Evaluation of Cardiac Shunts With 4D Flow Cardiac Magnetic Resonance: Intra- and Interobserver Variability

Background: In the last decade, the capacity of magnetic resonance (MR) to evaluate congenital anomalies has improved substantially. To date, only a few studies have evaluated the value of 4D-flow MRI in shunt assessments.

Purpose/hypothesis: To assess the intra- and interobserver variability of 4D-flow MRI in patients diagnosed with cardiac/extracardiac shunt. Secondarily, to assess the feasibility of directly measuring the shunt and to determine the prognostic correlation with the pulmonary-to-systemic (Qp/Qs) flow ratio.

Study type: Retrospective.

Population: In all, 18 patients with cardiac shunt diagnosis.

Field strength/sequence: 1.5 T/4D phase-contrast MRI.

Assessment: Pulmonary and systemic flows were measured at different locations to assess the internal consistency by two observers (twice by one, and once by the other). The Qp/Qs ratio was calculated. When feasible, direct flow was quantified by planimetry.

Statistical tests: Spearman's rho correlation coefficient was used to assess the relationship between pulmonary/systemic flows measured at different levels and to compare the jet characteristics with prognostic data as right ventricle volume. Intra- and interobserver variability were determined by Bland-Altman plots and interobserver correlation.

Results: The most common shunt type (n = 10; 55.5%) was ostium secundum atrial septal defect (ASD). Direct visualization and quantification of shunt flow was possible in all studies. Pulmonary and systemic flows showed a strong correlation between these measures (Spearman's rho [r] of 0.872 and 0.899). The mean Qp/Qs ratio was 1.61(0.62). Mean flow rate was 2.01(1.68) l/min. The mean jet diameter was 11.88 (5.44) mm. Intraobserver (r = 0.97) and interobserver correlation (ICC = 0.95) for the Qp/Qs calculation were both excellent. Direct measurement of flow was strongly correlated (r = 0.98; ICC = 0.95). Correlation was strong between Qp/Qs and direct jet flow (r = 0.76 and 0.77), Qp/Qs and mean jet diameter (r = 0.79 and 0.94), and Qp/Qs with jet area (r = 0.77 and 0.94).

Data conclusion: Measurement of the Qp/Qs ratio and direct shunt quantification using 4D-flow MRI was feasible, and highly reproducible. Internal consistency was excellent, with low intra- and interobserver variability. Correlation between the Qp/Qs ratio, direct flow measurement, mean diameter, and jet area was strong.

Evidence level: 3 TECHNICAL EFFICACY: Stage 2 J. Magn. Reson. Imaging 2020;52:1055-1063.