ESTRO 38 Abstract book

S992 ESTRO 38

1 Kyoto University Graduate School of Medicine, Department of Radiation Oncology and Image-Applied Therapy, Kyoto, Japan ; 2 Kyoto University Graduate School of Medicine, Department of Information Technology and Medical Engineering Human Health Sciences, Kyoto, Japan Purpose or Objective Recently, single isocenter volumetric modulated arc therapy (VMAT) has arisen as an alternative to whole brain radiotherapy when treating multiple brain metastases (BMs). The aim of this study was to evaluate the clinical validation of knowledge-based planning (KBP) for multiple BMs. Material and Methods Consecutive 56 patients treated with single isocenter VMAT for multiple BMs between October 2015 and September 2018 in our institute were enrolled into this study. The patient cohort was divided into two groups according to the number of BMs; model training (50 cases with 2-4 BMs) and validation (6 cases with 12-20 BMs). Four KBP models (M1-4) with different number of BMs per case were constructed using RapidPlan™ (Varian Medical Systems, Palo Alto, USA): M1, using all 50 cases with 2-11 BMs; M2, 40 cases with 2-6 BMs; M3, 30 cases with 2-4 BMs; and M4, 20 cases with 2-3 BMs. Then, the four KBP models were applied to each case in the validation group with the same beam arrangements as in a clinical plan (CL). The total number of BMs in the validation group was 96, whereas the clinical plans consisted of 88 planning target volumes (PTVs). One time of optimization per KBP model was performed for each case. The dosimetric differences between the KBP plans and CLs were assessed for PTVs and organs at risk (OARs). OARs included brain-PTVs, brainstem, chiasm, optic nerves, eyes, lens and skin. The following dose-volumetric parameters were recorded: PTV D 2% , D 50% and D 98% , the maximum dose to OARs and V 20Gy of brain-PTVs. Friedman test were performed for statistical analysis at α = 0.05, followed by Wilcoxon signed-rank test with Bonferroni correction for multiple comparisons as a post hoc test. Statistical significance was set at p < 0.01. Results Three to seventeen (median, 7) optimizations were required until the CLs were obtained. Interquartile ranges were smaller in the KBPs than in the CLs for any dose- volumetric parameters (Fig. 1). The difference between the median values in the CLs and those in the KBPs was within ± 1.7% points and ± 0.4% points for PTV D 50% and D 98% , respectively. No significant difference was observed among these dose-volumetric parameters ( p >0.05). Although the PTV D 2% was significantly higher in the KBPs than in the CLs ( p <0.01), the difference of its median value between the CLs and the KBPs ranged from 1.7 to 2.3% points. The maximum dose of brain-PTVs in the KBPs were decreased by 4.9% points (range, 1.5-11% points), compared with the CLs. The median values of difference in the maximum dose to other OARs and V 20Gy of brain-PTVs between the CLs and KBPs were ranged from -5.6% points to +0.5% points.

Figure 1: Dose volume histogram for the results obtained for PTV and hippocampi when optimizing utilizing distinct PTVX structures. X stands for the distance in mm from the PTVX structure to the hippocampi. Conclusion Optimizing using PTV7 structure as objective structure and evaluating final results on PTV implies maintaining PTV coverage and the achievement of an important reduction on hippocampi maximum dose. EP-1828 Treatment plan comparison between SBRT techniques for recurrent nasopharyngeal carcinoma Y. Lin 1 , H. Ho 1 1 Chi Mei Medical Center, Department of Radiation Oncology, Tainan, Taiwan Purpose or Objective To evaluate the potential benefit of HyperArc (HA) in previous treated, recurrent nasopharyngeal cancer treated with stereotactic body radiation therapy (SBRT). Material and Methods The twenty patients with recurrent nasopharyngeal cancer who were treated using CyberKnife (CK) for salvage treatment were enrolled. The median dose for the previous treatments for the twenty patients was 70Gy in 35 fractions. CK was delivered with a median 35 Gy in 5 fractions. The HA treatment plans were created for each patient to meet the same treatment plan criteria for CK. These two SBRT treatment plans were compared with target coverage, sparing of organs at risk, and dose distribution metrics, including conformity index, heterogeneity index, dose gradient index, and high/intermediate dose spillage. Monitor units (MU) were compared to assess as delivery efficiency. Results The HA plans consistently exhibited similar CTV and PTV coverage and significantly reduced the dose to organs at risk. The mean CTV coverage for CK and HA was 97.3% and 98.3%; the mean PTV coverage for CK and HA was 94.7% and 96.1%. Using HA plans, mean doses to the spinal cord, brainstem, optic nerves, and optic chiasm were reduced by 64%, 62%, 61% and 68%, respectively. The conformity and heterogeneity metrics of the HA plans were significantly better than the CK plans. With HA plans, the mean high dose spillage volumes were decreased by 54%. In average, the HA plans resulted in 56% less MUs than the CK plans (HA, 20643 MUs vs. CK, 47464 MUs). Conclusion Excellent sparing of organs at risk and good dosimetric distribution make HA an attractive SBRT technique for the treatment of recurrent nasopharyngeal cancer. EP-1829 Clinical validation of knowledge-based planning for multiple brain metastases N. Kishi 1 , M. Nakamura 1,2 , H. Hirashima 1 , N. Mukumoto 1 , K. Takehana 1 , M. Uto 1 , Y. Matsuo 1 , T. Mizowaki 1

Conclusion