A Prospective Evaluation of the Effects of Carotid Endarterectomy on Internal Carotid Artery Hemodynamics

Introduction: Decision to pursue endarterectomy (CEA), the gold standard surgical technique for treatment of asymptomatic internal carotid artery (ICA) stenosis, is based primarily on indirect estimation of the percentage of stenosis (PS) via duplex ultrasound (DUS) velocities. However, PS does not account for key hemodynamic indices such as cerebral blood flow (CBF), pressure gradient (PG), wall shear stress (WSS), or oscillatory shear index (OSI). CBF and collateral circulation are well defined risk metrics for stroke and stroke severity. Additionally, WSS and OSI have been associated with both atherosclerotic initiation and plaque rupture/embolization in the carotid vasculature. Moreover, PG is highly correlated with stenosis severity and stroke risk. However, measurements of PG rely on invasive catheter-based techniques and thus are often not obtained. Fractional flow reserve (FFR) has been used to predict hemodynamic significance of coronary artery lesions, however this technique relies on hyperemic conditions and thus is less feasible in the cerebrovascular space. Computational fluid dynamics (CFD) has been used to non-invasively assess PG, along with WSS, OSI, CBF, and collateralization. Our group is conducting an ongoing prospective study utilizing magnetic resonance imaging (MRI) informed CFD to assess differences in plaque morphology and hemodynamics in patients with high grade asymptomatic ICA stenosis. The objective of this work, which is an arm of our prospective study, is focused on evaluating the effects of CEA on hemodynamics pre-and post-operatively in patients with high grade asymptomatic ICA stenosis to better inform surgical decision making in this patient population.

Materials and

Methods: 3D geometric models were created in CRIMSON using a path planning and vessel contouring technique. PC-MRI derived waveforms were prescribed at the CCA inlet and ECA outlet for each patient’s geometric model. Each ICA outlet was coupled to a three-element Windkessel model, whose parameters were iteratively estimated to match the patient’s blood pressure. Blood was modeled as an incompressible Non-Newtonian fluid with a dynamic viscosity of 0.004 kg m−1s−1 and a density of 1,060 kg m-3. Incompressible Navier-Stokes equations were used to solve for velocity and pressures. In the subset of patients who underwent CEA, the same MRI protocol was then repeated and post-operative CFD models were created via the aforementioned methodology. ICA, ECA, and hemispheric CBF were collected from PC-MRI and ICA hemodynamics (PG=Pdistal/Pproximal, WSS, OSI) were collected, at the stenosis, from CFD. The above metrics when then compared pre-operatively and post-operatively in the cohort of patients who underwent CEA. Figure 1 shows a summary of the model inputs and outputs.

Results, Conclusions, and Discussions: Fifteen patients were included (average age = 71.3 years, 53.3% male) and five underwent CEA. Average PSV was 381.3 cm/s and average end diastolic velocity (EDV) was 114.6 cm/s. Average area of the maximum stenosis was 5.6 mm2, average PS was 77.3%. Patients who underwent CEA had higher PSV (439.3 vs 306.6 cm/s, p=0.002) and EDV (132.3 vs 90.6 cm/s, p=0.014) than the non-CEA cohort. Neither area of maximum stenosis nor PS differed between two. Table 1 depicts flow and hemodynamics measurements in CEA and non-CEA patients. There was a large variability in average PG and TAWSS across the cohort. However, there were no differences in hemodynamics between the CEA and no CEA patients. In patients who had CEA, only 2 (40.0%) had an “at-risk” PG (PG < 0.88). and a wider range of TAWSS was seen in this cohort. Table 2 depicts pre-operative and post-operative hemodynamic information for the three patients who had complete measurements. Post-operatively, ipsilateral ICA flow increased, however there was no difference in ipsilateral ECA flow or ipsilateral CBF. There was a trend for higher PG and lower TAWSS post-operatively, although these did not meet statistical significance given the large variability in the pre-operative cohort. However, post-operatively OSI increased (Figure 2).

This MRI-informed CFD protocol provides a thorough non-invasive assessment of flow and hemodynamics in ICA stenosis patients. Although decision for CEA is based on DUS velocity (which is in-line with current guidelines), patients with severe ICA stenosis have wide variability in hemodynamics. Patients with similar PSVs may have quite different PGs across the lesion. Our results showed that only a subset of patients undergoing CEA had “at-risk” PG. Although CEA improves ICA flow, no change in WSS or PG was observed. Moreover, over-sizing of the diseased vessel during the patch-technique in CEA may contribute to elevated OSI and thus unfavorable post-operative hemodynamics.

Although CEA leads to improved flow, it may lead to poor post-operative hemodynamics. Given the variability in the hemodynamics in ICA stenosis patients, they may have different stroke risk profiles. Improved metrics of determining need for CEA are warranted.

Acknowledgements (Optional):