Digital Twins of Patients with Complex Univentricular Physiology

Background: Several congenital heart defects result in the need for univentricular palliation, as no biventricular repair can be achieved either by intervention or surgery. In those patients, the superior and inferior vena cava are connected to the pulmonary artery via staged surgeries, facilitating the so-called total cavopulmonary connection (TCPC).

Due to the complex and heterogeneous patient-specific morphology of the TCPC, as well as the often intensive treatment history of a given patient, e.g., implantation of metallic stents, image-based diagnosis via CT or MRI is severely constrained. Similarly, generalized treatment strategies are often not feasible due to the heterogeneity in both anatomy and function. Thus, personalized strategies for diagnosis as well as for treatment-planning are required for improving the outcome of this patient group.

Method: By means of image processing, the three-dimensional, anatomical representation of the TCPC can be reconstructed from different imaging sources, as for example MRI, CT, and angiography. This anatomical representation alone allows to mitigate imaging constraints, such as uneven distribution of the contrast agent during CT, or susceptibility artifacts caused by metallic stents. Furthermore, the anatomical model can be enhanced by functional information using computational fluid dynamics. This approach allows to assess the entire patient-specific hemodynamics, including parameters such as pressure gradients, the power loss, or the distribution of hepatic blood toward the pulmonary arteries. Once these digital twins are parameterized and validated, they can be altered virtually to mimic different treatment strategies, such as interventional implantation of a covered stent or surgical implantation of a vascular graft. Then, the immediate functional outcome of different treatment strategies can be assessed and compared against each other to identify the optimal treatment for a given patient.

Conclusion: Digital twins are a promising approach to overcome the obstacles imposed by the heterogeneity found in patients with univentricular palliation. The anatomical models alone, especially when superimposed to image data, provide additional information for treatment planning, as the entire complex anatomy can be visualized, and constraints imposed by surrounding tissue can be assessed. The outcome prediction of different treatment strategies holds the promise to provide additional objective information, allowing to better stratify treatment risks and potential outcomes on a patient-specific level.

Publication History

Article published online:
28 January 2023