Optimal Sensor Selection for Motion-Corrected Supine Breast MRI With a Wearable Coil.

Abstract

Supine breast MRI offers several advantages over conventional prone MRI but requires correction of motion artifacts. This study aimed to identify the optimal combination of motion sensors for achieving the best motion correction quality using a wearable radio frequency coil ("BraCoil"). T2-weighted (T2w) 2D and T1-weighted (T1w) 3D sequences were acquired in 10 volunteers and 17 exams. Images were reconstructed using the GRICS motion correction algorithm, using data from a respiratory belt or MRI-compatible accelerometers. The motion models were evaluated by comparing the predicted motion to dynamic MRI. The resulting images were also assessed using both quantitative metrics and radiological scoring. Additionally, the accuracy of motion models calculated from high-resolution 2D T2w images was compared to those derived from low-resolution 3D T1w images. Motion predicted from 2D T2w images matched the dynamic MRI significantly better than that predicted from low-resolution 3D T1w images. Radiological scoring showed that using all accelerometers significantly improved image quality compared to using the belt. A single accelerometer explained only 45-70% of the total sensor data, and combining all sensors led to superior correction overall. No optimal sensor position was identified. The motion model calculated from 2D T2w images, using the data from all accelerometers, provides the best motion correction quality for supine breast MRI using a wearable coil. Selecting the appropriate physiological sensors minimizes motion artifacts and can potentially enhance diagnostic accuracy.