(E-192) Optimizing the Position of a Wearable Ultrasound Device for Measuring Cerebral Blood Flow

An aneurysm or hemorrhage can seriously harm the vasculature, especially in the blood vessels that supply blood to the brain. About 10 in 100,000 adults experience these conditions annually in the US, and survivors are often hospitalized. During this time, the patient is at high risk of vasospasm (a narrowing of arteries). Medical professionals check the blood flow of these patients to monitor for vasospasm. However, monitoring may occur less often than optimally; often, only once per day.

Medical professionals use transcranial Doppler ultrasound (TCD) for detecting and imaging blood flow. TCD is a noninvasive method of detecting the blood flow velocity within the cerebral vasculature. It enables the calculation of parameters such as the Lindegaard Ratio. The Lindegaard Ratio compares blood velocity in the middle cerebral artery (MCA) to the velocity of blood in the internal carotid artery (ICA). This parameter can provide insight into how at-risk a given patient is for vasospasm. The long-term goal of this research is to design an ultrasound patch for use in a TCD system to detect blood flow in relevant arteries in the brain. The system would continuously monitor blood flow; therefore, a medical professional does not need to be physically present when a patient experiences a vasospasm.

TCD was done on ten volunteers using a Doppler BoxX TCD. The subjects were brought into the lab and had a transducer placed in three locations: 1) on their temple to record the MCA velocity, 2) under the jawbone (i.e., submandibular window) to record ICA velocity, and 3) along the neck inferiorly to #2, which was also used for ICA velocity. Two photos were taken from a lateral view, and two from a frontal view to evaluate the depth, placement, and angle of the transducer when imaging the neck.

Blood flow velocities and the Lindegaard ratio were calculated from TCD data. Two Lindegaard ratios were calculated per person. One Lindegaard ratio was calculated using the blood velocity in the ICA seen when imaging from the submandibular window. The other was calculated by imaging along the neck using a TCD transducer.

Two angles were estimated from the pictures using the Image J software: θ₁ is the angle measured from the lateral view, and θ₂ is the angle measured from the frontal view. A qualitative model of signal quality as a function of the angle θ₁ was made using the photos.

The Lindegaard ratios and ICA velocities were compared using a two-tailed paired t-test (α=0.05).

The comparison of the velocity of Lindegaard ratios showed no significant difference between patients (p = 0.54 for the velocity and p = 0.48 for the Lindegaard), suggesting that imaging along the neck is a potential alternative to imaging in the submandibular window for calculating the Lindegaard Ratio. The velocities and Lindegaard Ratios of the two imaging locations were correlated (R²=0.79 for the velocity and R²=0.55 for the Lindegaard) and a regression model was obtained.

The ICA was found between 25.5 mm and 30 mm. The average θ₁ angle was 50.6° with a standard deviation of 6.8°. The average θ₂ angle was 41.6° with a standard deviation of 11.8°.  The angle θ₁ showed greater consistency versus θ₂. The variation is due to θ₂ being dependent on anatomical differences between volunteers. Therefore, θ₂ is a function of the patch location. The θ₁ angle had a more substantial effect on the signal to noise ratio. This angle influenced the next iteration of wearable patches. The piezo elements were placed at an angle of 45° because it falls within the confidence interval of θ_1. The model showed that as θ₁ increases, the signal quality decreases, leading to more noise in the signal. The experiment yielded conflicting results about the effects of a decrease in θ₁. As θ₁ decreased, the signal quality improved in two volunteers, but the signal quality had a negligible change in two other volunteers. Therefor signal quality cannot be guaranteed outside of the confidence interval for θ₁.

There are two ways to use the velocity calculated. The first is a substitution of the velocity of blood in the ICA found in the neck for the velocity of blood found in the ICA at the submandibular window in the Lindegaard Ratio. The alternative is to use the ICA velocity found in the neck and the regression model found prior to estimate the ICA velocity seen in the submandibular region. The ICA velocity found in the neck could be used to estimate the ICA velocity seen in the submandibular region. The estimated velocity can subsequently be used to estimate the Lindegaard Ratio.