Children born with severe heart malformations may need to be converted to a Fontan circulation with a total cavo pulmonary connection (TCPC), where all venous blood is routed passively to the lungs. The pulmonary blood flow distribution is however hard to predict. An unbalanced pulmonary blood flow distribution may result in clinical complications and the need for post-surgical invasive interventions. Pulmonary blood distribution can be predicted using patient-specific computational fluid dynamic (CFD) simulations based on CMR imaging. This is however a resource intensive process that few centers can manage independently. CMR acquisition often requires that the child is under anesthesia during imaging, and the CFD simulations may require complex software managed by engineers at external locations.
The purpose of the study is to develop a streamlined process to simulate the TCPC pulmonary blood distribution, allowing surgeons to evaluate different designs before surgery. We aim to provide a clinically integrated in-center process for performing CMR under mild sedation, without anesthesia or contrast agent and deliver an accurate CMR-based predictive pre-surgical simulation, both on the same day before surgery.
Two patients (age 2.5 and 3years, 14 and 12kg) in the second stage of Fontan implementation ("Glenn") and planned for TCPC surgery were prospectively included in the study. The patients were mildly sedated with Dexmedetomidine, enabling a maximum CMR-time of <60 minutes. A contrast-free CMR acquisition of anatomy (SSFP) and phase-contrast (PC) flow was used to build a patient-specific flow simulation model using commercial CFD software. The model was modified according to the surgeon’s plan for the TCPC surgery, and post-surgical pulmonary flow was calculated.
High quality pre-surgical CMR acquisitions of anatomy and flow was obtained. Simulations of the pre-surgical left/right pulmonary blood distribution (Figure A) were 48/52% and 32/68% and PC flow 48/52% and 39/61% respectively. Processing of the CMR data and performing all simulations, including the post-surgical TCPC (Figure B), could be completed in-center within 4-5 hours after the CMR examination, including provisioning for overnight 3D-printing of the Glenn and TCPC vessels. All results were ready by the start of the pre-surgical conference in the morning of the following day.
We have demonstrated that a same-day CMR and pre-operative predictive simulation of Fontan pulmonary blood flow can be achieved in-center, without anesthesia or contrast agent required during the CMR acquisition and without super computer access. This shows how CMR in combination with computational modelling can be implemented in clinical routine for improving surgical planning without extensive resources.