A novel noninvasive assessment of portal pressure from computational biofluid mechanics in patients with portal hypertension.

BACKGROUND AND AIMS

To introduce and assess a novel method for portal pressure measurement based on biofluid mechanics in portal hypertensive patients undergoing surgery.

METHODS

The research was a multi-center, retrospective study, conducted on patients who underwent surgery and measurement of free portal pressure (FPP). There were 118 patients included and 21 patients excluded due to the failure or poor results of Doppler ultrasound, and 97 patients were screened. We used patients' CT images, Doppler ultrasound results of the portal system, blood density and viscosity to reconstruct their portal system and simulate its internal blood flow. According to the patient's physical property, geometry, and boundary conditions, the Navier-Stokes equations were solved by FLUENT software, and virtual free portal pressure (vFPP) was calculated. Finally, the Bland-Altman Limits of Agreement, intraclass correlation coefficient (ICC), and the Lin's concordance correlation coefficient were performed to evaluate the numerical correlation between the vFPP and FPP.

RESULTS

All patients enrolled in this study underwent the surgery, and the FPP of patients was measured during the surgery, with a mean FPP of 22.8 ± 3.3 mmHg (range: 13-33 mmHg). Meanwhile, according to computational biofluid mechanics, all patients' vFPP was calculated. Then, we further explored whether there was a close relationship between vFPP and FPP in the whole population. For the analysis of Bland-Altman Limits of Agreement, the mean value of difference was - 0.1569 (95% CI: - 0.4305 to 0.1167); lower limit of agreement: - 2.8176 (95% CI: - 3.2868 to - 2.3484); upper limit of agreement: 2.5038 (95% CI: 2.0346 to 2.9730). The ICC was 0.9215 (95% CI: 0.8848 to 0.9468). Furthermore, the Lin's concordance correlation coefficient showed a numerical correlation between the vFPP and FPP, which was 0.9205 (95% CI: 0.8840 to 0.9459). All these results confirmed that our vFPP model could provide an accurate prediction of FPP in patients.

CONCLUSIONS

The vFPP of patients calculated by computational biofluid mechanics was significantly correlated with the FPP of portal hypertensive patients, which would be a novel, non-invasive, and accurate method for the assessment of portal pressure in surgical patients.