ESTRO 36 Abstract Book

S365 ESTRO 36 _______________________________________________________________________________________________

3 Radiotherapy Institute Friesland, Radiation Oncology, Leeuwarden, The Netherlands 4 The Netherlands Cancer Institute, Radiotherapy, Amsterdam, The Netherlands Purpose or Objective Delineation variation of esophageal tumors remains a large source of geometric uncertainty. Recent studies trying to decrease esophageal tumor delineation variation by adding PET/CT information showed no significant reduction. With future studies focusing on dose escalation, a correct gross tumor volume (GTV) delineation will become more important. In the present study, we investigated the inter- and intra-observer variation in esophageal GTV delineation and the impact of fiducial markers on this variability. Material and Methods Ten esophageal cancer patients were included in this study, with at least 2 fiducial markers implanted at the cranial and caudal tumor border, visible on planning CT (pCT). Two GTV delineation phases were performed by five dedicated gastrointestinal radiation-oncologists independently. In-house designed software digitally removed the fiducial markers from the pCTs. Phase 1 consisted of a GTV delineation on the resulting marker- less pCTs (denoted pCT-m). Phase 2 consisted of a GTV delineation on the original (e.g., with markers) pCTs (denoted pCT+m). Phase 1 and 2 were executed twice to determine intra-observer variability. As general metrics for inter-observer variability; volumes, generalized conformity indices (CI gen ; 1 indicating full overlap and 0 indicating no overlap) and delineation variation (cranio- caudal (CC) variation and overall observer variation expressed as standard deviations (SDs)) were calculated. Intra-observer variability was determined in terms of overall observer variation. A Wilcoxon signed-rank test was performed to compare delineation phases (significance level α=0.05). Results For all delineated volumes, the results of the mean volumes, CI gen, CC delineation variation and overall observer variation, of all 10 patients are listed in Table 1. Inter-observer overall observer variation (1SD) was 0.64 cm and 0.28 cm in phase 1 and phase 2, respectively ( p =0.027 ) . CC delineation variation (1SD) was 1.48 cm for phase 1 and 0.43 cm for phase 2 ( p =0.004 ) . CI gen was significantly larger in phase 2 (0.54 vs 0.68; p =0.006 ) . Intra-observer overall observer variation (1SD) was 0.44 cm and 0.26 cm, for phase 1 and phase 2 respectively, significantly reduced with the aid of fiducial markers ( p =0.006).

Conclusion Large inter- and intra-observer variation in delineation of esophageal tumors was found when using marker-less pCTs. Presence of fiducial markers at CC tumor borders, significantly improved all investigated delineation variation parameters. Since fiducial markers only assist with the determination of the cranial and caudal tumor border, the largest variation reduction (1SD = 1.48 cm to 0.43 cm) was logically seen in the CC delineation variation. Our results endorse the use of fiducial markers in esophageal GTV delineation, consequently reducing geometric uncertainty in esophageal cancer radiotherapy. PO-0698 Phase II study of induction chemotherapy followed by radiochemotherapy in pancreatic cancer M. Fiore 1 , R. D'Angelillo 1 , B. Floreno 1 , P. Trecca 1 , L. Trodella 1 , A. Iurato 1 , L. Trodella 1 , R. Coppola 2 , S. Ramella 1 1 Campus Biomedico University, Department of Radiation Oncology, Roma, Italy 2 Campus Biomedico University, Department of General Surgery, Roma, Italy Purpose or Objective To evaluate the safety and efficacy of induction gemcitabine and oxaliplatin followed by a high weekly dose of gemcitabine-based radiochemotherapy in patients with borderline resectable or unresectable locally advanced pancreatic cancer. Material and Methods From January 2012 through January 2015, 41 patients with pancreatic cancer were included in the study. In all cases an accurate pre-treatment staging was performed. Patients received chemotherapy with gemcitabine 1000 mg/mq and oxaliplatin 100 mg/mq every 14 days for four doses. Patients without disease progression after restaging received conformal radiation therapy and concurrent gemcitabine at the dose of 600 mg/mq weekly. Results Further to the results of the pre-treatment workup, seven patients (17%) were excluded from the protocol because of the evidence of metastatic disease, and thus a total of 34 patients were consequently enrolled. Induction chemotherapy was well tolerated. Five patients (14.7%) experienced disease progression after induction chemotherapy.Twenty-seven patients completed the therapy protocol. Only haematological grade 3-4 toxicities were observed. Nineteen patients were evaluated through surgical exploration and fifteen patients (55.5%) underwent surgical radical resection. With a median