ESTRO 38 Abstract book

S190 ESTRO 38

the IMRT group were accepted (24 (88.9%) vs 16 (64%), respectively; p=0.03). Conclusion The high frequency of protocol deviations underlines the importance of a QA program in clinical trials. Further work should assess the impact of deviations on patient outcomes. IMRT might allow dose escalation without increasing normal tissue complications. OC-0382 Patterns of local failure after SBRT for pancreatic cancer: implications of target volume design X. Zhu 1 , C. Yangsen 1 , Z. Xianzhi 1 , S. Yuxin 1 , J. Xiaoping 1 , Q. Shuiwang 1 , C. Fei 1 , J. Zhen 1 , F. Fang 1 , G. Lei 1 , Z. Huojun 1 1 Changhai Hospital, Radiation Oncology, Shanghai, China Purpose or Objective To identify prognostic factors and patterns of local failure in patients with pancreatic cancer receiving stereotactic body radiation therapy (SBRT) plus chemotherapy as initial treatment, for the optimal design of target volumes encompassing a majority of local recurrences. Material and Methods Consecutive patients with resectable or borderline resectable but medically inoperable due to comorbidities and locally advanced pancreatic cancer undergoing SBRT and chemotherapy in our center were reviewed. Local recurrences were plotted with respect to the celiac trunk (CT), superior mesenteric artery (SMA) and splenic artery (SA) on 1 CT scan of a template patient. Results Five hundred and ten patients were included. Median follow-up of the entire group was 21.8 months (range: 3.1- 54.9 months). Two hundred and seventeen patients had locoregional recurrences while local and distant progressions were found in 293 patients. One hundred and sixty-nine (33.2%) and 144 (28.2%) patients had recurrences closer to the CT and SMA, respectively, while both invasions of the CT and SMA were found in 115 patients (22.5%). Additionally, 33 patients (6.5%) and 49 patients (9.6%) had recurrences at the hepatic hilum and invasions of the SA, respectively. Besides these patterns of failure, 138 patients (27.1%) also experienced retroperitoneal progressions. The mean distance to the CT, SMA and retroperitoneal recurrence was 9.0mm, 8.3mm and 11.7mm, respectively. Multivariable analysis demonstrated that advanced pancreatic cancer, recurrences at both the CT and SMA and the hepatic hilum, CA19-9 non-responders and BED 10 <60Gy were predictive of worse survival. Moreover, failures stratified by volumes of recurrences in the radiation field were also analyzed. In-field and outside-the-field recurrences alone were found in 127 (24.9%) 111 patients (21.8%), respectively, while 51 (10.0%) and 67 patients (13.1%) had in-field plus outside-the-field recurrences and marginal plus outside- the-field recurrences, respectively. Compared with patients with BED 10 <60Gy, fewer patients with BED 10 ≥60Gy had in-field recurrence alone (68 patients vs. 59 patients, P=0.003) and in-field plus outside-the-field recurrences (29 patients vs. 22 patients, P=0.027). No differences were found in the incidences of margin recurrence alone (57 patients vs. 97 patients, P=0.108), outside-the-field recurrence alone (40 patients vs. 71 patients, P=0.128) and marginal plus outside-the-field recurrences (22 patients vs. 45 patients, P=0.091) between patients receiving BED 10 <60Gy and BED 10 ≥60Gy, respectively. Conclusion Areas closer to the CT, SMA and retroperitoneal space were at a high risk of local recurrences. Non-uniform and enough expansions from gross tumor volumes may be necessary and the splenic vessels abutting to the tumor might also be included in the target volume without compromise of dose constraints of organs at risk.

In this cohort of 1102 patients with EC who were treated with nCRT followed by surgery at four high volume institutes, it was demonstrated that the total radiation dose has a major impact on pathological response, especially in AC patients. OC-0381 Benchmark case in the ongoing PRODIGE 26 trial : quality assurance of dose escalated radiatherapy J. Boustani 1 , R. Rivin Del Campo 2 , J. Blanc 3 , D. Peiffert 4 , K. Benezery 5 , R. Pereira 6 , E. Rio 7 , E. Le Prisé 8 , G. Créhange 1 , F. Huguet 2 1 Centre Georges François Leclerc, Radiation Oncologist, Dijon, France ; 2 Tenon Hospital, Radiation oncology, Paris, France ; 3 Centre Georges François Leclerc, Biostatistics, Dijon, France ; 4 Institut de Cancérologie de Lorraine, Radiation Oncology, Vandoeuvre-Lès-Nancy, France ; 5 Centre Antoine Lacassagne, Radiation oncology, Nice, France ; 6 Centre Guillaume Le Conquérant, Radiation Oncology, Le Havre, France ; 7 Institut de Cancérologie de l’Ouest, Radiation Oncology, Saint Herblain, France ; 8 Centre Eugène Marquis, Radiation Oncology, Rennes, France Purpose or Objective The ongoing phase II/III PRODIGE 26 trial compares chemoradiotherapy with or without dose escalation in patients with locally advanced or unresectable oesophageal cancer. The results of a benchmark case procedure are reported here to evaluate the protocol compliance of participating centers as part of the radiation therapy quality assurance (RTQA) program. Material and Methods Volume delineation, target coverage, and dose constraints to the OARs were assessed on treatment plans performed by each participating center and compared to parameters defined in the protocol (Table 1). Centers were classified in three categories: per protocol (PP), minor acceptable deviation (MiD), or major unacceptable deviation (MaD). A plan was rejected if ≥ 4 MiD or one MaD were found. Results Thirty-seven centers submitted 52 plans. Among them, 17 (32.7%) were PP, 30 (57.7%) presented MiD, and 11 (21.1%) had MaD. Overall, 12 (23%) plans were rejected. One plan was rejected because the GTV was not correctly delineated. There were 8 (15.4%) MiD for CTV1 margins and 11 (21.2%) for CTV2 margins, due mostly to insufficient margins. The OARs delineation and the PTVs margins were respected in all cases. As for the target volume coverage, 10 plans had under-dosage among which 3 (5.8%) were major. Six plans had over-dosage among which 1 (1.9%) was major. When considering the maximal dose (Dmax) to the spinal cord, there were 2 (3.8%) plans with MiD and 1 (1.9%) plan with MaD. Regarding the lungs V20 and V30, 1 (1.9%) and 3 (5.8%) plans presented MiD and MaD, respectively. As for the heart V40, only 35 (67.3%) plans were PP, 13 (25%) had MiD and 4 (7.7%) had MaD. The liver V30 and the kidneys V20 were respected in all plans.Overall, 52% (n=27) of all treatments were planned with IMRT, while 48% (n=25) were planned with 3D-CRT. Deviations for target volume and OARs between 3D-CRT and IMRT plans are illustrated in Figure 1. The median PTV2 D2% was significantly lower in the IMRT group (p=0.01). The spinal cord Dmax was significantly lower in the IMRT group (median: 37.0 Gy (15.0-48.0), vs. 41.0 Gy (33.0- 46.0), respectively; p=0.003). The lungs V20 did not differ between groups (p=0.31), whereas the median V30 was significantly lower in the IMRT group, 9.0% (6.0-18.0) vs. 13.0% (7.0-28.0), respectively (p=0.003). The median heart V40 was significantly lower in the IMRT group (25.0% (14.0-36.0) vs. 30.0% (17.0-58.0), respectively; p=0.02). The liver V30 and the kidneys V20 did not differ significantly between groups. Overall, there were 13 (48.1%) and 4 (16%) PP plans in the IMRT and 3D-CRT groups (p=0.01), respectively. Significantly more plans in