ESTRO 36 Abstract Book

S335 ESTRO 36 _______________________________________________________________________________________________

PO-0641 A novel voxel based homogeneity index: clinical implications for WBRT A.H. Thieme 1 , C. Stromberger 1 , P. Ghadjar 1 , A. Grün 1 , S. Zschaek 1 , V. Budach 1 1 Charité-Universitätsmedizin Berlin, Klinik für Radioonkologie und Strahlentherapie, Berlin, Germany Purpose or Objective Homogeneity of a treatment plan is considered to be important as underdosage may decrease the tumor control probability and overdosage may result in excess toxicity if the target volume contains organs at risk. Historically, homogeneity indices only used a few data points of a dose- volume histogram (e.g Dmin, Dmax, D5, D95) for calculation without the possibility to decide if inhomogeneity arises from under- or overdosage. We have introduced a novel voxel based homogeneity index (VHI) which uses the entire information present in the three- dimensional dose distribution. In contrast to traditional indices it is possible to determine to what extent under- and overdosage contribute to inhomogeneity. We analyzed the performance of the VHI versus traditional indices and whether VHI results were associated with treatment outcomes in patients who underwent whole-brain radiation therapy (WBRT). Material and Methods The VHI uses the deviations from the prescribed dose in each voxel of the target volume to calculate a characteristic score. The score is without dimension and independent of the target volume size, fractionation scheme and prescribed dose. We retrospectively analyzed patients who underwent therapeutic WBRT between July 2009 and May 2016. VHI results were compared with results of traditional indices. Overall survival of patients was assessed and compared with the underdosage part of the voxel homogeneity index (VHI_Underdosage) using Kaplan Meier plot analysis and Log-Rank test. Results A total of 517 WBRT plans were analyzed. 266 (51.5%) of patients were male and 251 (48.5%) female. The median age was 61.5 years, the median prescribed dose was 30.8 Gy and the median number of voxels was 115.137 per target volume with a resolution of 15.625 mm³ per voxel. While established indices underestimated inhomogeneity in certain cases, the VHI showed a significant higher sensitivity. Differences between radiation technologies (e.g. VMAT and 3D-CRT) could easily be differentiated with the new index, while existing indices were not able to do so. A difference in survival was observed according to the respective VHI_Underdosage. At the end of the first year the number of patients who died with an index < 1.9 was 142 (74.3%) compared with 239 (82.7%) with an index >= 1.9 (p = 0.038 by Log-Rank test).

Results The DTI analysis was successfully performed pre and post RT. Differences in FA, RD, and AD between pre and post RT MRI were assessed. The superimposition of the fiber tracts maps with the relative information about post treatment alterations on the dosimetric information of the simulation CT was finally obtained. Conclusion We show the feasibility of a standardized tract-based dosimetric analysis, potentially useful to establish possible relations between dose-volumes data and the variation of DTI indices in WM structures. PO-0640 Which measurement type should be used for disease control of brain metastasis, volumetric or linear? A. Fischedick 1 , G. Fischedick 1 , U. Haverkamp 2 1 Clemens Hospital, Radiation Oncology, Münster, Germany 2 University Clinic Muenster, Radiation Oncology, Muenster, Germany Purpose or Objective Which quantification method for disease tracking of brain metastasis after stereotactic radiosurgery (SRT) is correct, remains a controversial topic. The RANO-BM standards involving linear 2d measurement are considered to be the gold standard. However, 3d measurements are believed to be more accurate but thorough scientific evidence is still missing. This study set out to clarify this by analysing 55 patients with brain metastasis before and after SRT. Material and Methods Measurements were performed with OsiriX on gadolinium contrasted T1 weighted MRI images and analysed using Microsoft Excel. The RANO-BM criteria were applied to the 2d measurements whereas 3d measurements were categorized according to Matthew J. et al 2012. Results This study contained 26 males and 29 females with an average age of 61 years. The average survival rate post SRT was 13,9 months categorized into partial responds at 17 months, stable disease at 10,5 months and progressive disease at 9 months. The categorization for both 2d and 3d measurements were the same in 81,8% of the cases. However, 18,2% of the 3d measurements were identified as partial response whereas the 2d measurement placed them in the stable disease category. In every instance this could be explained by either weak gadolinium uptake, cystic morphology or small size of the metastasis leading to wrong volumetric measurements. Conclusion We conclude that the results obtained from 2d and 3d measurements are highly comparable and that no benefit from 3d tracking could be observed. Therefore, we recommend the use of the RANO-BM criteria in 2d linear measurements for clinical use.

Conclusion