Evaluating Biomarkers as Predictors of Cancer Therapy Cardiotoxicity: All you Need is a Meta-Analysis? Reply

With improvements of cancer therapy and increasing numbers of long-term survivors, the timely diagnosis and management of cardiovascular complications is highly relevant to prevent morbidity and mortality in cancer patients. Biomarkers have been tested as diagnostic tools for cardiotoxicity in various settings, but aggregate evidence has been missing. We are thankful for the recent appreciation of our meta-analysis, which depicts the existing data for troponins and natriuretic peptides concerning this matter. We indicated a predictive role of troponin for cardiac dysfunction following anthracycline chemotherapy, human epidermal growth factor receptor 2 inhibitor therapy, and high-dose chemotherapy regimens, based on 61 prospective studies.1 One of the newer investigations is a recent study by Demissei et al.,2 which we were asked to re-consider for aggregate analysis. We have re-calculated our results accordingly and included this study. The updated meta-analysis further supports our general finding for troponin as a marker for cardiotoxicity (Figure 1A), but convincing evidence for a role for (N-terminal pro) brain natriuretic peptide remains warranted (Figure 1B). Cardiotoxicity is not limited to conventional chemotherapy, but may also result from modern targeted and immune therapies, particularly immune checkpoint inhibitors (ICI).3 Troponin elevations are associated with ICI-related myocarditis (sensitivity of 94%), and correlate with adverse outcomes if sustained.4 Most available evidence is from retrospective analyses and small case series. Given the lack of multiple systematic evidence for troponin testing and event rates under ICI therapy, ICI therapy regimens had to be excluded from this current meta-analysis based on our pre-specified inclusion criteria.1 Previous cancer studies have been limited, at least in part. This relates to the heterogeneity of cancer entities, cancer therapy, and variations regarding biomarkers and echocardiography parameters. Within our meta-analysis, questions have been raised regarding e.g. the different types of biomarker assays and cut-off values for left ventricular dysfunction. We note that our subgroup analysis regarding these issues did not provide any detectable differences.1 Also, data on long-term outcomes and clinical endpoints in patients with increased biomarkers and ejection fraction reductions are largely missing currently, but a small number of individual studies shows an increased mortality in cancer patients with elevated cardiac biomarkers.5 Therefore, we agree that only large cohesive registries with robust clinical endpoints will be able to provide definite evidence for cardio-oncology patients, which is one of the most relevant findings of our analysis.1