Cardiac tomography 3D post-processing to assess percutaneous pulmonary valve implantation feasibility

Abstract

Percutaneous pulmonary valve implantation (PPVI) remains limited in patients with large native right ventricle outflow tract (RVOT). New 3D post-processing analysis derived from cardiac tomography (CT) may improve pre-procedural assessment. We retrospectively selected 15 patients who had invasive balloon sizing (IBS) and printed cardiac models with thermoplastic polyurethane (TPU) by laser sintering before PPVI. RVOT long-axis curvilinear reconstruction (LACR) and 3D volume rendering images were produced from CT scan by interventionists blinded to outcome, to assess minimal diameter and shape of the expected valve landing zone (proto-meso diastolic phase in 13 patients and systolic phase in two, Aquarius 3D, Tokyo) (Fig. 1). Median IBS was 26.0 [24.4–27.9] mm. LACR and IBS diameters were well correlated (r = 0.67, P = 0.007; r2 = 0.55, P = 0.002) with a mean bias of 2.8 mm. Printed model and IBS diameters were moderately correlated (r = 0.55, P = 0.04, r2 = 0.50, P = 0.003) with a mean bias of 0.9 mm (Fig. 2). Assessing size and shape of the models, 3 interventionists rated the expected complexity of PPVI from non-feasible (n = 3), challenging (n = 3), and straightforward (n = 9). Analysis of outcome confirmed non-feasible in 3, challenging procedures in 1, and uneventful in 11 (r = 0.85, P = 0.0001). Median fluoroscopy time was 18.6 min [16.9–26.2]. Expected complexity was correlated with fluoroscopy time (r = 0.76, P = 0.003). RVOT CT scan 3D reconstruction is feasible by interventionists and the cardiology team before PPVI and offered assessment of landing zone shape and diameter. CT scan acquisition in diastole underestimated the RVOT diameter. Nevertheless, it allowed accurate predicting of feasibility and complexity of PPVI.