Multi-Shot EPI Reconstruction for Motion and Distortion Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing MR imaging techniques, particularly diffusion-weighted imaging (DWI), are vulnerable to motion-induced artifacts and geometric distortions, especially in multi-shot EPI, which affect image quality and spatial resolution, particularly in regions with magnetic field inhomogeneities like the head and neck.

Innovation Solution

A method for MR imaging that utilizes a multi-shot or multi-acquisition EPI sequence with overlapping k-space segments, combined with phase-encoding polarity variation and motion-correction algorithms, to derive and correct geometric distortions using geometry distortion maps and B0 maps, enhancing image reconstruction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multi-shot EPI is used to increase spatial resolution, then image resolution is improved, but sensitivity to patient motion increases leading to image artifacts

Engineering Contradiction:
Improvespatial resolutionVSAvoidmotion sensitivity
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent divides the k-space acquisition into multiple segments or shots, where each shot acquires a portion of k-space. This segmentation allows for longer individual readout durations per shot while still achieving high overall resolution when all segments are combined, thereby reducing motion sensitivity compared to single-shot EPI while maintaining high spatial resolution through the segmented multi-shot approach

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements motion correction algorithms that use information from the acquired data to detect and correct motion artifacts. By analyzing the acquired k-space segments and applying correction transformations based on detected motion, the system feedback-adjusts the image reconstruction to compensate for patient motion, thereby maintaining reliability while using multi-shot EPI for high resolution

Inventive Principle:
Principle #23Feedback

2Productivity

If EPI readout is used to reduce motion sensitivity, then acquisition speed is improved, but geometric distortions increase due to main magnet field inhomogeneities and eddy currents

Engineering Contradiction:
Improveacquisition speedVSAvoidgeometric accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent performs preliminary distortion correction by acquiring B0 field maps before the main EPI acquisition and using these maps to pre-calculate and apply geometric distortion corrections during image reconstruction. This preliminary action compensates for the geometric distortions that will occur during the fast EPI readout, allowing the system to maintain both high acquisition speed and geometric accuracy

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces B0 field maps as an intermediary element that mediates between the fast EPI readout and the final geometrically accurate image. These field maps serve as a reference that allows the system to calculate distortion fields and apply corrective transformations, thereby acting as an intermediary that enables both rapid acquisition and geometric precision

Inventive Principle:
Principle #24Intermediary (Mediator)

3Manufacturing precision

If readout segmentation is used to reduce geometric distortions, then geometric accuracy is improved, but readout duration per segment increases leading to T2* decay losses

Engineering Contradiction:
Improvegeometric accuracyVSAvoidT2* signal decay
Core Design Contradiction:
Manufacturing precisionVSLoss of energy

Solution Approach 1:

The patent optimizes the readout parameters for each segmented acquisition by adjusting echo spacing, readout duration, and gradient strengths to minimize T2* decay effects. By carefully controlling these parameters within each segment and using parallel imaging techniques to accelerate acquisition, the system achieves geometric accuracy through segmentation while minimizing signal loss from T2* decay

Inventive Principle:
Principle #35Parameter changes

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach achieves high-resolution, motion-resistant MR imaging with reduced geometric distortions, suitable for applications requiring precise geometric accuracy, such as head imaging and radiation therapy planning.

Implementation Method 1

acquiring gradient echo signals from the object using a multi-shot or multi-acquisition EPI imaging sequence

Methodology Applied
Scientific EffectMagnetic field gradients: Magnetic Field

Implementation Method 2

the diffusion of protons (of water molecules) along the direction of the diffusion gradient reduces the amplitude of the acquired MR signals

Methodology Applied
Scientific EffectDiffusion: Diffusion

Implementation Method 3

additional diffusion-encoding (or diffusion-weighting) magnetic field gradients of different strengths and/or directions are applied prior to the signal readout

Methodology Applied
Scientific EffectMagnetic field gradients: Magnetic Field

Implementation Method 4

the image quality can be low and the spatial resolution is limited in single-shot diffusion-weighted EPI. These resolution losses and also geometric distortions are mainly caused by main magnet field inhomogeneities (B0) and T2*- and/or T2-relaxation during the readout

Methodology Applied
Scientific EffectMagnetic field inhomogeneities: Magnetic Field

Implementation Method 5

EPI requires rapid switching of magnetic field gradients causing a particularly high level of eddy currents

Methodology Applied
Scientific EffectEddy currents: Eddy Currents

Data Source

PatentEP4679122A1Gradient echo imaging with compensation of geometric distortions
Publication Date: 2026.01.14 KONINKLIJKE PHILIPS NV
  • EP4679122A1 patent drawingFigure 1
  • EP4679122A1 patent drawingFigure 2
  • EP4679122A1 patent drawingFigure 3~4

AI summary

The invention discloses a method of magnetic resonance imaging of an object (10), comprising: acquiring gradient echo signals from the object (10) using a multi-shot or multi-acquisition echo planar imaging, EPI, sequence, wherein each shot (21-24) or acquisition corresponds to a different k-space segment, wherein the different k-space segments only partially overlap in a central portion of k-space; deriving geometry distortion information by combining the echo signals of different shots (21-24) or acquisitions; and reconstructing a magnetic resonance, MR, image by combining the echo signals of the multiple shots (21-24) or acquisitions, correcting for motion between the different shots (21-24) or acquisitions based on the overlap of the different k-space segments, and using the geometry distortion information to reduce geometric distortions in the reconstructed MR image. An imaging system is further disclosed, comprising a computational system configured to perform the method.