ESTRO 2023 - Abstract Book

S533

Sunday 14 May 2023

ESTRO 2023

MO-0641 Late cardiac events from lymphoma treatment -a meta-regression analysis of dose/response L. Nygård 1 , I.R. Vogelius 1 , S.M. Bentzen 2 , L. Specht 1 1 Copenhagen University Hospital, Rigshospitalet, Department of Oncology, Copenhagen, Denmark; 2 University of Maryland, School of Medicine, Department of Epidemiology and Public Health, Baltimore, MD, USA Purpose or Objective Treatment of lymphoma is highly curable but there can be long-term cardiotoxicity associated with both radiation and chemotherapy. We performed a meta-regression analysis of long-term cardiac events from lymphoma treatment. The objective was to identify dose-response relationships for both anthracycline and radiotherapy on the risk of cardiotoxic events subdivided in congestive heart failure (CHF), ischemic heart disease (IHD), and valvular heart disease (VHD). Materials and Methods Data on lymphoma patients was sought in PubMed with inclusion of papers published between January 2000 until December 2021. Patient numbers should be above 100. Data should contain cardiac outcomes in lymphoma patients, radiation doses to the heart and anthracycline doses. Cardiac outcomes should be reported for CHF, IHD and/or VHD for a study to be included. Results We found eleven eligible papers. All three cardiac outcomes had relevant cumulative incidence twenty-five years post treatment (CHF around 5%, IHD 5-10% and VHD 10-15 %). The excess relative risk (ERR) of CHF for anthracyclines per 100mg/m2 was 74% (CI: 55% to 94%). For radiotherapy, ERR of CHF per Gy of mean heart dose was 6.1%/Gy (CI: 4.3% to 7.8%). Corresponding numbers for the other endpoints were: IHD: No significant effect of anthracycline dose, ERR=5.5%/Gy (CI: 3.1% to 7.9%). VHD: ERR=25%/100mg/m2 (CI: 13% to 36%) and ERR=10%/Gy (CI: 6% to 15%). There were no signs of a “safe” lower dose level for neither anthracyclines nor radiotherapy dose on the affected endpoints. No interaction between the two cardiotoxic agents could be identified. Figure 1 shows the fit for CHF and Figure 2 shows all models of ERR versus dose.