Objective To systematically review the literature and assess the diagnostic value of biomarkers in detection of late-onset left ventricular (LV) dysfunction in childhood cancer survivors (CCS) treated with anthracyclines.

Methods We systematically searched the literature for studies that evaluated the use of biomarkers for detection of LV dysfunction in CCS treated with anthracyclines more than 1 year since childhood cancer diagnosis. LV dysfunction definitions were accepted as an ejection fraction <50% or <55% and/or a fractional shortening <28%, <29% or <30%. Contingency tables were created to assess diagnostic accuracies of biomarkers for diagnosing LV dysfunction.

Results Of 1362 original studies screened, eight heterogeneous studies evaluating four different biomarkers in mostly asymptomatic CCS were included. In four studies, an abnormal N-terminal pro-B-type natriuretic peptide (NT-proBNP, cut-off range 63–125 ng/L) had low sensitivity (maximally 22%) and a specificity of up to 97% for detection of LV dysfunction. For troponin levels, in five studies one patient had an abnormal troponin value as well as LV dysfunction, while in total 127 patients had LV dysfunction without troponin elevations above cut-off values (lowest 0.01 ng/mL). Two studies that evaluated brain natriuretic peptide and nitric oxide were underpowered to draw conclusions.

Conclusions In individual studies, the diagnostic value of NT-proBNP for detection of LV dysfunction in CCS is limited. Troponins have no role in detecting late-onset LV dysfunction with cut-off values as low as 0.01 ng/mL. Further study on optimal NT-proBNP cut-off values for rule out or rule in of LV dysfunction is warranted.