ESTRO 2020 Abstract book

S1099 ESTRO 2020

range=18-23). The maximum axillary separation was >18 cm in 17/30 patients, 14 of which (82%) had a V36<90% while the remaining 3 had a V36<95%.Mean V36 was 78% for patients with separation ≥18 cm, whereas in patients with a separation <18 cm mean V36 was 95.9%.The deepest point of the PTV in patients who had V36<90% was located at a median of 9.8 cm (range= 8.8 -12.8). Depth was the reason of low coverage in all the 15 patients. Miss from the medial field border also contributed to poor dosimetry in 8/15 patients. A strong negative correlation was found between V36 and separation (R=0.735). The maximum depth of PTV showed a strong positive correlation with separation (R=0.618). Conclusion The dose distribution obtained from an anterior beam might be inadequate. The relationship between reduced PTV coverage and separation appears to be more robust when separation is 18 cm or greater. The maximum axillary separation is a simple parameter that can be used to predict the need to utilize other techniques to ensure adequate treatment of regional lymph nodes. PO‐1876 A visual grading analysis‐based audit for MR simulation sequence development L. McDaid 1 , L. Cooper 2 , T. Edwards 2 , A. McPartlin 2 , S. Bonington 3 , C. Eccles 1 1 The Christie NHS Foundation Trust, Radiotherapy, Manchester, United Kingdom ; 2 The Christie NHS Foundation Trust, Proton Beam Therapy Centre, Manchester, United Kingdom ; 3 The Christie NHS Foundation Trust, Radiology, Manchester, United Kingdom Purpose or Objective To assess recently implemented MR Simulation protocols used for treatment planning in head and neck patients being treated as part of a newly introduced proton beam radiotherapy (PBT) service. Material and Methods Based on a review of the literature, an imaging protocol comprising four sequences was implemented. Imaging was performed using a 1.5 Tesla (T) magnet, utilising a combination of 32-channel body array, 44-channel spinal array and two large flexible 2-channel transmit receive coils. Table 1 details the sequences used, but in short included 2- and 3 – dimensional acquisitions using turbo spin echo (TSE), driven equilibrium (DRIVE), spectral pre- saturation with inversion (SPIR) and mDixon sequences. All patients were imaged axially from superior orbital margin to sternal notch in the treatment position. To evaluate image quality and suitability, an audit was undertaken, using a visual grading analysis (VGA). The VGA was based on a 4-point scale, ranging from “very clear” to “not visible” for a series of pre- determined structures (table 2) by 3 multidisciplinary team members. Results Sixteen patient datasets were available for review. The preliminary analysis demonstrated a scan time of 31.26 minutes, with a mean ‘in-room’ time of 64.07 minutes (range 24-101 minutes). Using VGA tool, one radiographer, one radiologist and one clinical oncologist reviewed images (Table 1). Overall, an average of 69.4% of the sequences and structures were reported as “very clear” or “clear” (range 57.1-87.3%) (Table 2). The T2W TSE mDixon and T1 SPIR with intravenous contrast were the 2 sequences that scored the highest. The 3D T2W DRIVE sequence came out as unclear or had structures that weren’t visible in nearly 43% of the cases.

Conclusion Based on our analysis, the most appropriate sequences for planning PBT for head and neck tumours are a combination of T2W TSE mDixon and T1 SPIR with intravenous contrast. Considering clarity of named anatomical structures has facilitated an optimisation process of individual sequences. Using interdisciplinary VGA as an audit tool, we have implemented a continuous quality improvement (CQI) process for PBT treatment planning in head and neck cancers using MR. We anticipate this process will soon be implemented for other treatment sites at our institution. This philosophy of CQI will ensure sequence selection is evidence-based, the inclusion of appropriately selected sequences, and the removal of sequences felt to be less useful. PO‐1877 Survival and dosimetric parameters in stage III NSCLC patients undergoing radical chemo‐ radiotherapy I. Remmerts De Vries 1 , M. Ronden 1 , P. De Haan 1 , F. Spoelstra 1 , N. Haasbeek 1 , M. Dahele 1 , I. Bahce 2 , S. Senan 1 , W. Verbakel 1 1 Amsterdam UMC, Radiotherapy, Amsterdam, The Netherlands ; 2 Amsterdam UMC, Lung, Amsterdam, The Netherlands Purpose or Objective The standard of care treatment of stage III NSCLC is concurrent chemo-radiotherapy (cCRT). In less fit patients or patients that refuse cCRT it is sequential chemo- radiotherapy (sCRT) or radiotherapy (RT) alone. We explored the relationship between dosimetric parameters and survival after radical treatment. The main endpoint was overall survival (OS), calculated from the last day of RT. Material and Methods Patients were considered eligible if they were treated for stage III non-small-cell lung cancer (NSCLC) between 2015- 2017 and received radical intent cCRT, sCRT or RT alone with a total dose of ≥50 Gy delivered in ≥15 fractions. Patients were excluded if they underwent surgery, SABR or previously had radiotherapy on the thorax. The planning target volume (PTV) was the internal target volume (ITV) plus a 1 cm margin. Plans aimed to limit the volume of total lung receiving at least 20 Gy (V20) ≤35%, and limited the contralateral lung V5 as much as possible. In this period no hard constraints were used for limiting doses to the heart or esophagus. The following prognostic factors were collected: PTV, mean lung dose, total lung V20, V10, V5, contralateral lung V5, esophagus V40, V50, V60, V65, mean heart dose, heart V25, V40. OS was measured using Kaplan-Meier graphs. Predictors of OS were assessed by Cox proportional hazard analysis and expressed as hazard ratios and 95% confidence intervals. Results Between 2015-2017, a total of 129 stage III NSCLC patients underwent treatment to a dose of ≥50 Gy at our