ESTRO 37 Abstract book

S1240

ESTRO 37

Material and Methods Our in-house developed system for fully automated, multi-criteria EBRT treatment planning was extended with an option for optimisation of HDR dose distributions. The test cohort consisted of 22 planning CTs with catheter reconstructions and delineations for low- intermediate risk prostate cancer patients who were previously treated with 4 x 9.5 Gy HDR brachytherapy. Data of 5 patients was used to configure the autoplanning system. The fixed configuration was then used to also automatically generate treatment plans for the other 17 patients. Automatically generated plans were compared to the corresponding clinical plans. All evaluations were performed in the clinical TPS. Clinically applied hard constraints were used for automated planning: D 1cc < 80% of the prescribed dose (PD) for rectum and bladder, and D 1% < 120% of PD for the urethra. The objectives in order of priority were to 1) prostate V 100% = 95%, 2) minimise urethra D 1% , 3) improve conformality. For comparison, both the clinical and automated plans were rescaled to match 95% coverage. Results Automated plan optimisation took on average 19.5 seconds (range: 5.5 – 43.2 seconds), including computation of the dose kernels. The autoplans of 21/22 patients showed 95% tumour coverage within the imposed hard constraints. Coverage of patient 4 was lower than desired (94.8%), though slightly higher than in the clinical plan (94.5%), both due to the urethra constraint. For 21/22 patients the autoplan showed a reduction in the most important OAR objective, urethra D 1% (see figure). The absolute mean and maximum reductions were 1.8 Gy and 4.1 Gy, respectively.

However, the highest dose points are found inside the catheters volume and not in contact with the breast tissue. To solve this uncertainty, the model excluded the highest doses in the dose-volume histogram (DVH) corresponding to the volume of the catheters in the CTV. From this modified DVH, EQD2 was calculated for each voxel with dose d i (EQD2 i ) using linear quadratic (LQ) model. If incomplete repair effect between fractions is added, EQD2 i becomes:

Where d i is the dose in every voxel, n is the number of is the incomplete repair factor calculated with the monoexponential repair model. For a scheme of 2 fractions per day, h 2 = exp(-µx). EQD2g is defined as EQD2 i weighted by its total volume fraction w i at the modified DVH: 90 patients (8 fractions x 4 Gy: 53 patients; 7 fractions x 4.3 Gy: 37 patients) treated with accelerated partial breast irradiation (APBI) using high dose rate (HDR) interstitial brachytherapy were evaluated. EQD2g was calculated using the values α/β = 4 Gy, mean reparation time T 1/2 = 1.5 h, and time interval between fractions delivered the same day x = 6 hours. Results The highest excluded doses are (median ± standard deviation) 19 ± 2 Gy (range: 13.4 - 26.1 Gy). EQD2g obtained with the current model was (median ± standard deviation): 87 ± 8 Gy (range: 67 - 106 Gy). Conclusion The model presented in this study took into account the “hot spot” dose regions in the CTV in a robust way. Calculated EQD2g reveals large inter-patient dosimetry heterogeneity. However, it exhibits several limitations. As we are considering hot spot regions in brachytherapy (more than four times the PD), the LQ model is used in extreme circumstances. Furthermore, the highest doses excluded from the DVHs are supposed to be inside catheters, while optimisation can make the radioactive source stay much longer in selected catheters. Despite these limitations, the conclusions on the high variability in APBI dosimetry in brachytherapy are still effective. EP-2243 Fast automated multi-criteria planning for HDR brachytherapy explored for prostate cancer S. Breedveld 1 , A. Bennan 1 , S. Al-Uwini 1 , D. Schaart 2 , I.K. Kolkman-Deurloo 1 , B. Heijmen 1 1 Erasmus Medical Center Rotterdam Daniel den Hoed Cancer Center, Radiation Oncology, Rotterdam, The Netherlands 2 Delft University of Technology, Radiation Science & Technology, Delft, The Netherlands Purpose or Objective To develop an automated treatment planning workflow for high dose rate (HDR) brachytherapy, compatible with our clinical treatment planning system (TPS). In this workflow, the patient-CT with catheter reconstructions and dwell positions are imported from the clinical TPS into our in-house developed system for automated multi- criteria dwell time optimisation. The optimised dwell times are then imported in the clinical TPS. The aims of automation are planner-independent, enhanced plan quality and short optimisation times. fractions, and h m

Conclusion Fast, automated multi-criteria treatment planning for prostate HDR brachytherapy is feasible. For 21/22 patients, the autoplan met the coverage condition for the prostate with consistent reduction in near-maximum urethra dose (most important clinical objective). Plan optimisation took less than 20 seconds on average. EP-2244 Accuracy Verification of Dose Calculation Algorithm Based-on TG-43 and AcurosBV on HDR Brachytherapy P. Nakkrasae 1 1 Faculty of Medicine Siriraj Hospital- Bangkok- Thailand, Department of Radiology- Division of Radiation Oncology, Bangkok, Thailand Purpose or Objective To verify dosimetric accuracy of dose calculation algorithm based-on TG-43 and Varian Acuros TM BV on high dose rate (HDR) brachytherapy in in-house heterogeneous