Session Item

The effect of radiation exposure on cardiovascular disease has been studied in several publications, across many patient diagnoses and with varying definitions of cardiovascular disease. For example, in Breast cancer the rate of major coronary events increased linearly by 7.4% per gray of mean radiation dose to the heart.  

The purpose of the research is to estimate the association between radiotherapy exposure and cardiovascular toxicity after radiotherapy. In particular the aim is to establish quantitative associations between normal tissue exposure and incidence/prevalence of side effects. The rationale for the project is to improve our understanding of such associations, which will in turn allow us to better tailor our radiotherapy to minimize the risk of the observed cardiovascular toxicity while maintaining or improving the probability of durable disease control. 

This is a non-randomized retrospective study with a large sample size to estimate radiation dose-effect relationship with cardiotoxicity, specifically myocardial infarction (MI) defined by ICD-10 code I21. Eligible patients were diagnosed with cancer and received curative intent radiotherapy at Rigshospitalet between 2009 and 2016. Manual delineations of organs at risk are time consuming and thus effectively limited to small populations. Therefore, heart annotation was conducted using RootPainter3D trained from 933 manually delineated CT scans in a hematological toxicity study. All delineations were reviewed by a researcher and questionable cases were discussed with a physician.  

Of the patients considered for inclusion, 5544 were eligible for analysis. The diagnoses of greatest interest with regard to heart exposure are Breast (n = 2139), NSCLC (n = 411), Upper GI (n = 406), and SCLC (n = 106). Of those 2.6% of Breast (n = 55), 10.1% of NSCLC (n = 45), 7.1% of Upper GI (n = 29), and 8.5% SCLC (n = 9) experienced MI following radiotherapy treatment. 

This abstract represents preliminary findings for a large retrospective cohort study. This study supports a direction away from mean organ at risk analysis and toward distributional dose quantification. Additional organs at risk should be segmented and accounted for. Furthermore, quantification of dose to substructures of the heart may provide further insights. While this is a large study compared to the existing literature on this topic, there is an opportunity to expand the timeframe as well as location, which may improve generalizability  and power of the study.