Robust Cardiac-gated MRI using Ultrasound Sensors

点击翻译按钮获取中文摘要

Project summary For cardiac MRI pulse sequences to function properly, a signal or trigger must typically be obtained to help synchronize the acquisition process with the cardiac motion. As a result of such synchronization, reconstructed images represent a given cardiac phase, mostly free of cardiac-related motion. Improper detection of such triggers would cause data from different cardiac phases to mix into a same image, leading to motion artifacts and loss of image quality. Most commonly, electrocardiogram (ECG) leads are employed to detect the electrical activity of the heart and to generate such triggers, allowing the MRI acquisition and the heart motion to remain in sync. Besides a small risk for skin burns, one of the main challenges with using ECG leads for cardiac MRI comes from the magnetohydrodynamic (MHD) effect, which tends to distort the ECG signal. Because of the Lorenz force, and in a manner similar to the Hall effect, the positive and negative ions in blood tend to curve in different directions as they move through a magnetic field, creating a charge separation and an electric field that tends to corrupts the ECG signal. Because the aorta carries a large amount of fast-moving blood, it tends to be the main source of corrupting electrical fields, and the problem becomes more significant at higher field strength. Scanner operation, especially gradient switching, can also further corrupt ECG signals. Real-time cardiac imaging, as opposed to cardiac-gated imaging, does not necessarily require reliable triggers, but typically suffers from lower spatial and temporal resolution. Presumably for this reason, cardiac- gated imaging represents the overwhelming bulk of clinical cardiac MR (CMR) scans. Alternatives to ECG leads for CMR gating include pulse oximetry, which typically takes the form of a sensor clipped around a finger. While it has the advantage of being insensitive to the MHD effect, pulse oximetry and other forms of peripheral gating detect heartbeats with a delay, the time for the pressure wave to reach the detection point, as opposed to ECG detection that is more direct and immediate. With a perfectly regular heart, such delay would be of little to no consequence, but cardiac patients often have irregular heartbeats and immediate detection of a contraction allows the MRI acquisition to react in a more sensible way, with better outcomes on image quality. The present project introduces an alternative to ECG monitoring for cardiac-gated MRI which detects cardiac motion directly. An ultrasound-based sensor is proposed, along with 3D-printed capsule for quick application to the skin and realistic clinical workflow. Before the patient table slides into the scanner, as the magnetic field is weak and the MHD effect is minimal, we will validate our approach with ECG detection as the reference standard. As the patient goes into the scanner, the ECG trace will be degraded but our MR- compatible sensor will remain unaffected. With patients in the scanner, we will demonstrate an improved detection of cardiac triggers and improved image quality using our sensors as compared to using ECG leads.