Abstract 5713: Accuracy of Right Ventricular Volume and Function Analysis in Tetralogy of Fallot using Knowledge Based Reconstruction
Measurement of right ventricular (RV) volume and function by two-dimensional (2D) echo is inaccurate but 3D methods require much tracing. Knowledge Based Reconstruction (KBR) can rapidly 3D reconstruct the RV without requiring border tracing or performing imaging processing. Instead KBR employs a database to create a 3D surface to fit points that the user enters at valves, apex, septum, and free wall. We measured RV volume using KBR from magnetic resonance images in 20 patients with repaired tetralogy of Fallot (TOF) at end diastole (ED) and end systole (ES). We entered points in long and short axis and/or oblique views. The true volume was computed by manually tracing RV borders for 3D reconstruction by the piecewise smooth subdivision surface method. The reference database had 110 patients including 54 TOF and normal subjects. The KBR values agreed closely with true values for ED volume (Figure⇓) (r=0.972), ES volume (r=0.986), and ejection fraction (EF; r=0.922). KBR slightly overestimated ED and ES volume and EF (Table⇓). There was no bias in the error by Bland Altman analysis (p=NS for ED and ES volume and EF). Accurate volumes were achieved in two runs and entry of 57±15 points. KBR provides accurate measurement of RV volume and EF with minimal user input. KBR is a clinically feasible alternative to full manual tracing of heart borders from image data.