Abstract Book

ESTRO 37

S382

PO-0744 LBP is predictive for long-term cardiac complications in breast cancer patients. J. Chalubinska-Fendler 1 , Ł. Graczyk 2 , G. Piotrowski 3 , B. Tomasik 4 , K. Wyka 5 , M. Spych 1 , A. Papis-Ubych 6 , W. Fendler 4 , J. Fijuth 1 1 Medical University of Lodz, Radiotherapy Department, Lodz, Poland 2 Radomian Oncology Centre, Radiotherapy Department, Radom, Poland 3 Voivodeship Centre for Traumatology and Oncology, Cardiology Department, Lodz, Poland 4 Medical University of Lodz, Department of Biostatistics and Translational Medicine, Lodz, Poland 5 Medical University of Lodz, Department of Pediatrics- Oncology- Hematology and Diabetology, Lodz, Poland 6 Regional Oncohematology Centre, Department of Radiotherapy and General Oncology, Lodz, Poland Purpose or Objective Radiation-induced cardiac toxicity (RICT) appears as soon as 2-3 years after radiotherapy (RT) and no good early predictive markers have been identified thus far. We decided to evaluate the potential of lipopolysaccharide binding protein (LBP) as a biomarker of RICT in patients with breast cancer. Material and Methods A prospectively recruited cohort of 52 women (mean age 59.25+/-8,78 years ) underwent RT in 2014 for cancer localized in the left breast. RT was performed using 3D planning, with a total dose of 50Gy and 10-16Gy boost to tumor site. Serum concentrations of LBP were evaluated using ELISA (Abnova, Taiwan) kits before RT, 24 hours after administering the last RT fraction and 1 month after RT. Dose-volume histogram (DVH) data were derived from RT plans. In 26 women who reached the planned follow- up time of 3 years, cardiac ECHO examination was performed by a specialist in cardiology. Results Significant, positive correlations were found between post-treatment LBP concentrations and heart and lung DVH: Left Anterior Descending artery dose (LAD) mean R =0.34; p=0.01; Left Ventricle dose (LAD) mean R=0.32; p=0.020, Mean Lung Dose (MLD) R=0.29; p=0.039. Post- treatment concentrations of LBP maintained their correlations with the following DVH parameters: LAD mean R=0.28; p=0.04 and MLD R=0.28; p=0.047. In ECHO examination we noted that the E/E’ parameter (ratio between early mitral inflow velocity and mitral annular early diastolic velocity) correlated positively with both post-treatment and follow-up LBP levels (Figure 1: R=0.46, p=0.016; Figure 2: R=0.50; p=0.01) confirming incipient diastolic dysfunction. A non-significant trend for a negative LBP and ejection fraction correlation was also observed (R=-0.24;p=0.22 and R=-0.27;p=0.19). The correlation between LBP and the E/E’ index persisted for post-treatment LBP concentrations after adjusting for patient age, type of surgery and mean dose to the left ventricle (beta=0.45; p=0.049) and was close to significance for follow-up LBP levels (beta=0.43; p=0.055). Conclusion LBP concentrations reflect the biological impact of radiation during chest RT. Post-treatment LBP levels may be predictive for early cardiac complications in cancer survivors. PO-0745 Postmastectomy radiation leads to increased mortality in patient with nodal micrometastasis G. Eastwick 1 , M. Daugherty 2 , J. Bogart 1 , A. Shapiro 1 1 Suny Upstate Medical University, Radiation Oncology, Syracuse- NY, USA 2 Suny Upstate Medical University, Urology, Syracuse- NY, USA

whole heart (WH), Right atrium (RA), Left atrium (LA), Right ventricle (RV) and left ventricle (LV), divided in 5 different segments (anterior, inferior, apical, septal and lateral). For each substructure of all patients, a multicontour structure (MCS) based on a majority vote algorithm was then created. A consensus based delineation (CBD) of all substructures was then developed by an independent team of two blinded operators (ID and GF). Dice similarity coefficients (DSC) between volumes delineated by each different operator and multicontour delination were collected and reported, as well as DSC between independent consensus and multicontour. See Figure 1 for a case contour example.

Results Considering WH, RA, LA, RV and LV, mean DSCs between single substructures delineated by each operator and corresponding MSCs ranged between 0.78 and 0.96. Furthermore, analyzing each delineated patient, mean DSC between substructures delineated by all single operators and corresponding MSC ranged between 0.84 and 0.94. For all these heart substructures, mean DSC between CBD and corresponding MSC ranged from 0,89 to 0,97. Considering LV segments, mean DSC between single segments delineated by each operator and corresponding MSC ranged from 0,38 to 0,77; within each patient, mean DSC between single operator delineation and MSC ranged from 0,54 to 0,72. For LV segments, mean DSC between CBD and corresponding MSC ranged from 0,58 to 0,63. See Table 1 below for details.

Conclusion Overall, results showed good overlapping between heart chambers contours, both comparing MSC with single operator delineation or CBD; LV segments showed a considerably worse result. However, the current experience showed low interobserver variability of heart substructure delineation, allowing a reliable dosimetric assessment of these volumes within SAFE2014 trial.