Univariate regression models with 95% confidence intervals (CIs) are shown demonstrating the inverse exponential relationship between pulmonary vascular resistance (PVR) and (A) pulmonary artery flow (Qp) and (B) left-to-right shunt. Given the patient numbers, the CIs are quite wide; however, all patients with an elevated Qp or shunt in excess of 2.5/1 have a PVR 3.5 Woods Units (WU)/m². Of note, only a small number of patients had an elevated PVR.

Univariate regression models with 95% CIs are shown demonstrating the direct exponential relationship between PVR and (A) mean pulmonary artery pressure (MPAP) (mm Hg) and (B) the mean transpulmonary gradient. The CIs for MPAP are very wide, and the MPAP can be in excess of 40 mm Hg with the PVR remaining within an acceptable range. This illustrates that pulmonary artery pressure measurements alone provide an incomplete picture, and clinical decisions should not be based on pressure parameters alone. Abbreviations as in Figure 1.

Scatterplot and corresponding receiver-operator characteristic (ROC) curves for shunt magnitude (left-to-right shunt [Qp/Qs]) (A, B) and Qp (C, D). All patients with a PVR 3.5 WU/m² had a Qp/Qs <2.5/1 with an area under the ROC curve of 0.94 or a Qp <6.05 l/min/m², area under the ROC curve of 0.90. With future validation these values could be applied clinically, where those with a Qp of 6.05 l/min/m² or a shunt 2.5/1 measured by phase contrast cardiac magnetic resonance could proceed to surgery without invasive assessment. Abbreviations as in Figure 1.