Imaging-Based Tissue Mechanics Prediction Without Elastography

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing non-invasive methods for determining mechanical tissue characteristics, such as classical elastography, are inadequate for assessing tissues like cartilage due to limited mechanical excitations, and invasive methods like arthroscopy carry risks, necessitating a need for accurate non-invasive means to assess tissue state.

Innovation Solution

A system using machine learning architecture to determine mechanical tissue characteristics from in-vivo imaging data, including structural properties, without requiring mechanical excitations, by employing imaging modalities like MRI and photoacoustic imaging, and utilizing a constitutive artificial neural network to derive mechanical responsiveness.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If classical elastography is used to assess tissue mechanically, then non-invasive imaging is achieved, but measurement precision is insufficient for stiff tissues like cartilage

Engineering Contradiction:
Improvenon-invasive assessment capabilityVSAvoidmechanical excitation detection accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent replaces mechanical excitation-based elastography with a machine learning system that processes structural imaging data (MRI, CT, X-ray) to predict mechanical tissue characteristics. The ML model learns mappings from structural images to mechanical properties, eliminating the need for mechanical sound wave excitations while maintaining non-invasive assessment capability.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The invention changes the input parameters from mechanical response signals (in elastography) to structural imaging parameters (pixel intensities, anatomical features). By training the ML model on paired structural images and mechanical property measurements, the system predicts mechanical characteristics directly from structural parameters, achieving both non-invasiveness and improved precision for stiff tissues.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If arthroscopy is performed to obtain accurate mechanical tissue characteristics, then measurement precision improves, but object-affected harmful factors increase due to invasiveness

Engineering Contradiction:
Improvemechanical tissue characteristic accuracyVSAvoidsurgical and anesthetic complications
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent replaces invasive mechanical probing (arthroscopy) with non-invasive structural imaging combined with machine learning prediction. The ML model, trained on data including arthroscopy measurements, can now predict mechanical properties from benign structural images, eliminating surgical risks while preserving measurement accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The machine learning model serves as an intermediary that translates structural imaging data into mechanical property predictions. This intermediary enables the system to leverage the accuracy of invasive measurements without requiring direct invasive contact, bridging the gap between non-invasive imaging and accurate mechanical assessment.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Manufacturing precision

If machine learning architecture is trained using invasive mechanical testing data, then manufacturing precision of the model improves, but loss of substance occurs during sample preparation

Engineering Contradiction:
Improvemodel training accuracyVSAvoidtissue sample material
Core Design Contradiction:
Manufacturing precisionVSLoss of substance

Solution Approach 1:

The patent performs preliminary structural imaging of tissue samples before mechanical testing. These pre-acquired structural images are paired with subsequent mechanical measurement data to train the ML model. This preliminary imaging captures the structural state that corresponds to the mechanical properties, enabling the model to predict mechanical characteristics from structural data alone, thereby reducing the need for additional tissue consumption during testing.

Inventive Principle:
Principle #10Preliminary action

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

Enables accurate, non-invasive assessment of tissue state by determining mechanical tissue characteristics indicative of mechanical responsiveness, suitable for assessing conditions like osteoarthritis risk in cartilage without invasive procedures.

Implementation Method 1

images of a tissue region of interest have been acquired using an imaging modality suitable for in-vivo imaging of the tissue region of interest

Methodology Applied
Scientific EffectMagnetic resonance imaging:

Implementation Method 2

employing imaging modalities like MRI and photoacoustic imaging

Methodology Applied
Scientific EffectPhotoacoustic imaging: Photoacoustic Effect

Data Source

PatentEP4239569B1System for determining mechanical tissue characteristics from imaging data
Publication Date: 2026.01.28 HELMHOLTZ ZENTRUM HEREON GMBH
  • EP4239569B1 patent drawingFigure 1
  • EP4239569B1 patent drawingFigure 2
  • EP4239569B1 patent drawingFigure 3

AI summary

The invention relates to a system (100) for determining a mechanical tissue characteristic (Ψ). The system comprises a providing unit (101) for providing a set of images of a tissue region of interest, wherein the provided set of images comprises image features indicative of one or more structural properties (α, α0, CO, PG) of the tissue region of interest and has been acquired using an imaging modality suitable for in-vivo imaging of the tissue region of interest. The system further comprises a determining unit (102) for determining, based on the provided set of images, a mechanical tissue characteristic (Ψ) indicative of a mechanical responsiveness of the tissue region of interest. The determining unit comprises a machine learning architecture (201, 202) adapted to receive the provided set of images as an input and to provide the mechanical tissue characteristic as an output.