ESTRO 36 Abstract Book

S145 ESTRO 36 2017 _______________________________________________________________________________________________

on MRI are potentially smaller than CT-based volumes, which could lead to lower dose to organs at risk (OARs) and, in turn, reduction of RT-induced toxicity. The purpose of this study is to ascertain potential reduction in target volume and OAR dose. Material and Methods 23 breast cancer (cTis-3N0M0) patients from the MILANO trial (NL50046.041.14) were scanned in supine position on 1.5 T, arms abducted, after SN biopsy and breast- conserving surgery. MRI included a 3-dimensional (3D) T1- weighted (T1w) spoiled gradient echo (T1-SPGR) anatomical scan and two T2w fast spin echo (FSE) techniques for LN detection, which were co-registered. Axillary levels were delineated, using ESTRO guidelines [Offersen et al. 2015, IJROBP ], as well as OARs, including the lungs, heart, chest wall (CW), brachial plexus (BP), and humeral head (HH). LNs were identified by 4 observers, and delineated. Encompassing LN volumes – and after 5 mm isotropic expansion of the LNs – were related to axillary levels. In 5 patients (17–26 LNs), elective RT of 16 x 2.66 Gy = 42.56 Gy, delivered by 13 intensity- modulated RT beams, was simulated on MRI for two situations: (i) axillary levels I-IV, and (ii) all individual LN- based targets (1 mm PTV margin). For this, pseudo-CT scans were generated by bulk assignment of Hounsfield units on MRI for water, lungs and air. OAR dose parameters in both (i) and (ii) were compared. Results A median of 26 axillary LNs were delineated per patient. Compared to the respective axillary levels, LN-based target volumes, even after 5 mm isotropic expansion, are considerably smaller [table 1]. Coverage of all targets was excellent ( V 95% > 99%, V 107% = 0; all PTVs) in (i) and (ii). For elective RT on LN-based PTVs [figure 1], dose to all OARs was substantially reduced compared to standard elective RT: the average reduction of mean dose to lungs, heart, and HH was 2.3 Gy, 2.2 Gy, and 13.3 Gy, respectively; reduction of maximum dose to the BP and CW was 25.5 Gy and 9.4 Gy.

Conclusion LN-based target volumes on MRI are considerably smaller than axillary levels, conventionally delineated on CT according to the ESTRO guidelines. Addition of dedicated MRI in regional RT planning leads to reduced OAR dose, and a potential reduced RT-associated toxicity for breast cancer patients . In the near future, this will be investigated for more patients, and these results will be available at ESTRO 36. Moreover, MR imaging of lymph vessels is being investigated. Introduction of MRI-guided regional RT, by direct visualization and delineation of individual LNs and OARs, and future use on the MRL, may reduce RT-induced toxicity. PV-0282 Out-of-plane motion correction in orthogonal cine-MRI registration M. Seregni 1 , C. Paganelli 1 , J. Kipritidis 2 , G. Baroni 1,3 , M. Riboldi 1 1 Politecnico di Milano University, Dipartimento di Elettronica- Informazione e Bioingegneria, Milano, Italy 2 University of Sydney, Radiation Physics Laboratory- Sydney Medical School, Sydney, Australia 3 Centro Nazionale di Adroterapia Oncologica, Bioengineering Unit, Pavia, Italy Purpose or Objective Online motion monitoring in MRI-guided treatments currently relies on the acquisition of 2D cine-MRI images that are registered to the planning anatomy 1 . However, out-of-plane motion (OOPM) cannot be measured and it could affect the accuracy of 2D-2D registration