Determination of a degree of deformity of at least one vertebral bone

Patent metadata

Field	Value
Publication number	US-11074688-B2
Application number	US-201916678046-A
Country	US
Kind code	B2
Filing date	Nov 8, 2019
Priority date	Nov 21, 2018
Publication date	Jul 27, 2021
Grant date	Jul 27, 2021

Abstract

For processing a medical image, medical image data representing a medical image of at least a portion of a vertebral column is received. The medical image data is processed to determine a plurality of positions within the image. Each of the plurality of positions corresponds to a position relating to a vertebral bone within the vertebral column. Data representing the plurality of positions is processed to determine a degree of deformity of at least one vertebral bone within the vertebral column.

First claim

The invention claimed is:

1. A method of determining a degree of deformity of at least one vertebral bone, the method comprising:
receiving medical image data, the medical image data representing a medical image of at least a portion of a vertebral column;
determining, from the medical image data, a plurality of positions, each of the plurality of positions corresponding to a position relating to a vertebral bone within the vertebral column;
generating, from at least the plurality of positions, a multilabel mask representing a model of the vertebral column, the model including a region of interest around each of the plurality of vertebral bones;
extracting one or more vertebrae measurements from the multilabel mask and the plurality of positions; and
determining, from the extracted one or more vertebrae measurements, the degree of deformity of the at least one vertebral bone within the vertebral column.

2. The method according to claim 1 , wherein the plurality of positions each comprises a coordinate within the medical image, each coordinate representing a central point of the respective vertebral bone.

3. The method according to claim 1 , further comprising determining the presence of one or more vertebral fractures on the basis of the determined degree of deformity.

4. The method according to claim 1 , wherein generating the multilabel mask further comprises:
determining, for each of the plurality of positions, a corresponding anatomical feature; and
assigning, to each of the regions of interest, a label representing the respective corresponding anatomical feature.

5. The method according to claim 1 , wherein extracting one or more vertebrae measurements comprises:
determining one or more sagittal points for each of vertebral bones; and
determining a Mahalanobis distance between each of the one or more sagittal points and corresponding center coordinates of each of the vertebral bones.

6. The method according to claim 1 , wherein extracting one or more vertebrae measurements comprises:
determining a mineral bone density value for each of the regions of interest, the mineral bone density values being determined based on Hounsfield Unit values within the respective regions of interest.

7. The method according to claim 1 , wherein determining the degree of deformity comprises:
for each region of interest, determining an estimate of loss of height based on a comparison between the one or more vertebrae measurements a first region of region of interest and one or more adjacent regions of interest; and
determining the degree of deformity at least partly on the basis of the estimated loss of height.

8. The method according to claim 7 , wherein the estimate of loss of height comprises an estimate of loss of height for anterior, mid, and posterior portions of the corresponding vertebra.

9. The method according to claim 1 , further comprising determining a fracture classification or a severity of a fracture on the basis of the determined degree of deformity.

10. A data processing system for determining a degree of deformity of at least one vertebral bone, the data processing system comprising:
a processor configured to:
receive medical image data, the medical image data representing a medical image of at least a portion of a vertebral column;
process the medical image data to determine a plurality of positions within the image, each of the plurality of positions corresponding to a position relating to a vertebral bone within the vertebral column;
generate, from at least the plurality of positions, a multilabel mask representing a model of the vertebral column, the model including a region of interest around each of the plurality of vertebral bones;
extract one or more vertebrae measurements from the multilabel mask and the plurality of positions; and
process the one or more vertebrae measurements to determine a degree of deformity of at least one vertebral bone within the vertebral column.

11. The data processing system according to claim 10 , wherein the processor is configured to implement a trained deep image-to-image neural network to determine the plurality of positions.

12. The data processing system according to claim 10 , wherein the processor is configured to:
determine, for each of the plurality of positions, a corresponding anatomical feature; and
assign, to each of the regions of interest, a label in the multilabel mask representing the respective corresponding anatomical feature.

13. The data processing system according to claim 10 , wherein the processor is configured to:
process each of the regions of interest to determine one or more sagittal points for each of the corresponding vertebral bones; and
determine a Mahalanobis distance between each of the one or more sagittal points and the corresponding center coordinates.

14. The data processing system according to claim 10 , wherein the processor is configured to determine a mineral bone density value for each of the regions of interest, the mineral bone density values being determined based on Hounsfield Unit values within the respective regions of interest.

15. A non-transitory computer readable storage medium comprising a computer program, the computer program being executable by a processor, the computer program including program code for:
receiving medical image data, the medical image data representing a medical image of at least a portion of a vertebral column;
determining, from processing the medical image data, a plurality of positions within the image, each of the plurality of positions corresponding to a position relating to a vertebral bone within the vertebral column;
generating, from at least the plurality of positions, a multilabel mask representing a model of the vertebral column, the model including a region of interest around each of the plurality of vertebral bones;
determining one or more sagittal points for each of the corresponding vertebral bones;
determining a Mahalanobis distance between each of the one or more sagittal points and corresponding center coordinates of each of the plurality of vertebral bones; and
determining, from at least the Mahalanobis distances, a degree of deformity of at least one vertebral bone within the vertebral column.

Assignees

• Siemens Healthcare Gmbh

Inventors

• Chabin Guillaume
• Sperl Jonathan
• Kärgel Rainer
• Grbic Sasa
• Ionasec Razvan
• Comaniciu Dorin

Classifications

• G06V10/143

  Sensing or illuminating at different wavelengths · CPC title

• G06V10/457

  by analysing connectivity, e.g. edge linking, connected component analysis or slices · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/454

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title