A Personalized Pulmonary Circulation Model to Non-Invasively Calculate Fractional Flow Reserve for Artery Stenosis Detection

Autor: Zhi-Cheng Jing, Yu-Ping Zhou, Xiji Xu, Heye Zhang, Zhengyang Chen, Dhanjoo N. Ghista, Xiujian Liu, Tao Wu, Xin Jiang
Rok vydání: 2022
Předmět:
Zdroj: IEEE Transactions on Biomedical Engineering. 69:1435-1448
ISSN: 1558-2531
0018-9294
DOI: 10.1109/tbme.2021.3119188
Popis: Objective: Fractional Flow Reserve (FFR) is regarded as a fundamental index to assess pulmonary artery stenosis. The application of FFR can increase the accuracy of detection of pulmonary artery stenosis. However, the invasive examination may carry a number of physiological risks for patients. Therefore, we propose a personalized pulmonary circulation model to non- invasively calculate FFR of pulmonary artery stenosis. Method- ology: We employed a personalized pulmonary circulation model to non-invasively calculate FFR using only computed tomography angiogram (CTA) data. This model combined boundary conditions estimation and 3D pulmonary artery morphology reconstruction for CFD simulation. First, we obtained patient-specific boundary conditions by adapting the right ventricle stroke volume and main pulmonary artery pressure feature points (systolic, diastolic, and mean pressure). Secondly, the 3D pulmonary artery morphology was reconstructed by threshold segmentation. The CFD simulation was then performed to obtain pressure distribution in the entire pulmonary artery. Finally, the FFR in pulmonary artery stenoses was calculated as the ratio of distal pressure and proximal pres- sure. Results: To validate our model, we compared the calculated FFR with measured FFR by pressure guide wires examination of 8 patients. The FFR calculated by our model showed a good agreement with measured FFR by pressure guide wires exami- nation. The average accuracy rate was 91.41%. Conclusion: The proposed personalized pulmonary model is capable of reasonably non-invasively calculating FFR with sufficient accuracy. Significance: FFR calculated in our model may contribute to non-invasive detection of pulmonary artery stenosis and to the assessment of invasive interventions.
Databáze: OpenAIRE