Abstract Book

S54

ESTRO 37

Material and Methods We analyzed data from 281 EC patients treated with preCRT followed by surgery at our institution between 2002-2016. The endpoints for analysis were postoperative pulmonary and cardiac complications, which were prospectively scored according to the international definition of Low et al (Ann Surg 2015). We assessed complications as a Comprehensive Complication Index (CCI) above 1750, or 300 in case of pleural effusions requiring prolonged chest drainage after surgery. For each individual radiotherapy plan, the lungs and heart were manually delineated according to the current algorithm and recalculated dose-volume histograms (DVH) were exported. Radiation dose varied between 36 Gy and 50.4 Gy, delivered in fractions of 1.8 Gy or 2 Gy. An univariable logistic regression analysis was performed to study the predictive value of lung and heart dosimetric variables (volumes receiving 5 Gy (V5) to 55 Gy (V55) and mean dose) for postoperative pulmonary and cardiac complications. The Area Under the Curve (AUC) of the receiver operating characteristic curve was calculated. Results Respectively 153 (54%) and 90 (32%) of the 281 patients developed pulmonary and cardiac complications. Normal- tissue complication probability modelling identified that the volume of the lung receiving 30 Gy (V30), 35 Gy (V35) and 40 Gy (V40) was statistically significantly associated with the risk of postoperative pulmonary complications (respectively AUC=0.59, 0.58 and 0.57; p=0.025, 0.025 and 0.033; Fig. 1). The volume of the lung receiving 5 Gy (V5) and 10 Gy (V10) was associated with postoperative cardiac complications (respectively AUC=0.56 and 0.59; p=0.047 and 0.015; Fig. 2). There was a trend to more cardiac complications with increasing mean lung dose (AUC=0.57; p=0.054). No cardiac dose volume parameter was correlated to neither cardiac nor pulmonary complications.

Conclusion Our data suggest that the volume of the lung receiving doses higher than 30 Gy is an important dosimetric parameter in the development of postoperative pulmonary complications in EC patients treated with preCRT followed by surgery. Moreover, the volume of the lung receiving doses from 5 to 10 Gy is predictive for postoperative cardiac complications, whereas no correlation was seen between heart dose and cardiac complications. This emphasizes that lung-heart interactions are important in dose-modelling. In that respect, further research should focus on optimizing the delivery of the radiation treatment to decrease the volume of the lung receiving both low and high radiation doses. Multivariable analysis, taking into account clinical and treatment-related factors, is currently ongoing. PV-0101 Impact of cardiac dose on survival and toxicity after neo-adjuvant treatment for esophageal cancer. E. Oldehinkel 1 , M.J. Van der Heiden 1 , J.C. Beukema 1 , M. Dieters 1 , A.C.M. Van den Bergh 1 , P. Van Luijk 1 , J.A. Langendijk 1 , C.T. Muijs 1 1 University Medical Center Groningen, Radiation Oncology, Groningen, The Netherlands Purpose or Objective Several factors are known to influence survival and toxicity in patients treated with tri-modality therapy for esophageal cancer (EC). As survival improves there is a rising concern about treatment-related toxicity due to cardiac radiation dose. However, evidence remains scarce and most literature is based on retrospective patient cohorts. We evaluated the role of cardiac dose by assessing its relation with overall survival (OS) and cardiac toxicity in a prospective patient cohort. Material and Methods We selected all 137 patients treated with neo-adjuvant chemoradiation (neoCRT) included in our prospective data registration program (PDRP) for patients treated with curative intent for EC at our department until March, 2017. The prescribed dose to the PTV was 41.4 Gy delivered in 23 fractions with a hybrid IMRT or VMAT technique. Chemotherapy administered was carboplatin and paclitaxel. Treatment outcome and toxicity were scored at 3 months and yearly thereafter. The heart and its substructures (LA, RA, LV, RV) were contoured using an automatic delineation tool based on the atlas by Feng et al. The lungs were considered as one organ. To analyze the influence of radiation dose on OS and