ESTRO 2020 Abstract book

S418 ESTRO 2020

reoptimization. Patient outcomes, treatment factors, and daily adaptation were evaluated. Results Median follow-up from completion of SMART was 9.2 months (range, 1.1-46.2). Among all fractions, the reoptimized plan was chosen for 93% of fractions by the treating physician. Median progression free survival (PFS) was 6 months (95% CI, 3.9 to 8.2) with 6-month and 12- month PFS rates of 43.2% (28.0-56.5), and 22% (11.3-34.8), respectively. The overall local control rate was 70.5%. There was limited acute toxicity with no grade 3 acute toxicities. Late toxicity included two (4.6%) grade 3 toxicities (gastrointestinal ulcers requiring intervention), and three (6.8%) grade 2 toxicities (duodenal perforation, antral ulcer and gastric bleed). Conclusion This is the largest reported experience of 50 Gy in 5 fractions to inoperable pancreas cancer. SMART enables the delivery of ablative radiation doses with comparable toxicity to CT based, non-ablative SBRT delivery. Further studies evaluating target volume modifications to decrease local recurrence and validating clinical outcomes are warranted. OC-0685 MRI-linac-based head-and-neck radiotherapy: a new concept for optimal neck sparing T. Van Heijst 1 , J. Mases 1 , S.L. Meyer Viol 1 , E. Brand 1 , C.H.J. Terhaard 1 , P.H. Doornaert 1 , M.E.P. Philippens 1 , C.P.J. Raaijmakers 1 1 UMC Utrecht, Radiotherapy, Utrecht, The Netherlands Purpose or Objective Elective radiotherapy (RT) in cN0 head and neck (HN) cancer is a proved strategy for regional control. Nevertheless, large treated volumes are associated with elective RT. CT is used to define targets, lacking sufficient sensitivity to identify individual lymph nodes (iLNs). In contrast, MRI can be used to directly delineate the iLNs for elective RT planning. MRI/RT hybrid modalities, such as the 1.5 T MRI linear accelerator (MRI-linac, Unity, Elekta) could facilitate new RT strategies targeting the iLNS in the neck with on-line MRI capability. Aim is to investigate potential reduction of organ at risk (OAR) dose by comparing standard CT-based elective RT to MRI-linac- based RT of iLNs in the neck. Material and Methods Ten cases of larynx carcinoma T1-4N0 were retrospectively selected. For treatment planning, HN imaging in an immobilization mask was performed on CT (Brilliance, Philips) and on a 3 T MRI with two flexible receive coils (Ingenia, Philips). A T2-weighted turbo spin echo sequence with multiple-Dixon water-fat separation was registered to the planning-CT, and used to delineate LN levels II- IV (CTV), OARs and all iLNs in the LN levels. A 3-mm margin to the CTV was applied to generate the conventional PTV (cPTV). The encompassing volume of all iLNs with 3-mm margin created the so-called selective PTV (sPTV). Numbers of iLNs and volumes of cPTV and sPTV were compared. Three strategies (35x1.55Gy=54.25Gy) were investigated: (A) conventional elective RT on the cPTV, by volumetric-arc RT on a 6-MV linear accelerator (linac); (B) MRI-linac highly selective RT: intensity-modulated RT (IMRT) targeting the sPTV, performed on a 7-MV MRI-linac; (C) Cascaded MRI-linac selective RT: IMRT on cPTV and sPTV, with a lower dose prescribed to cPTV (35x1.03 Gy=36 Gy). Strategy C serves as intermediate technique which targets iLNs but also irradiates the LN levels. In order to show solely the effect of irradiating iLNs vs. entire LN levels, primary tumor dose was not taken into account.

Treatment planning software Monaco (Elekta) was used for strategies A (5.10), and for B and C (5.40), which incorporates the magnetic field. V95% was aimed at >99%. Dose plans were compared by assessing the mean OAR dose in two ways: B vs. A and C vs. A. Results An average of 35 (range: 25-58) iLNs were identified per patient. An average 85% target volume reduction of the sPTV (22.5cc) compared to the cPTV (155.7cc) was achieved. Mean OAR dose (table 1, figure 1) was lowered by 44% on average per OAR in strategy B vs. A and by 24% per OAR for strategy C vs. A.