Magnetic resonance imaging (MRI) with self-navigation and self-registration

摘要

Three-dimensional (3D) projections of nuclear magnetic resonance (NMR) signals are acquired from a liver experiencing NMR in response to a 3D multi-echo non-Cartesian pulse sequence. The projections are reconstructed into two sets of images having different resolutions. Bins associated with the different positions to which the liver moves during respiration are identified in lower resolution images, and then higher resolution images are binned into the position dependent bins based on navigator data in the lower resolution images. A combined image for a bin is made from images located in the bin and then registered to a reference image. An overall combined image is made by summing the combined bin images. Quantized data for a contrast agent concentration in the liver is produced using signal intensity in the overall combined image. The quantized value may describe a liver perfusion parameter. A diagnosis may be made from the quantized value.

主权项

1. A method, comprising:
controlling a magnetic resonance imaging (MRI) apparatus to acquire a set of three dimensional (3D) projections from a liver that moves in a superior/inferior direction due to respiration, where the set of 3D projections are acquired using a 3D multi-echo non-Cartesian acquisition; controlling the MRI apparatus to reconstruct members of the set of 3D projections in the superior/inferior direction into a corresponding set of first 3D images using a compressed sensing reconstruction, where members of the set of first 3D images have a first resolution; controlling the MRI apparatus to reconstruct members of the set of 3D projections in plane into a corresponding set of second 3D images using a compressed sensing reconstruction, where members of the set of second 3D images have a second resolution that is less than the first resolution; identifying one or more in-plane navigator voxels in members of the set of second 3D images; producing a plot of the signal intensity of the one or more in-plane navigator voxels in two or more of the set of second 3D images as a function of time; identifying a motion time course for the liver based, at least in part, on the plot of the signal intensity; partitioning the motion time course into a set of position-dependent intervals; associating a member of the set of first 3D images with a member of the set of position-dependent intervals based, at least in part, on the motion time course and the one or more in-plane navigator voxels; producing a 3D image for an interval from the set of position-dependent intervals using members of the set of first 3D images associated with the interval, where the 3D image for the interval has a third resolution; using non-linear registration to register the 3D image for the interval to a 3D reference image associated with a reference interval in the set of position-dependent intervals; and producing a combined 3D image from 3D images associated with two or more different intervals, where the combined 3D image has a fourth resolution.