System and method for predicting vertebral artery dissection

Patent metadata

Field	Value
Publication number	US-11103142-B2
Application number	US-201916372996-A
Country	US
Kind code	B2
Filing date	Apr 2, 2019
Priority date	Apr 2, 2019
Publication date	Aug 31, 2021
Grant date	Aug 31, 2021

Abstract

A method of determining a risk probability of vertebral artery dissection (VAD) in a patient, including receiving medical image information of the patient and clinical report information of the patient; extracting at least one biomarker corresponding to a vertebral artery segment included in the medical image information; extracting patient history information from the clinical report information; and determining the risk probability of VAD using a deep learning classification model based on the extracted at least one biomarker and the extracted patient history information.

First claim

What is claimed is:

1. A method of determining a risk probability of vertebral artery dissection (VAD) in a patient, the method comprising:
receiving, by at least one processor, medical image information of the patient and clinical report information of the patient;
extracting, by said at least one processor, at least one biomarker corresponding to a vertebral artery segment included in the medical image information, wherein the at least one biomarker is determined based on a centerline of the vertebral artery segment that is extracted from the medical image information, and a lumen of the vertebral artery segment that is segmented using the centerline;
extracting, by said at least one processor, patient history information from the clinical report information; and
determining, by said at least one processor, the risk probability of VAD using a deep learning classification model based on the extracted at least one biomarker and the extracted patient history information.

2. The method of claim 1 , wherein the extracting of the at least one biomarker further comprises:
extracting, by said at least one processor, the centerline of the vertebral artery segment;
segmenting, by said at least one processor, the lumen of the vertebral artery segment using the extracted centerline; and
extracting, by said at least one processor, the at least one biomarker based on the centerline and the lumen.

3. The method of claim 2 , wherein the extracting of the centerline further comprises:
determining, by said at least one processor, an initial centerline by performing a minimum cost path analysis; and
refining, by said at least one processor, the initial centerline to generate a refined centerline.

4. The method of claim 3 , wherein the minimum cost path analysis is performed, by said at least one processor, based on at least two seed points corresponding to the vertebral artery segment, and a cost image generated from the medical image information.

5. The method of claim 4 , wherein the at least two seed points include a first seed point manually annotated at a beginning of the vertebral artery segment, and a second seed point manually annotated at an end of the vertebral artery segment.

6. The method of claim 3 , further comprising:
generating, by said at least one processor, a plurality of curved multi plan reconstruction (CMPR) images using the initial centerline;
performing, by said at least one processor, the minimum cost path analysis on the CMPR images to extract a revised centerline; and
converting, by said at least one processor, the revised centerline to an original image space of the medical image information to obtain the refined centerline.

7. The method of claim 3 , wherein the refined centerline is used to generate a plurality of CMPR images, and wherein the plurality of CMPR images are used to segment the lumen.

8. The method of claim 3 , wherein the lumen is segmented, by said at least one processor, from the medical image information using a deep learning segmentation model.

9. The method of claim 1 , wherein the patient history information is extracted, by said at least one processor, from the clinical report information using a language model.

10. The method of claim 1 , wherein the patient history information includes at least from among a gender of the patient, a body mass index of the patient, a drug history of the patient, and a blood test result of the patient.

11. A apparatus for determining a risk probability of vertebral artery dissection (VAD) in a patient, the apparatus comprising:
at least one memory configured to store program code; and
at least one processor configured to read the program code and operate as instructed by the program code, the program code including:
receiving code configured to cause the at least one processor to receive medical image information of the patient and clinical report information of the patient;
first extracting code configured to cause the at least one processor to extract at least one biomarker corresponding to a vertebral artery segment included in the medical image information, wherein the at least one biomarker is determined based on a centerline of the vertebral artery segment that is extracted from the medical image information, and a lumen of the vertebral artery segment that is segmented using the centerline;
second extracting code configured to cause the at least one processor to extract patient history information from the clinical report information; and
determining code configured to cause the at least one processor to determine the risk probability of VAD using a deep learning classification model based on the extracted at least one biomarker and the extracted patient history information.

12. The apparatus of claim 11 , wherein the first extracting code further comprises:
third extracting code configured to cause the at least one processor to extract the centerline of the vertebral artery segment;
segmenting code configured to cause the at least one processor to segment the lumen of the vertebral artery segment using the extracted centerline;
fourth extracting code configured to cause the at least one processor to extract the at least one biomarker based on the centerline and the lumen.

13. The apparatus of claim 12 , wherein the third extracting code further comprises:
initial centerline determining code configured to cause the at least one processor to determine an initial centerline by performing a minimum cost path analysis;
refining code configured to cause the at least one processor to refine the initial centerline to generate a refined centerline.

14. The apparatus of claim 13 , wherein the minimum cost path analysis is performed based on at least two seed points corresponding to the vertebral artery segment, and a cost image generated from the medical image information.

15. The apparatus of claim 14 , wherein the at least two seed points include a first seed point manually annotated at a beginning of the vertebral artery segment, and a second seed point manually annotated at an end of the vertebral artery segment.

16. The apparatus of claim 13 , further comprising:
generating code configured to cause the at least one processor to generate a plurality of curved multi plan reconstruction (CMPR) images using the initial centerline;
analysis code configured to cause the at least one processor to perform the minimum cost path analysis on the CMPR images to extract a revised centerline; and
converting code configured to cause the at least one processor to convert the revised centerline to an original image space of the medical image information to obtain the refined centerline.

17. The apparatus of claim 13 , wherein the refined centerline is used to generate a plurality of CMPR images, and wherein the plurality of CMPR images are used to segment the lumen.

18. The apparatus of claim 13 , wherein the lumen is segmented from the medical image information using a deep learning segmentation model.

19. The apparatus of claim 11 , wherein the at least one biomarker includes at least one from among a geometry of the vertebral artery segment, a cervical distortion, and stiffness of a vessel wall.

20. A non-transitory computer-readable medium storing instructions that, when executed by at least one processor of a device for determining a risk probability of vertebral artery dissection (VAD) in a patient, cause the at least one processor to:
receive medical image information of the patient and clinical report information of the pati

Assignees

• Tencent America LLC

Inventors

• Tang Hui
• Du Nan
• Tu min
• Wang Kun
• Han Lianyi
• Fan Wei

Classifications

• A61B5/055

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

• G06N3/045

  Combinations of networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · 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