120 Chapter 8 cardiac complications after CRT for EC[11]. On the other hand, AF is quite common in the elderly population and is frequently seen in the perioperative period after intrathoracic surgery, which is relevant to mention as most patients are treated with CRT in the neo-adjuvant setting [14]. In our retrospective study (chapter 3), the number of patients with newly diagnosed AF (N=8 (4%)) was relatively low and consequently insufficient for reliable modelling procedures. In our cross-sectional study (Chapter 4), we found a difference in the onset of AF between the irradiated patient group and the non-irradiated surgical group who were on average older and thus we expected a higher AF rate after surgery alone. However, the opposite was found, with six patients diagnosed with AF in the nCRT versus two patients in the surgery alone group. Although the numbers are very small, we tried to look further into the mechanisms behind this toxicity as the study was designed as an hypothesis generating pilot study. Radiation dose to the left atrium was associated with a new onset of AF when corrected for age. In this study, we performed several imaging techniques, which enabled us to correct for other possible confounders, such as a wide atrium (left atrial volume index(LAVI)), another well-known risk factor for the development for AF in the general population. In the univariate analysis, both mean left atrium dose and LAVI showed borderline significant association with AF. However, when LAVI and radiation dose to the left atrium were combined in a multivariable model, both parameters were significantly associated with AF, and the model showed a high discriminative power with a high area under the curve (AUC 0.93). These findings suggest that clinical and baseline risk factors are important confounders that may affect toxicity endpoints and its relationship with radiation dose distributions. Our results are in line with those found by Song et al, who reported on 677 oesophageal cancer patients who underwent surgery after neoadjuvant CRT and tried to find risk factors for new onset AF [15]. In the multivariate analysis, only higher age (p<0.00) and higher prescribed radiation dose (p=0.03) remained significant. However, in this study, radiation exposure to the heart itself was not considered. Regarding the association between radiation dose to (substructures) of the heart and the onset of AF, only two papers have been published so far. The first study was performed in lung cancer patients [16], and reported the maximum dose in the sinoatrial (SA) node located in the right atrium to have the best association with the onset of AF.
RkJQdWJsaXNoZXIy MTk4NDMw