ESTRO 2020 Abstract book

S881 ESTRO 2020

spatial patterns of the dose distribution. Material and Methods Prospectively collected clinical and dosimetric data of 206 patients treated at a single institution were retrospectively analysed. These patients were treated with a variety of fractionation regimens between 2001 and 2013 (18 Gy in 3, 19 Gy in 2, and 17 Gy in 2 fractions), and had an average urethral toxicity rate of 12.6% defined as the first urethrotomy within 4 years after radiotherapy completion. Due to the tubular nature of the urethra, DSM was constructed for each patient by sampling points at a fixed radius distance of 3.5 mm (representative of a 22 Fr gauge three-way indwelling urinary catheter) from the interpolated centroid of the urethral contour, and then by unfolding along the anterior axis of the organ. Doses in each pixel were converted to Biological Effective Dose (BED) using an α/β=5 Gy. DSM were compared between patients with and without toxicity, average doses calculated and difference in spatial patterns were analysed with a 2-sample t-test (p<0.05) by aligning the DSM at the most inferior contour (CERR/Matlab). Results The average DSMs showed patterns of higher dose distribution closer to the apex on the right side of the urethra regardless of whether toxicity occurred (Fig 1a,b). Overall, the toxicity group was found to be irradiated with higher doses with a maximum difference of 20 Gy closer to the apex (Fig 1c). Looking at the pixel-wise difference of the mean dose maps of both groups (Fig 1c), the patterns on the urethra right side were confirmed to be statistically significantly correlated to urethral toxicity (Fig 1d).

processing methods. These features were mainly describing geometric aspects of the tumour rather than relationship/combination of grey levels. Out of the reproducible features, Spearman’s rank correlation coefficient identified 73 features robust to various pre- processing techniques. Figure1 shows an overview of the robustness of the texture features investigated in this work.

Figure1.Robustness of standardized radiomic features computed from T2-weighted MR images of patients with proven soft-tissue sarcomas. Conclusion We evaluated the robustness of different MR-based standardized radiomics features when various pre- processing schemes are used. Our results showed less than 100 features are both robust and reproducible. Further work is needed to investigate more the individual impact of these pre-process techniques on the discovered features.

References : 1.

Vallières M et al.A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities.The Cancer Imaging Archive, 2015 Clark K et al.The Cancer Imaging Archive(TCIA):Maintaining and Operating a Public Information Repository.Journal of Digital Imag, 2013 Whybra P et al.Assessing radiomic feature robustness to interpolation in 18F-FDG PET imaging.Sci Rep, 2019 https://arxiv.org/abs/1612.07003v7

2.

3. 4.

PO-1538 Dose-surface-maps patterns of risk of urinary toxicity for prostate patients treated with HDRB boost V. Panettieri 1,2 , Z. Anderson-Cutting 3 , J. Crosbie 3 , J. Millar 1,4 1 Alfred Health, Radiation Oncology, Melbourne, Australia ; 2 Monash University, Department of Medical Imaging and Radiation Sciences, Clayton, Australia ; 3 RMIT University, School of Science, Melbourne, Australia ; 4 Monash University, Central Clinical School, Clayton, Australia Purpose or Objective In the treatment of prostate cancer with HDR Brachytherapy (HDRB) boost several studies have shown correlation between urethral dose and the risk of urinary toxicity. However, clear guidelines for this organ are still lacking due to the variety of fractionation regimens and organ contouring methods used. This might limit traditional DVH-toxicity relationship estimates, due to the loss of spatial information in the 3D-dose distribution in the organ. The aim of this work was to use a pixel-wise comparison of urethral Dose-Surface-Maps (DSM) obtained in patients with and without urinary toxicity, in order to evaluate if the risk of urethra stricture was associated with distinct

Conclusion The risk of urethral strictures may be correlated not only to the dose-level provided but also to the area of the urethra irradiated. These results provide insight on local dose effects and provide guidance in the preparation and delivery of HDRB boost for prostate cancer. Additionally, the methodology of DSM, as outlined in this study could be extended to the growing use of hypofractionated external beam radiotherapy treatments focusing on limiting urethral toxicity, and potentially safely allowing additional dose-escalation.