has the potential for optimizing the standard of care delivered to patients requiring retreatment. PO-1626 Cervical cancer IGRT: A comparison of offline, online and MRI guided adaptive planning strategies I. White 1 , D. McQuaid 2 , E. Goodwin 2 , J. Mohajer 2 , S. Lalondrelle 3 1 Institute Of Cancer Research And Royal Marsden Hospital NHS Trust, Clinical Oncology, London, United Kingdom ; 2 Royal Marsden Hospital NHS Trust, Physics, Sutton, United Kingdom ; 3 Royal Marsden Hospital NHS Trust, Radiotherapy, Sutton, United Kingdom Purpose or Objective Quantify benefit of different adaptive radiotherapy (ART) strategies, using dose accumulation to assess target and OAR doses. Define optimal adaptive planning solutions for

cervical cancer EBRT. Material and Methods

Two hundred and fifty VMAT plans were created in RayStation 7.0R (RaySearch, Stockholm) to model