Feasibility, reproducibility, and accuracy of echocardiographic right ventricular systolic function assessments in childhood cancer survivors at risk for heart failure

Heidi Ostler,Lin Liu,Khang Tong,Maria T. Acuero,Juliana Gomez‐Arostegui,Seth Degner,Sun Choo,Fraser Golding,Sanjeet Hegde,Dennis J. Kuo,Hari K. Narayan

DOI: https://doi.org/10.1111/echo.15905

2024-08-21

Echocardiography

Abstract:Echocardiography and cardiac magnetic resonance imaging performed within 60 days of one another were analyzed in 58 anthracycline‐exposed childhood cancer survivors. Right ventricular echocardiographic measurements were feasible and reproducible. Although measures did not correlate with CMR RVEF, participants with RV dysfunction had worse automated GLS (p = .007) and FWLS (p = .09). Purpose We sought to assess the feasibility, reproducibility, and accuracy of conventional and newer echocardiographic measures of right ventricular (RV) systolic function in adolescent and young adult childhood cancer survivors treated with anthracyclines. Methods Echocardiography and cardiac magnetic resonance imaging (CMR) were acquired ≤60 days apart in prospectively recruited survivors and RV functional measures were quantitated by blinded observers. Repeat quantitation was performed in a subset to evaluate reproducibility. For each echocardiographic measure, Spearman correlations with CMR measures were calculated, and values in participants with CMR RV ejection fraction (RVEF) ≥48% and RVEF <48% were compared using two sample Wilcoxon rank‐sum tests. Results Among 58 participants, mean age was 18.2 years (range 13.1–25.2) and five participants had CMR RVEF <48%. Intra‐ and inter‐observer coefficients of variation were 8.2%–10.1% and 10.5%–12.0% for adjusted automated strain measures, and 5.2%–8.7% and 2.7% for 3D RVEF, respectively. No echocardiographic measures were significantly correlated with CMR RVEF; only tricuspid annular plane systolic excursion was correlated with CMR RV stroke volume (r = .392, p = .003). Participants with RV dysfunction had worse automated global longitudinal strain (−20.3% vs. −23.9%, p = .007) and free wall longitudinal strain (−23.7% vs. −26.7%, p = .09). Conclusions Echocardiographic strain and 3D RV function measurements were feasible and reproducible in at‐risk childhood cancer survivors. Although not associated with CMR RVEF in this population with predominantly normal RV function, automated strain measurements were more abnormal in participants with RV dysfunction, suggesting potential clinical utility of these measures.

cardiac & cardiovascular systems