Abstract Book

S1091

ESTRO 37

planning era. These were used to reconstruct cardiac doses using five reconstruction methods: 1) a field superposition method which estimates the mean dose to organs at risk based on irradiated areas and as % of the prescribed dose. ( %Prescribed ) 2) a simple patient-specific approach, where the mean dose to the heart is estimated from the percentage cardiac area exposed within the 2D simulation fields ( %Heart ) 3) a method estimating dose to OAR based on an anthropomorphic phantom ( Phantom ) 4) a representative CT technique based on male and female anatomical data sets ( RepCT ) 5) a “2D to 3D” method using deformable image registration ( Navigator ) The estimated cardiac doses were compared to those calculated on each patient’s individual CT (“True Value” method) and the standard error of prediction for each method was estimated using linear regression. Results The results can be seen on Table 1. The %Prescribed was the quickest method to use but it had the largest standard error of prediction (SEP) of 5.8 Gy. The Phantom is the most widely used method in the literature and with a SEP of 2.5 Gy. The %Heart method was simple to use and for the MHD it had a low SEP of 1.6 Gy. The RepCT and Navigator methods were the most labour intensive and the SEP was 1.9 and 1.3 Gy respectively. The %Heart method can only estimate the mean dose to the whole heart while the %Prescribed can estimate the dose to the heart and cardiac substructures. The Phantom, the RepCT and the Navigator have also the potential to estimate the dose to substructures of the heart but additionally, they can provide volumetric data.

Conclusion Neglecting occurring dose variations by assuming a constantly delivered planned dose or averaging of the delivered dose leads to deviations and uncertainties in EQDx and α/β predictions. This leads to an underestimation of the actually applied biological dose. Furthermore, the assumption of a mean dose (spatial or temporal) causing a certain observed outcome rather than considering dose variation, results in an underestimation of the derived α/β value. Both factors can especially affect normal tissues in dose gradient regions, and our findings suggest that the magnitude of these effects should be investigated in patient cases. We conclude that local analysis of the dose can reduce uncertainties in dose response modeling which are induced by spatial as well as temporal averaging and this may create an additional benefit from daily imaging information. EP-2004 Cardiac Radiation Dose Reconstruction in the Study of Late Effects:A Comparison of Different Methods G. Ntentas 1 , M.C. Aznar 2 , D.C. Hodgson 3 , R.M. Howell 4 , M.V. Maraldo 5 , S. Ahmed 3 , A. Ng 3 , S.C. Darby 1 , D.J. Cutter 1 1 University of Oxford, Nuffield Department of Population Health, Oxford, United Kingdom 2 The University of Manchester, School of Medical Sciences - Faculty of Biology- Medicine and Health, Manchester, United Kingdom 3 Princess Margaret Cancer Centre, Department of Radiation Oncology, Toronto, Canada 4 The University of Texas, Radiation Dosimetry Services, Houston, USA 5 Rigshospitalet, Department of Oncology, Copenhagen, Denmark Purpose or Objective Radiation-related cardiovascular disease can occur several decades after radiotherapy. Cardiac radiation dose information is not easily available for most historic patient cohorts. Therefore, to investigate radiation- related cardiac late effects, it is necessary to reconstruct the doses delivered to the heart retrospectively, often without individual CT planning scans. Several reconstruction methods have been published and their dose prediction accuracy has been questioned over time. Here we evaluate their performance in Hodgkin lymphoma (HL) patients. Material and Methods Fourteen patients treated with CT-based radiotherapy for mediastinal HL were selected for this study. Two- dimensional digitally reconstructed radiographs were reconstructed to mimic simulation films, which are available for patients treated in the past, during the 2D