An interactive simulation tool for patient-specific clinical decision support in single-ventricle physiology

作者：Conover Timothy; Hlavacek Anthony M; Migliavacca Francesco; Kung Ethan; Dorfman Adam; Figliola Richard S; Hsia Tain Yen^*

来源：The Journal of Thoracic and Cardiovascular Surgery, 2018, 155(2): 712-721.

DOI：10.1016/j.jtcvs.2017.09.046

• single ventricle physiology
• computer modeling and simulation
• clinical decision support system
• fontan circulation
• Glenn or superior cavopulmonary circulation

Objective: Modeling of single-ventricle circulations has yielded important insights into their unique flow dynamics and physiology. Here we translated a state-of-the-art mathematical model into a patient-specific clinical decision support interactive Web-based simulation tool and show validation for all 3 stages of single-ventricular palliation.
Methods: Via the adoption a validated lumped parameter method, complete cardiovascular-pulmonary circulatory models of all 3 stages of single-ventricle physiology were created within a simulation tool. The closed-loop univentricular heart model includes scaling for growth and respiratory effects, and typical patient-specific parameters are entered through an intuitive user interface. The effects of medical or surgical interventions can be simulated and compared. To validate the simulator, patient parameters were collected from catheterization reports. Four simulator outputs were compared against catheterization findings: pulmonary to systemic flow ratio (Qp:Qs), systemic arterial saturation (SaO2), mean pulmonary arterial pressure (mPAp), and systemic-venous oxygen difference (SaO2-SvO2).
Results: Data from 60 reports were used. Compared with the clinical values, the simulator results were not significantly different in mean Qp: Qs, SaO2, or mPAp (P > .09). There was a statistical but clinically insignificant difference in average SaO-SvO2 (average difference 1%, P < .01). Linear regression analyses revealed a good prediction for each variable (Qp: Qs, R-2 = 0.79; SaO2, R-2 = 0.64; mPAp, R-2 = 0.69; SaO2-SvO2, R-2 = 0.93).
Conclusions: This simulator responds quickly and predicts patient-specific hemodynamics with good clinical accuracy. By predicting postoperative and postintervention hemodynamics in all 3 stages of single-ventricle physiology, the simulator could assist in clinical decision-making, training, and consultation. Continuing model refinement and validation will further its application to the bedside.

• 出版日期2018-2