Koninklijke Philips Corporation N.v.

Philips Intellectual Property & Standards Gmbh

Ingmar Graesslin

Peter Boernert

Peter Vernickel

Ulrich Wolfgang Katscher

Christoph G Leussler

Paul Royston Harvey

Kay Nehrke

Joachim Schmidt

Bernhard Gleich

Johannes Adrianus Overweg

Thomas Erik Amthor

Christian Findeklee

Arne Ewald

Ferdinand Schweser

Rafael Wiemker

Steffen Weiss

Johan Samuel Van Den Brink

Sascha Krueger

Oliver Lips

Holger Eggers

Eberhard Sebastian Hansis

Falk Uhlemann

Thomas Netsch

Jochen Keupp

Hanno Heyke Homann

Peter Koken

Axel Saalbach

Ingo Schmale

Karsten Sommer

Tim Philipp Harder

Jörn Borgert

Tom Brosch

Tobias Richard Schaeffter

Rudolf Mathias Johannes Nicolaas Lamerichs

Tobias Ratko Voigt

Tanja Nordhoff

Wilhelmus Reinerius Maria Mens

Nick Flaeschner

Sven Biederer

Olaf Dössel

Liqin Wang

Yajing Zhang

Chenguang Zhao

Graham Michael Place

Cornelis Jacobus Hendrikus Blom

Sebastian Alfred Seitz

Nick Flaschner

The present disclosure relates to a medical imaging method, comprising: receiving () a set of subject parameters descriptive of a subject; in response to inputting () the set of subject parameters into a trained deep neural network, DNN, receiving () from the trained DNN a predicted task; presenting the task to the subject; controlling () an MRI system () for acquiring fMRI data from the subject in response to the predicted task performed by the subject during the acquisition.

fMRI task settings with machine learning

The invention provides for a medical imaging system () comprising a memory () storing machine executable instructions () and a configured artificial neural network (). The medical imaging system further comprises a processor () configured for controlling the medical imaging system. Execution of the machine executable instructions causes the processor to receive () magnetic resonance imaging data (), wherein the magnetic resonance imaging data is BOLD functional magnetic resonance imaging data descriptive of a time dependent BOLD signal () for each of a set of voxels. Execution of the machine executable instructions further causes the processor to construct () a set of initial signals () by reconstructing the time dependent BOLD signal for each of the set of voxels using the magnetic resonance imaging data. Execution of the machine executable instructions further causes the processor to receive () a set of modified signals () in response to inputting the set of initial signals into the configured artificial neural network. The configured artificial neural network is configured for removing physiological artifacts from the set of initial signals.

Functional magnetic resonance imaging artifact removal by means of an artificial neural network

The present disclosure relates to a medical imaging method for enabling magnetic resonance imaging of a subject () using a set of imaging parameters of imaging protocols, the method comprising: receiving information related to the subject; using a predefined machine learning model for suggesting at least one imaging protocol for the received information, wherein the imaging protocol comprises at least part of the set of imaging parameters and associated values; providing the imaging protocol.

Method for enabling magnetic resonance imaging of a subject

When predicting required component service in an imaging device such as a magnetic resonance (MR) imaging device (), component parameters such as coil voltage, phase lock lost (PLL) events, etc. are sampled to monitor system components. Voltage samples are filtered according to their temporal proximity to coil plug-in and unplug events to generate a filtered data set that is analyzed by a processor () to determine whether to transmit a fault report. A service recommendation is received based on the transmitted report and includes a root cause diagnosis and service recommendation that is output to a user interface ().

Device and context specific signal pattern analysis for magnetic resonance imaging equipment component diagnostics and predictive maintenance

The invention relates to a magnetic resonance imaging data processing system () for processing motion artifacts in magnetic resonance imaging data sets using a deep learning network () trained for the processing of motion artifacts in magnetic resonance imaging data sets. The magnetic resonance imaging data processing system () comprises a memory () storing machine executable instructions () and the trained deep learning network (). Furthermore, the magnetic resonance imaging data processing system () comprises a processor () for controlling the magnetic resonance imaging data processing system. Execution of the machine executable instructions () causes the processor () to control the magnetic resonance imaging data processing system () to: receive a magnetic resonance imaging data set (), apply the received magnetic resonance imaging data set () as an input to the trained deep learning network (), process one or more motion artifacts present in the received magnetic resonance imaging data set () using the trained deep learning network ().

Deep learning based processing of motion artifacts in magnetic resonance imaging data

A medical imaging system for acquiring medical image data from an imaging zone. The medical imaging system includes a memory for storing machine executable instructions and medical imaging system commands. The medical imaging system further includes a user interface and a processor. Execution of the machine executable instructions causes the processor to: receive scan parameter data for modifying the behavior of the medical imaging system commands; receive metadata descriptive of imaging conditions from the user interface; store configuration data descriptive of a current configuration of the medical imaging system in the memory; calculate an error probability by comparing the metadata, the configuration data, and the scan parameter data using a predefined model, wherein the error probability is descriptive of a deviation between the metadata and between the configuration data and/or the scan parameter data; perform predefined action if the error probability is above a predetermined threshold.

Automated consistency check for medical imaging

In an MR imaging method and apparatus, a portion of a body placed in an examination volume of an MR device is subjected to an imaging sequence of RF pulses and switched magnetic field gradients. The imaging sequence is a stimulated echo sequence including i) two preparation RF pulses (&#x3b1;) radiated toward the portion of the body during a preparation period (), and ii) reading RF pulses (&#x3b2;) radiated toward the portion of the body during an acquisition period () temporally subsequent to the preparation period (). FID signals (I) and stimulated echo signals (I) are acquired during the acquisition period () with equal T*-weighting. A Bmap indicating a spatial distribution of the RF field of the preparation RF pulses within the portion of the body is derived from the acquired FID (I) and stimulated echo (I) signals.

[object Object]

A magnetic resonance imaging system () acquires magnetic resonance data () from a subject () that may include an electrically conductive object (e.g. an implant or a medical device). The magnetic resonance imaging system includes a radio-frequency transmitter () for generating a radio-frequency transmit field for acquiring the magnetic resonance data using a radio-frequency antenna (). The radio-frequency transmitter has multiple transmit channels. The radio-frequency antenna comprises multiple antenna elements () each adapted to connect to an antenna element. The amplitude and phase values of the RF transmit field of each of the transmit channels can be selected such that the magnetic field generated by the RF antenna is minimized at the location of the electrically conductive object, thereby reducing RF heating of the object.

Reducing the radio-frequency transmit field in a predetermined volume during magnetic resonance imaging

The invention relates to a method of characterizing the RF transmit chain of a magnetic resonance imaging scanner () using a local transmit/receive coil system (), comprising a first local NMR probe and a first local magnetic resonance coil, the first NMR probe being spatially located in immediate neighborhood to the first coil, a local receive coil system (), comprising a second local NMR probe and a second local magnetic resonance coil, the second NMR probe being spatially located in immediate neighborhood to the second coil, wherein the transmit chain comprises an external MR coil (), the method comprising: determining with the first magnetic resonance coil, a first MR signal phase evolution of the local RF transmit field generated by MR excitation of the first probe using the first magnetic resonance coil by measuring the RF response of the first probe upon said excitation, determining with the second magnetic resonance coil a second MR signal phase evolution of the local RF transmit field generated by MR excitation of the second probe using the external MR coil () by measuring the RF response of the second probe upon said excitation, calculating a phase offset between the first and second MR signal phase evolution.

Method of characterizing

A magnetic resonance sequence includes a repetitively applied radiofrequency pulse capable of causing a specific absorption rate (SAR) hot spot. The composition of the repetitive pulse is varied to generate versions of the repetitive pulse such that the SAR hot spot changes locations with subsequent applications of the repetitive pulse. To generate versions of the pulse, a pilot scan is performed to generate a patient model. A simulation of the SAR response to each of the versions of the repetitive pulse is performed to determine the location of SAR hot spot(s). A plurality of versions of the repetitive pulse is selected to be used in the magnetic resonance sequence.

SAR hotspot reduction by temporal averaging in parallel transmission MRI

In a method and apparatus to enable increased RF duty cycle in high field MR scans, a specific energy absorption rate (SAR) calculation processor calculates the local and global SAR or even a spatial SAR map. By incorporating additional information as, e.g. patient position, the SAR calculation accuracy can be increased as well as by using more patient specific pre-calculated information (e.g. based on different bio meshes), the so called Q-matrices. A sequence controller maybe provided to create a global SAR optimal RF pulse. After the optimal RF pulse is applied, the SAR and its spatial distribution are determined. SAR hotspots are also determined. Q-matrices within an appropriate radius around the hotspots are averaged and added to a global Q-matrix in a weighted fashion. After the global Q-matrix is updated, a new optimal RF pulse is created.

Real-time local and global SAR estimation for patient safety and improved scanning performance

When generating an MR image using a multi-channel transmit coil arrangement, SAR is reduced by employing a number of different RF pulses in a single scan. Each RF pulse exhibits a different performance and/or accuracy, resulting in different RF pulse-specific SAR values. As a result, the RF pulses differ slightly in actual excitation pattern, Bwaveform and/or k-space trajectory, etc. The average SAR over a single scan is thus reduced compared to a fixed RF pulse, without compromising image quality.

SAR reduction in parallel transmission by K-space dependent RF pulse selection

The invention relates to a method of acquiring MRI image data comprising the following steps: performing a 3-dimensional B1 mapping of a first volume using a first voxel size, selecting an MRI protocol, performing the B1-shim in accordance with the MRI protocol, performing the MRI protocol to acquire MRI imaging data of a second volume using a second voxel size, wherein the first voxel size is larger than the second voxel size, wherein the first volume is larger than the second volume, and wherein the second volume is contained within the first volume.

B1-mapping and B1L-shimming for MRI

A magnetic resonance system includes a magnetic resonance scanner having a multi-channel RF transmit system; an RF shimming processor configured to compute complex amplitude values for operative channels of the multi-channel RF transmit system based on a figure-of-merit including a weighted combination of a Buniformity measure and a total power measure. An RF transmit sub-system is configured to drive the multi-channel RF transmit system to excite magnetic resonance using the complex amplitude values computed by the RF shimming processor for the operative channels of the multi-channel RF transmit system.

RF shimming with RF power regularization using a multi-channel RF transmit system for MRI

A multi-channel RF transmitter arrangement comprising a plurality of RF transmitter elements like RE antennas, antenna elements, coils or coil elements, for generating an RF field, especially for use in a magnetic resonance imaging system for exciting nuclear magnetic resonances, and a method for generating such an RF field wherein the RF transmitter elements are segmented in a plurality of segments at least along the direction of one or more of the main magnetic field of the MRI system, the z-direction or the longitudinal direction.

Method and RF transmitter arrangement for generating RF fields

Growing community of inventors

Boenningstedt, Germany

Ingmar Graesslin