General Electric Company

The Medical College of Wisconsin, Inc.

Leland Stanford Junior University

Yale University

New York Society for the Relief of the Ruptured and Crippled, Maintaining the Hospital for Special Surgery

New York Society for the Ruptured and Crippled Maintaining the Hospital for Special Surgery

Kevin Matthew Koch

Andrew S Nencka

Suryanarayanan Sivaran Kaushik

Kevin Franklin King

Hollis Potter

Matthew F Koff

Dattesh Dayanand Shanbhag

Rakesh Mullick

Graeme Colin McKinnon

Brian A Hargreaves

Richard Scott Hinks

Florian Wiesinger

Dawei Gui

David Thomas Gering

Sheshadri Thiruvenkadam

Reed Frederick Busse

Sudhanya Chatterjee

Robert Donald Peters

Weitian Chen

Zachary Slavens

Eric M Printz

Lutfi Tugan Muftuler

Gaspar Delso

Robin De Graaf

Juan Liu

Casey Anderson

Diego Hernando Arribas

Ali Bakhshinejad

Robin A Karr

Bradley J Swearingen

Andrew Scott Nencka

Suryanarayanan Sivaram Kaushik

Systems and methods for training and implementing a machine learning algorithm to generate feature maps depicting spatial patterns of features associated with osteolysis, synovitis, or both. MRI data, including multispectral imaging data, are input to the trained machine learning algorithm to generate the feature maps, which may indicate features such as a location and probability of a pathology classification, a severity of synovitis, a type of synovitis, a synovial membrane thickness, and other features associated with osteolysis or synovitis. In some implementations, synovial anatomy are segmented in the MRI data before inputting the MRI data to the machine learning algorithm. These segmented MRI data may be generated using another trained machine learning algorithm.

Neural network classification of osteolysis and synovitis near metal implants

Aneurysms are classified and quantitatively analyzed based on medical image data acquired from a subject. In general, one or more algorithms are implemented to automatically classify, or otherwise diagnose, and measure aneurysms and their change over time. These algorithms make use of artificial intelligence and deep learning to develop quantitative analytics that can be consolidated into diagnostic reports.

Systems and methods for generating classifying and quantitative analysis reports of aneurysms from medical image data

Systems and methods for quantitative susceptibility mapping ('QSM') using magnetic resonance imaging ('MRI') are described. Localized magnetic field information is used when performing the inversion to compute quantitative susceptibility maps. The localized magnetic field information can include multi-resolution subvolumes obtained by segmenting, or dividing, a field shift map. In some instances, a trained machine learning algorithm, such as a trained neural network, can be implemented to convert the localized magnetic field information into quantitative susceptibility data. These local susceptibility maps can be combined to form a composite quantitative susceptibility map of the imaging volume.

Multi-resolution quantitative susceptibility mapping with magnetic resonance imaging

Described here are systems and methods for using a magnetic resonance imaging ('MRI') system to estimate parameters of spectral profiles contained in multispectral data acquired using multispectral imaging ('MSI') techniques, such as MAVRIC. These spectral profile parameters are reliably extracted using an iterative perturbation theory technique and utilized in a number of different applications, including fat suppression, artifact correction, and providing accelerated data acquisitions.

Voxelwise spectral profile modeling for use in multispectral magnetic resonance imaging

A method for correcting one or more artifacts within a multi-spectral magnetic resonance image is provided. The method includes acquiring a plurality of spectral bins each including a plurality of voxels and corresponding to a different frequency of MR signals emitted by an imaged object. The plurality of voxels of each spectral bin correspond to the frequency of the spectral bin so as to define a spatial coverage of the spectral bin. The method further includes expanding each spectral bin by increasing the spatial coverage of the spectral bin, and generating the multi-spectral magnetic resonance image based at least in part on the expanded spectral bins.

System and method for correcting one or more artifacts within a multi-spectral magnetic resonance image

Systems and methods for quantitative susceptibility mapping ('QSM') using magnetic resonance imaging ('MRf') and a localized processing technique are described. A field-shift map is processed based on localized regions of local field perturbations. These localized field-shift regions are processed using established QSM algorithms, or using direct dipole inversion techniques, to compute regional susceptibility distributions from the localized field shift information. When the localized regions correspond to subvolumes of the field-shift map, local susceptibility maps can be generated and combined to form a composite quantitative susceptibility map. By computing regional susceptibility distributions based on localized field-shift information, residual streaking artifacts in the susceptibility map are constrained to the individual volumes from which they originate, thereby eliminating their propagation through the image.

System and method for localized processing of quantitative susceptibility maps in magnetic resonance imaging

Systems and methods are provided for performing a calibration 'pre-scan' prior to acquiring data using a magnetic resonance imaging ('MRI') system performing a multi-spectral imaging (&#x201c;MSI&#x201d;) acquisition. Information from the calibration scan is used to optimize the scanning and data collection during the MSI scan. As a result, scan times and motion artifacts are reduced. In addition, image resolution can also be increased, thereby improving image quality. As one example, the MSI acquisition can be a MAVRIC acquisition. In general, the calibration data is used to determine the minimum number of spectral bins required to achieve acceptable image quality near a specific metallic implant or device.

Systems and methods for calibrated multi-spectral magnetic resonance imaging

Systems and methods for performing diffusion-weighted multi-spectral imaging ('MS!') with a magnetic resonance imaging ('MRI') system are provided, Diffusion-weighted images can thus be acquired from a subject in which a metallic object, such as an implant or other device, is present. In general, a two-dimensional or three-dimensional diffusion-weighted PROPELLER acquisition is performed to acquire data from multiple different spectral bins. Images from the spectral bins are reconstructed and combined to form diffusion-weighted composite images. Non-CPMG phase-cycling and split-blade PROPELLER techniques are combined with PROPELLER MSI metal artifact mitigation principles to this end.

Systems and methods for diffusion-weighted multi-spectral magnetic resonance imaging

A method includes: accessing MRI data acquired from a joint area, the MRI data including a series of spatially mapped spectral data points; generating MRI images of the joint area; receiving information encoding a region of interest that encompasses a suspected metal particle deposition area over at least one of the MRI images; constructing magnetic field maps using the MRI data, each representing off-resonance frequency shifts over the joint area; removing a background of off-resonance field inhomogeneity from the magnetic field map such that the region of interest is free from off-resonance field inhomogeneity; identifying clusters from the magnetic field maps with the background of off-resonance field inhomogeneity removed, the clusters defined over a first dimension of offset frequencies and a second dimension of cluster volumes; and computing a quantitative metric by combining information from the identified clusters according to both the first dimension and the second dimension.

System and apparatus for assessing metallosis using magnetic resonance imaging

A magnetic resonance imaging (MRI) system can include a magnetic resonance imaging (MRI) scanner, having a plurality of radio frequency (RF) receivers, and a processor. The MRI scanner can perform a full field of view (fFOV) scan on an anatomy area including an implant to acquire first multi-spectral MRI data associated with a plurality of frequency bins. The processor can generate, for each pair of a single RF receiver and a single frequency bin, a respective spectral sensitivity map using at least a portion of the fFOV multi-spectral MRI data. The MRI scanner can perform a reduced FOV (rFOV) scan to acquire second multi-spectral MRI data associated with the plurality of frequency bins. The processor can reconstruct one or more MRI images according to the rFOV using the rFOV multi-spectral MRI data and the spectral sensitivity maps.

Spectral bin un-aliasing for reduced field-of-view magnetic resonance imaging near metal implants

An imaging system and method are disclosed. An MR image and measured B0 field map of a target volume in a subject are reconstructed, where the MR image includes one or more bright and/or dark regions. One or more distinctive constituent materials corresponding to the bright regions are identified. Each dark region is iteratively labeled as one or more ambiguous constituent materials. Susceptibility values corresponding to each distinctive and iteratively labeled ambiguous constituent material is assigned. A simulated B0 field map is iteratively generated based on the assigned susceptibility values. A similarity metric is determined between the measured and simulated B0 field maps. Constituent materials are identified in the dark regions based on the similarity metric to ascertain corresponding susceptibility values. The MRI data is corrected based on the assigned and ascertained susceptibility values. A diagnostic assessment of the target volume is determined based on the corrected MRI data.

Method and system for improved classification of constituent materials

A method for attenuation correcting a PET image of a target includes locating a radiopaque structure by MRI scan of the target; fitting a model of the radiopaque structure to the MRI scan image; and correcting attenuation of the PET image based on the fitted model.

Method, apparatus, and article for pet attenuation correction utilizing MRI

A method for acquiring 3D multispectral MRI of a target includes scanning a spectrum of spectral windows with an MRI scanner, wherein each spectral window of the spectrum defines a continuously-differentiable distribution of frequencies around a scan frequency and adjacent scan frequencies are spaced apart by substantially uniform frequency offsets such that adjacent spectral windows substantially uniformly overlap, wherein selected adjacent spectral windows are scanned in consecutive passes, and nearest neighbor spectral windows within each pass are scanned at a maximum temporal spacing within the pass.

Method, apparatus, and article for frequency-overlapped 3-D multispectral magnetic resonance images

An apparatus for detecting and repairing a ring artifact in a multi-spectral magnetic resonance image includes an image processor, which is configured to obtain an off-resonance magnetic field map and a deblurred composite image, to calculate a spatial gradient of the image based on the magnetic field map, to kernel search the spatial gradient, to mask the image, based on the kernel search, in order to identify voxels affected by a ring artifact, and to apply a filter in order to smooth intensities of the voxels identified by the image mask.

Method and apparatus for ring artifact repair of magnetic resonance images

A system and method for generating MR phase contrast images near metal include an MRI apparatus that includes an MRI system having a plurality of gradient coils and an RF transceiver system and an RF switch controlled by a pulse module to transmit RF signals to an RF coil assembly. The MRI apparatus also includes a computer programmed to acquire a plurality of three-dimensional (3D) MR data sets and to generate a plurality of frequency images based on the plurality of 3D MR data sets. Each 3D MR data set is acquired using a central transmit frequency and a central receive frequency set to an offset frequency value that is distinct for each 3D MR data set. The computer is also programmed to convert the plurality of frequency images to a plurality of time domain images and to generate a phase image based on the plurality of time domain images.

Growing community of inventors

Milwaukee, WI, United States of America

Kevin Matthew Koch