Improvements in cross-sectional imaging techniques have allowed their utilisation in critically ill children with heart disease. Paediatric patients on veno-arterial extracorporeal membranous oxygenator may benefit from three-dimensional imaging for assessment of intracardiac anatomy and extracardiac vasculature. Patients with CHD may need veno-arterial extracorporeal membranous oxygenator in the immediate post-operative period for treatment of cardiogenic shock or as a bridge to cardiac transplantation. ^{Reference Perry, Brown and Misfeldt1}

Cardiac CT in a patient on a veno-arterial extracorporeal membranous oxygenator can be challenging. ^{Reference Hull, Young and Thacker2} Imaging techniques are affected by flow and cannula-related artefacts. Cannulae with flows needed for full mechanical circulatory support can also affect the correct timing of acquisition due to the unpredictability of contrast density. Full extracorporeal membranous oxygenator flows can lead to contrast dilution and non-diagnostic images. The technique of pausing or lowering the extracorporeal membranous oxygenator circuit has been described previously in paediatric patients with CHD while performing a non-gated chest CT. ^{Reference Hull, Schooler and Binkovitz3,Reference Lambert, Grus and Balik4} However, performing an electrocardiogram-gated cardiac CT in patients with CHD on a veno-arterial extracorporeal membranous oxygenator has not been previously described. Electrocardiogram-gated cardiac CT offers the advantage of functional information in addition to higher resolution and diagnostic accuracy. Key logistical considerations include patient transport and whether the patient’s haemodynamic state will allow for lower extracorporeal membranous oxygenator flows for the study so contrast density in the vasculature can be optimised. Each protocol is individualised to the patient and clinical question. Herein, we share our experience and lessons learned while establishing a successful clinical workflow to perform cardiac CT in paediatric patients on a veno-arterial extracorporeal membranous oxygenator. All cardiac CTs described in this report were performed using a dual-source CT scanner (SOMATOM Force, Siemens Healthcare, Forchheim, Germany).

The first case was a 4-week-old male with Trisomy 21 who underwent a right Blalock-Taussig-Thomas shunt placement for Tetralogy of Fallot with hypercyanotic spells. He experienced a cardiac arrest in the immediate postoperative period, requiring emergent veno-arterial extracorporeal membranous oxygenator cannulation. Cardiac CT was requested to evaluate the Blalock-Taussig-Thomas shunt and distal pulmonary artery architecture, which were not well visualised on echocardiography. A dedicated 22-gauge peripheral IV was used for injection of 3 mL/kg of Omnipaque 300 contrast. The region of interest was placed in the left ventricle, and the scan was triggered based on periodic monitoring slices. Extracorporeal membranous oxygenator flow was decreased for a duration of 20 s based on the managing intensivist’s discretion to minimise flow-related artefacts and increase contrast density in the vasculature during acquisition. A retrospective electrocardiogram-gated cardiac CT was acquired with a focused field of view. A retrospective acquisition strategy was selected rather than an ultrafast, high-pitch acquisition to minimise the potential for cannula movement attributable to very rapid transit through the scanner. The patient was placed head-first into the scanner to avoid manipulation of the extracorporeal membranous oxygenator cannulae. The CT demonstrated the location of the extracorporeal membranous oxygenator cannulae (Figure 1( a )) well and showed a contrast leak near the distal anastomosis of the Blalock-Taussig-Thomas shunt (Figure 1( b ) and ( c )). The endeavour involved participation from all clinical teams—intensivists, CT technicians, nursing, radiology, and cardiology—to conduct the scan in a safe and efficient manner. The patient was taken immediately to the operating room for revision of the distal shunt anastomosis. The patient tolerated the procedure well and was subsequently decannulated from the extracorporeal membranous oxygenator in the following days.

Figure 1.

Examples of datasets obtained from paediatric patients with congenital heart disease on VA-ECMO. ( a ) represents the location of the venous and arterial cannulae in a patient with a Blalock-Taussig-Thomas shunt on VA-ECMO. The patient was noted to have a contrast leak adjacent to the distal anastomosis of the BTT shunt ( b ), and ( c ) shows a three-dimensional rendering of the dataset. Cardiac CT helped identify an aneurysmal outpouching along the distal end of a Sano shunt and proximal branch pulmonary artery narrowing in a patient with hypoplastic left heart syndrome ( d ) and ( e ). Cardiac CT has also been utilised for anatomical assessment of ventricular assist device placement in a patient on VA-ECMO ( f ).

Figure 1. Long description

The image contains six different types of medical imaging scans and diagrams related to pediatric patients with congenital heart disease on veno-arterial extracorporeal membranous oxygenator. The first image (a) is a three-dimensional rendering showing the location of the venous and arterial cannulae in a patient with a Blalock-Taussig-Thomas shunt on veno-arterial extracorporeal membranous oxygenator. The second image (b) is a cross-sectional scan highlighting a contrast leak adjacent to the distal anastomosis of the Blalock-Taussig-Thomas shunt. The third image (c) is another three-dimensional rendering of the dataset. The fourth image (d) is a cross-sectional scan showing an aneurysmal outpouching along the distal end of a Sano shunt and proximal branch pulmonary artery narrowing in a patient with hypoplastic left heart syndrome. The fifth image (e) is a three-dimensional rendering illustrating the same anatomical features as in image (d). The sixth image (f) is a three-dimensional rendering used for anatomical assessment of ventricular assist device placement in a patient on veno-arterial extracorporeal membranous oxygenator.

The second case was a neonate with hypoplastic left heart syndrome status post-Norwood procedure and Sano shunt who developed haemodynamic instability in the immediate postoperative period requiring veno-arterial extracorporeal membranous oxygenator. A cardiac CT was performed with retrospective cardiac gating. The region of interest was placed in the systemic right ventricle, and the scan was manually triggered based on periodic monitoring. Pausing extracorporeal membranous oxygenator flow during the injection for less than one minute, administration of 3 mL/kg contrast, and head-first positioning of the scanner allowed for high-resolution visualisation of extracardiac vasculature. The exam showed focal narrowing of the branch pulmonary arteries (worse on the left) with an outpouching along the superior aspect of the distal Sano shunt at the insertion of the pulmonary arteries (Figure 1( d ) and ( e )). The patient was taken back to the operating room for revision of distal Sano shunt anastomosis into the pulmonary arteries. Intraoperatively, there was concern that the surgical packing material around the Sano shunt was causing compression at its distal end. Packing material was removed and the distal Sano was revised, with immediate improvement in cardiac output. The patient was decannulated from the extracorporeal membranous oxygenator on postoperative day 2.

The third case was performed on a 5-year-old patient with a fenestrated extracardiac Fontan with new onset severe ventricular dysfunction. The patient required veno-arterial extracorporeal membranous oxygenator cannulation for refractory decompensated cardiogenic shock. A cardiac CT was requested in preparation for ventricular assist device placement and eventual transplant. The CT acquisition was performed with the patient placed head-first into the scanner. The field of view was focused from the aortic arch to the diaphragm with retrospective cardiac gating and with a region of interest over the left ventricle. Three-dimensional volume rendering of the ribcage and heart with simulated placement of a left ventricular assist device at the apex was performed (Figure 1( f )).

The above cases demonstrate the feasibility of performing high-quality cardiac CT in patients with complex CHD on a veno-arterial extracorporeal membranous oxygenator. A collaborative multi-disciplinary approach among all stakeholders is crucial. Patient selection and assessment of the optimal time for the scan based on clinical condition are important, as pausing the circuit or lowering extracorporeal membranous oxygenator flow can significantly improve the quality of the dataset. At our centre, we use 3 mL/kg of contrast for patients <6 kg and 2 mL/kg for patients greater than or equal to 6 kg. Further dose adjustment was not deemed necessary since we were able to pause or lower extracorporeal membranous oxygenator flows. Techniques such as placement of the patient head-first in the scanner and retrospective acquisition to minimise movement with cannulae further improve the safety of carrying out the scan. Our experience with these patients can serve as a guide for other teams in the future in planning cardiac CTs in patients on mechanical support.