ESTRO 2020 Abstract book

S334 ESTRO 2020

irradiation. However, the consequences of partial splenic irradiation during stereotactic ablative radiotherapy (SABR) are not well understood. The objective of this study was to investigate changes in spleen volume (SV) during the course of stereotactic magnetic resonance-guided adaptive radiotherapy (SMART) for adrenal lesions. Material and Methods All patients treated with adrenal SMART at a single institution were retrospectively reviewed. Fractions were typically given every-other-day. No splenic dose constraints were used though doses were limited whenever possible. The spleen was recontoured on daily magnetic resonance (MR) scans for all patients to obtain SVs. A second observer verified the contours on a random basis. Splenic dosimetric parameters were obtained from daily on-couch post-adaptation SABR plans using the MRIdian planning system (ViewRay Inc.). The relationship between relative splenic volume change (SVC) and the cumulative mean dose (CMD), the overall percent volume of spleen receiving >10 Gy (V10) and the overall V20 were investigated with linear regression. Logistic regression was used for preliminary splenic dose thresholds. C-statistics were used for logistic regression goodness-of-fit. Results A total of 43 patients (218 total fractions) underwent SMART for adrenal lesions (27 left-sided) between 2016 to 2019. Prescription doses ranged from 24 Gy in 3 fractions to 60 Gy in 8 fractions (median: 50 Gy in 5 fractions). Mean SV prior to the first and last treatments were 223.0 cc and 203.3 cc, respectively (Wilcoxon signed-rank test p < 0.001). CMD ranged from 0.5 Gy-17.2 Gy. There were strong associations between the relative SVC and the CMD (-1.15% per Gy increase in CMD, p < 0.001), V10 (-1.88% per 10% increase in V10, p = 0.007) and V20 (-5.17% per 10% increase in V20, p = 0.006). Two patients who were subsequently treated for contralateral adrenal lesions had persistent SV reductions at 3 and 11 months after initial SABR. Logistic regression showed significant associations between all three dosimetric parameters and the probability of having a significant SV reduction, defined as >25%. Preliminary splenic dose thresholds are presented in Figure 1 below.

PD-0556 Using multiple planning scans to predict organ shape variability during RT for prostate cancer K. Furmanová 1 , R.G. Raidou 2 , N. Grossmann 2 , O. Casares- Magaz 1 , V. Moiseenko 3 , J.P. Einck 3 , L.P. Muren 1 1 Aarhus University Hospital, Danish Center for Particle Therapy, Aarhus, Denmark ; 2 TU Wien, Institute of Visual Computing & Human-Centered Technology, Wien, Austria ; 3 University of California San Diego, Radiation Medicine and Applied Sciences, San Diego, USA Purpose or Objective In radiotherapy (RT), treatment plans are usually designed based on a single CT scan acquired prior to treatment. However, organs in pelvic and abdominal regions are known to significantly change their shape over a course of therapy. These changes make dose distribution in the original plan non-representative of actual accumulated dose. Furthermore, organ shape has been advocated as a factor governing the development of normal tissue toxicity. We evaluated whether mean organ shapes from multiple CT scans acquired over several days can capture the organ shape variability occurring during the full treatment. The shape variability captured within this period can potentially provide a better basis for treatment plan evaluations and plan comparisons (e.g., proton vs. photon-based RT) or for an increased (or decreased) risk of complications. Material and Methods Twenty-four patients previously treated with VMAT for localized prostate cancer were included in this study. For each patient, cone-beam CTs (CBCT) were acquired during the delivery of 13 treatment fractions—first five days followed by weekly scans. Contouring of bladder and rectum was performed for all patients. The CBCTs were pre-registered on a per-patient basis by medical experts. To also include the organ motion, we preserved the relative mutual position of all organs. We used the Hilbert space-filling curve to linearize the 3D volumetric organ shapes into 1D vectors. The imaging data set for each patient was divided into two groups (first six CBCTs and subsequent seven CBCTs). The scans from the first group were used to compute mean shapes across two to six scans. Subsequently, we computed distances of the remaining seven scans to the mean shapes, calculated their variability, and compared it to the variability calculated with only one CBCT scan used (Fig. 1).

Conclusion Dose-dependent SVC was seen during the course of upper abdominal SABR for adrenal lesions, and preliminary splenic SABR dose thresholds are reported. Patients with high predicted splenic doses during SABR could be candidates for prophylactic vaccination to mitigate the effects of hyposplenism. Additional study, with long-term follow-up, is required to determine the relationship between SV reduction and changes in spleen function.