ESTRO 2021 Abstract Book

S1026

ESTRO 2021

(GR) was defined as TRG 3 or 4. Based on post-chemoradiotherapy T2-weighted axial MR images, two deep learning models were constructed to predict pCR and GR, respectively. The predictive performance of the deep learning models was evaluated in the testing set and was compared to that of a senior radiologist and radiation oncologist. Results The deep learning model showed an area under the receiver operating characteristic curve, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 0.76, 0.30, 0.96, 0.67, 0.87, and 85.0% for predicting pCR and 0.72, 0.54, 0.81, 0.60, 0.77, and 71.7% for predicting GR, respectively. The deep learning model had a superior predictive performance than the observers. Fair agreement between the ground truth and the model was shown for pCR prediction (kappa = 0.34) and GR prediction (kappa = 0.36). Conclusion The post-chemoradiotherapy T2-weighted axial MR image-based deep learning model showed acceptable performance in predicting pCR or GR in patients with rectal cancer, compared with human observers. PO-1243 Is there a role for perioperative radiotherapy in surgically resected stage IV rectal cancer? J. Kwon 1 , J. Kim 1 , B.H. Kim 2 1 Chungnam National University College of Medicine, Department of Radiation Oncology, Daejeon, Korea Republic of; 2 Seoul Metropolitan Government Seoul National University Boramae Medical Center, Department of Radiation Oncology, Seoul, Korea Republic of Purpose or Objective This study aimed to determine whether perioperative radiotherapy (RT) improves outcomes in stage IV rectal cancer patients treated with primary surgical resection and systemic chemotherapy and to identify predictive factors for selection of patients for these approaches. Materials and Methods We searched the Surveillance, Epidemiology, and End Results (SEER) database for patients diagnosed between 2010 and 2015 with stage IV rectal cancer, but without brain or bone metastases. After applying the exclusion criteria, a total of 26,132 patients were included in the analysis; propensity score matching was used to balance their individual characteristics. Results Overall, 3283 (12.6%) patients received perioperative RT; the 3-year overall survival (OS) rates were 43.6% in the surgery group and 50.5% in the surgery with RT group (p < 0.001). The survival benefit of RT was maintained after propensity score matching and multivariate adjustment (HR 0.70, 95% CI, 0.66-0.81, p < 0.001). Interaction testing of the prognostic variables revealed a significant interaction between RT and the presence of lung metastasis (p < 0.001): the benefit of RT was observed only in patients without lung metastases (3-year OS 52.1% vs. 44.1%, p < 0.001), but it was observed regardless of liver metastases. Additionally, we developed a web-based calculator (http://bit.do/mRC_surv) to provide individualized estimates of OS benefit based on the receipt of perioperative RT. Conclusion Perioperative RT significantly improved OS rates, especially in patients without lung metastases. We successfully developed a nomogram and web-based calculator that could predict survival benefit with the addition of RT for these patients. PO-1244 Severe surgical morbidity after chemoradiotherapy conformational-3D versus IMRT for rectal cancer M. Martin 1 , N. Giraud 2 , M. Capdepont 3 , C. Sarrade 2 , A. Viaouet 2 , D. Smith 4 , E. Terrebonne 4 , N. Frulio 5 , A. Rullier 6 , Q. Denost 3 , E. Rullier 7 , V. Vendrely 2 1 Haut-Lévêque Hospital, Radiation oncology, Bordeaux, France; 2 Haut-lévêque Hospital, Radiation oncology, Bordeaux, France; 3 Haut-lévêque Hospital, Colorectal surgery, Bordeaux, France; 4 Haut-lévêque Hospital, Oncology, Bordeaux, France; 5 Haut-lévêque Hospital, Radiology, Bordeaux, France; 6 Haut-lévêque Hospital, Pathology, Bordeaux, France; 7 Haut-lévêque Hospital, Colorectal surgery, bordeaux, France Purpose or Objective Pre-operative chemoradiotherapy (CRT) for locally advanced rectal adenocarcinoma is responsible for early and late toxicity. This radiation can either be conformational in 3D (3D-RT) or intensity-modulated (IMRT). Comparative retrospective studies of preoperative CRT between these techniques have shown a reduction of acute adverse effects in favor of IMRT. Nevertheless, there is to date a lack of study exploring the impact of IMRT on post-operative morbidity. The main objective of this study is assess the severe surgical morbidity rates depending on the pre-operative radiation technique (3D-RT or IMRT). The secondary objectives were the evaluation of toxicity during CRT for all patients and digestive effects subsisting more than one-year post-TME. Materials and Methods All patients treated by pre-operative CRT and surgery at Bordeaux University Hospital between January 2012 and September 2019 for LARC were considered. The primary endpoint was the severe surgical morbidity, defined as any grade 3-4 event according to the Dindo-Clavien classification during the 30 days following the surgery and was prospectively recorded in the medical files by the surgery team. Acute toxicity events during the CRT were graded following the CTCAEv.5 and were collected weekly. Late digestive toxicity was evaluated according to the LARS score. The questionnaires were updated between August and October 2019 for patients treated by TME with an interval superior to one year after the surgery. Results Among 155 patients included 79 (51%) were treated with 3D-RT and 75 (49%) with IMRT (median dose: 50 Gy; range, 25-60) with concomitant chemotherapy (CT) for 149 patients (96%). Surgery consisted in anterior