Numerical prediction of portal hypertension by a hydrodynamic blood flow model combing with the fractal theory

Portal hypertension (PH) can cause a series of complications, therefore, early prediction of PH is important. Traditional diagnostic methods are harmful to the human body, while other non-invasive methods are inaccurate and lack physical meaning. Combining various fractal theories and flow laws, we establish a complete portal system blood flow model from the Computed Tomography (CT) and angiography images. The portal vein pressure (PP) is obtained by the flow rate data from the Doppler ultrasound and the pressure-velocity rela-tionship is established by the model. Three normal participants and 12 patients with portal hypertension were divided into three groups. For the three normal participants (Group A), their mean PP calculated by the model is 1752 Pa, falling into the normal range of PP. The mean PP of three patients with portal vein thrombosis (Group B) is 2357 Pa; and for the 9 patients with cirrhosis (Group C), their mean PP is 2915 Pa. These results validate the classification performance of the model. Moreover, the blood flow model can give early warning parameters of the corresponding portal vein trunk and portal vein microtubules for thrombosis and liver cirrhosis. This model presents the complete process of blood flow from sinusoids to the portal vein, adapts to the diagnosis of portal hypertension of thrombosis and liver cirrhosis, and provides a new method for noninvasive portal vein pressure detection from the perspective of biomechanics.

Automated summary

By combining various fractal theories and flow laws, we established a blood flow model based on computed tomography and angiography images, which can accurately predict portal hypertension. The model successfully classified patients with portal vein thrombosis and cirrhosis, and provided early warning parameters for these conditions.