Background Cancer survivorship is an international priority and mortality from cardiac damage is one of the greatest concerns. Anthracycline chemotherapy is cardiotoxic but remains highly effective against cancer.

Methods This translational project involved the development of a pre-clinical rat model of progressive anthracycline-induced cardiotoxicity (1.25mg/m² doxorubicin weekly for 8 doses) and a concurrent clinical study of 30 cancer patients receiving 6 cycles of curative anthracycline therapy (doxorubicin 50mg/m²). A bespoke cardiac magnetic resonance imaging (CMR) protocol was developed which included longitudinal T1 mapping (modified look-locker) to estimate extracellular volume (ECV), T2 mapping (to quantify oedema) and strain analysis. A panel of circulating biomarkers (including inflammatory, ischaemic and fibrosis markers) was evaluated by multiplex ELISA. Longitudinal histological analysis was performed in the rats and correlated with imaging and biomarker findings.

Results Left ventricular function (LVEF) declined steadily during treatment in rats and humans. Persistent LV dysfunction was seen in 23% of patients 12 months after therapy. Peak Troponin I levels correlated with fall in (LVEF) and circulating levels translated well between rat and humans. However, levels did not peak until the final cycle of chemotherapy when significant LV decline had already occurred. There was no significant change in the other circulating biomarkers. ECV did not change significantly during treatment but lower baseline ECV correlated with a greater fall in LVEF. Microscopic changes were seen in the rat myocardium after 6 doses of doxorubicin but electron-microscopy revealed mitochondrial damage after just one dose.

Conclusions This translational approach enabled forward and back translation leading to the development of a clinically relevant pre-clinical cardiotoxicity model. Troponin I was the most informative circulating biomarker but peaked at the end of therapy too late to modify treatment and prevent LV decline. The model has the potential to be used to identify earlier biomarkers and evaluate cardioprotective strategies. The imaging findings showed that fibrosis cannot be detected on CMR within one year of chemotherapy but generated a new hypothesis that patients with ‘healthier hearts’ may be a greatest risk of drug-induced cardiotoxicity.