Right ventricular function assessment: comparison of geometric and visual method to short-axis slice summation method.

Background: Short-axis summation (SAS) method applied for right ventricular (RV) volumes and right ventricular ejection fraction (RVEF) measurement with cardiac MRI is time consuming and cumbersome to use. A simplified RVEF measurement is desirable. We compare two such methods, a simplified ellipsoid geometric method (GM) and visual estimate, to the SAS method to determine their accuracy and reproducibility.
Methods: Forty patients undergoing cine cardiac MRI scan were enrolled. The images acquired were analyzed by the SAS method, the GM (area and length measurement from two orthogonal planes) and visual estimate. RVEF was calculated using all three methods and RV volumes using the SAS and GM. Bland-Altman analysis was applied to test the agreement between the various measurements.
Results: Mean RVEF was 49 +/- 12% measured by SAS method, 54 +/- 12% by the GM, and 49 +/- 11% by visual estimate. There were similar bias and limits of agreement between the visual estimate and the GM compared to SAS. The interobserver variability showed a bias close to zero with limits of agreement within +/-10% absolute increments of RVEF in 35 of the patients. The RV end-diastolic volume by GM showed wider limits of agreement. The RV end-systolic volume by GM was underestimated by around 10 ml compared to SAS.
Conclusion: Both the visual estimate and the GM had similar bias and limits of agreement when compared to SAS. Though the end-systolic measurement is somewhat underestimated, the geometric method may be useful for serial volume measurements.