BACKGROUND: Right ventricular (RV) function is of interest in an array of cardiopulmonary diseases. First-pass radionuclide ventriculography (FP), gated blood-pool single photon emission tomography (GBPS) and cardiac magnetic resonance imaging (MRI) are three currently used non-invasive methods for evaluation of right-sided cardiac function. The aim of our study was to compare the agreement between these methods when measuring right-sided cardiac function. METHODS: Twenty-four healthy volunteers were included. Mean age was 44 years (range: 25-60) and 29% were females. All participants had FP, GBPS and breath-hold cine MRI performed according to standard protocols. RESULTS: Normal ranges for RV ejection fraction (RVEF) defined as mean +/- 2SD were 0.49-0.72, 0.44-0.66 and 0.40-0.69 when measured by MRI, FP and GBPS respectively. Bland-Altman analysis showed a mean difference (bias) between MRI and FP of 0.05 (95% CI: 0.03-0.08) and of 0.06 (95% CI: 0.02-0.10) between MRI and GBPS. No systematic bias was found between FP and GBPS. Normal values for RV end-diastolic volume index (RVEDVI) were 37-95 and 29-91 ml m(-2) when measured by MRI and GBPS respectively. The mean difference between RVEDVI was 6 ml m(-2) (95% CI: 0-11). CONCLUSIONS: (i) Normal values of RVEF differ between MRI, FP and GBPS with wide limits of agreement, accordingly it is difficult to evaluate changes over time if measured by different methods, (ii) RV volumes are in the same range when measured by MRI or GBPS but with wide limits of agreement, and (iii) if MRI is considered gold standard then FP is more accurate than GBPS for RVEF measurements.