Electronic device and method for diagnosing dementia with Lewy bodies or predicting morbidity to dementia with Lewy bodies

Patent metadata

Field	Value
Publication number	US-12414734-B2
Application number	US-202318230352-A
Country	US
Kind code	B2
Filing date	Aug 4, 2023
Priority date	Aug 5, 2022
Publication date	Sep 16, 2025
Grant date	Sep 16, 2025

Abstract

The electronic device for diagnosing dementia with Lewy bodies (DLB) or predicting morbidity to DLB according to the present invention includes a processor that measures cortical thicknesses for a plurality of regions of the brain by using brain MRI images of a normal group and a DLB patient group, generates a DLB pattern matrix by using a residual matrix according to a difference between the average cortical thickness and the cortical thickness for each region, applies a first cortical thickness matrix generated by using a brain MRI image of the subject to the DLB pattern to calculate a first DLB pattern score, and diagnoses the subject as DLB or predicting morbidity to DLB by using the first DLB pattern score.

First claim

The invention claimed is:

1. An electronic device for diagnosing dementia with Lewy bodies (DLB) or predicting morbidity to DLB, comprising:
a display;
a communicator configured to receive a first brain MRI image of a subject to be diagnosed, the first brain MRI image is a target brain MRI image;
a memory configured to store operating programs;
a processor configured to execute the programs to:
standardize and preprocess the first brain MRI image,
generate a first vector, the elements of the first vector corresponding to cortical thicknesses of predetermined regions, the cortical thicknesses being calculated using the first brain MRI image,
compute an inner product between the first vector and a second vector to generate a DLB pattern score, the second vector being a DLB pattern vector pre-stored in the memory,
diagnose the subject as having DLB or predict morbidity to DLB based on the DLB pattern score, and
display the result of the diagnosis on the display,

wherein the second vector is generated by:
measuring cortical thicknesses for the predetermined regions of the brain for second brain MRI images of a normal group and a DLB patient group,
generating a residual matrix, each element of the residual matrix being a difference between an average cortical thickness and the cortical thickness for each region in each of the second brain MRI image, and
calculating weights for each region of the brain, the weights reflecting relationships between brain regions, the elements of the second vector being the weights.

2. The electronic device of claim 1 , wherein the processor is configured to generate a covariance matrix representing a correlation between the cortical thicknesses for the each region by using the residual matrix, and generate at least one DLB pattern vector according to at least one principal component of the covariance matrix extracted through principal component analysis.

3. The electronic device of claim 2 , wherein the processor is further configured to select a DLB pattern vector as the second vector, the second vector having the lowest Akaike information criterion value among a plurality of DLB pattern vectors generated based on different principal components.

4. The electronic device of claim 1 , wherein the processor is configured to calculate the DLB pattern score as a standard score compared to the distribution of DLB pattern scores in the normal group to diagnose the subject as DLB or predict morbidity to DLB.

5. The electronic device of claim 4 , wherein age and gender of the normal group are matched by age and gender of the subject.

6. The electronic device of claim 4 , wherein the processor is configured to differentiate the subject from Parkinson's disease to diagnose the subject as DLB or predict morbidity to DLB, based on whether a cut-off value of the standard score is exceeded.

7. The electronic device of claim 6 , wherein the processor is configured to diagnose the subject as DLB or predict that the subject is afflicted with DLB, if the standard score is more than the cut-off value, and predict that the subject is afflicted with Parkinson's disease, if the standard score is less than or equal to the cut-off value.

8. The electronic device of claim 7 , wherein the processor is configured to predict that the subject's memory and spatial perception ability are reduced, if the subject is a REM sleep behavior disorder patient and the standard score is more than the cut-off value.

9. The electronic device of claim 6 , wherein the cut-off value is 1.

10. A method for diagnosing dementia with Lewy bodies (DLB) or predicting morbidity to DLB which is performed by a device for diagnosing DLB or predicting morbidity to DLB, the method comprising:
receiving a first brain MRI image of a subject to be diagnosed;
standardizing and preprocessing the first brain MRI image, the first brain MRI image is a target brain MRI image;
generating a first vector, the elements of the first vector corresponding to cortical thicknesses of predetermined regions, the cortical thicknesses being calculated using the first brain MRI image;
computing an inner product between the first vector and a second vector to generate a DLB pattern score, the second vector being a DLB pattern vector pre-stored in a memory;
diagnosing the subject as having DLB or predicting morbidity to DLB based on the DLB pattern score; and
displaying the result of the diagnosis;
wherein the second vector is generated by:
measuring cortical thicknesses for the predetermined regions of the brain for second brain MRI images of a normal group and a DLB patient group,
generating a residual matrix, each element of the residual matrix being a difference between an average cortical thickness and the cortical thickness for each region in each of the second brain MRI image, and
calculating weights for each region of the brain, the weights reflecting relationships between brain regions, the elements of the second vector being the weights.

11. The method of claim 10 , further comprising:
generating a covariance matrix representing a correlation between the cortical thicknesses for the each region by using the residual matrix; and
generating at least one DLB pattern vector according to at least one principal component of the covariance matrix extracted through principal component analysis.

12. The method of claim 11 , further comprising:
selecting a DLB pattern vector as the second vector, the second vector having the lowest Akaike information criterion value among a plurality of DLB pattern vectors generated based on different principal components.

13. The method of claim 10 , wherein the step of diagnosing the subject as DLB or predicting morbidity to DLB comprises:
diagnosing the subject as DLB or predicting morbidity to DLB by calculating the DLB pattern score as a standard score compared to the distribution of DLB pattern scores in the normal group.

14. The method of claim 13 , wherein age and gender of the normal group are matched by age and gender of the subject.

15. The method of claim 13 , wherein the step of diagnosing the subject as DLB or predicting morbidity to DLB comprises:
diagnosing the subject as DLB or predicting morbidity to DLB based on whether a cut-off value of the standard score is exceeded.

16. The method of claim 15 , wherein the step of diagnosing the subject as DLB or predicting morbidity to DLB comprises:
diagnosing the subject as DLB or predicting that the subject is afflicted with DLB, if the standard score is more than the cut-off value; and
predicting that the subject is afflicted with Parkinson's disease, if the standard score is less than or equal to the cut-off value.

17. The method of claim 16 , further comprising:
determining that the subject's memory and spatial perception ability are reduced, if the subject is a REM sleep behavior disorder patient and the standard score is more than the cut-off value.

18. The method of claim 15 , wherein the cut-off value is 1.

19. The method of claim 10 , further comprising:
combining with at least one selected from a group consisting of a clinical biomarker, an imaging biomarker and a blood biomarker of the subject to predict whether the subject has developed DLB or is afflicted with DLB.

Assignees

• Seoul Nat Univ R&Db Foundation

Inventors

• Lee Jee Young
• Kim Hee Jung
• Shin Jung Hwan
• Kim Yu Kyeong

Classifications

• A61B5/055

  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging · CPC title

• A61B5/7275

  Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor · CPC title

• A61B5/7246

  using correlation, e.g. template matching or determination of similarity · CPC title

• A61B2503/08

  Elderly · CPC title

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title