Cross-Comparison of 4-Dimensional Flow Magnetic Resonance Imaging and Intraoperative Middle Cerebral Artery Pressure Measurements Before and After Superficial Temporal Artery-Middle Cerebral Artery Bypass Surgery

BACKGROUND

The hemodynamic changes after superficial temporal artery (STA) to middle cerebral artery (MCA) bypass surgery are unclear.

OBJECTIVE

To clarify the hemodynamics by comparing flow parameters obtained by 4-dimensional (4D) flow magnetic resonance imaging (MRI) and intraoperative MCA pressure measurement.

METHODS

We recruited 23 patients who underwent STA-MCA bypass surgery for internal carotid artery (ICA) or MCA stenosis. We monitored intraoperative MCA, STA, and radial artery (RA) pressure. All patients underwent 4D flow MRI preoperatively and 3 wk after surgery to quantify the blood flow volume (BFV) of the ipsilateral ICA (BFViICA), contralateral ICA (BFVcICA), basilar artery (BFVBA), ipsilateral STA (BFViSTA), and contralateral STA (BFVcSTA). The sum of intracranial BFV was defined as BFVtotal. We compared BFV parameters and intraoperative pressure.

RESULTS

BFViSTA significantly increased after surgery (P < .001). BFViICA and BFVBA significantly decreased after surgery (BFViICAP = .005; BFVBAP = .02). No significant difference was observed between BFVcICA before and after surgery. As a result, BFVtotal postoperatively increased by 6.8%; however, no significant difference was observed. Flow direction at M1 changed from antegrade to unclear after surgery in 5 patients. Intraoperative MCA pressure and MCA/RA pressure ratio significantly increased after surgery (P < .001). We found a stronger positive correlation between MCA pressure increase ratio and BFVtotal increase ratio in patients with lower pre-MCA pressure (r = 0.907, P < .001).

CONCLUSION

The visual and quantitative assessment of 4D flow MRI revealed that intracranial blood flow changes complementarily after STA-MCA bypass surgery. 4D flow MRI may detect the improvement of cerebral perfusion pressure.

Evidence for non-Newtonian behavior of intracranial blood flow from Doppler ultrasonography measurements

Computational fluid dynamics (CFD) studies of intracranial hemodynamics often use Newtonian viscosity model to close the shear rate term in the Navier-Stokes equation. This is based on a commonly accepted hypothesis which state that non-Newtonian effects can be neglected in intracranial blood flow. This study aims to examine the validity of such hypothesis to guide future CFD studies of intracranial hemodynamics. Doppler ultrasonography (DUS) measurements of systolic and diastolic vessel diameter and blood velocity were conducted on 16 subjects (mean age 50.6). The measurements were conducted on the internal carotid (ICA), middle cerebral (MCA), and anterior communicating (AComA) arteries. Systolic and diastolic wall shear stress (WSS) values were calculated via the Hagen-Poiseuille exact solution using Newtonian and three different non-Newtonian models: namely Carreau, power-law and Herschel-Bulkley models. The Weissenberg-Rabinowitsch correction for blood shear-thinning viscosity was applied to the non-Newtonian models. The error percentage between the two sets of models was calculated and discussed. The Newtonian hypothesis was tested statistically and discussed using paired t tests. Significant differences (P < 0.0001) were found between the Newtonian and non-Newtonian WSS in ICA. In MCA and AComA, similar differences were found except in the systole and diastole for the Herschel-Bulkley and power-law models (P = 0.0669, P = 0.7298), respectively. The error between the Newtonian and non-Newtonian models ranged from - 27 to 30% (0.2 to 2.2 Pa). These values could affect the physical interpretation of IA CFD studies. Evidence suggests that the Newtonian assumption may be inappropriate to investigate intracranial hemodynamics. Graphical abstract The WSS estimation error resulting from using the Newtonian assumption compared to three non-Newtonian models for ICA, MCA, and AComA in systole and diastole conditions, based on TCCD measurements of 16 subjects. The error due to the Newtonian assumption ranged from 0.2 to 2.2 Pa (- 27 to 30%). These values could affect the physical interpretation of IA CFD studies.