SYSTEMS AND METHODS FOR ESTIMATING HISTOLOGICAL FEATURES FROM MEDICAL IMAGES USING A TRAINED MODEL

• US 20170300622A1
• Filed: 09/11/2015
• Published: 10/19/2017
• Est. Priority Date: 09/11/2014
• Status: Active Grant

• Current Assignee: Medical College of Wisconsin Incorporated
• Original Assignee: Medical College of Wisconsin Incorporated
• Inventors: Laviolette, Peter Sherman
• Granted Patent: US 11,154,212 B2
• Time in Patent Office: Days
• CPC Class Codes: A61B 2576/026 for the brain
• CPC Class Codes: A61B 5/055 involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
• CPC Class Codes: A61B 6/032 Transmission computed tomography [CT]
• CPC Class Codes: A61B 6/501 for diagnosis of the head, e.g. neuroimaging or craniography
• CPC Class Codes: A61B 6/5217 extracting a diagnostic or physiological parameter from medical diagnostic data
• CPC Class Codes: A61B 6/541 involving acquisition triggered by a physiological signal
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 7/0014 using an image reference approach
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

Systems and methods for estimating quantitative histological features of a subject'"'"'s tissue based on medical images of the subject are provided. For instance, quantitative histological features of a tissue are estimated by comparing medical images of the subject to a trained model that relates histological features to multiple different medical image contrast types, whether from one medical imaging modality or multiple different medical imaging modalities. In general, the trained model is generated based on medical images of ex vivo samples, in vitro samples, in vivo samples or combinations thereof, and is based on histological features extracted from those samples. A machine learning algorithm, or other suitable learning algorithm, is used to generate the trained model. The trained model is not patient-specific and thus, once generated, can be applied to any number of different individual subjects.

Claims

1. A method for estimating quantitative histological features of a tissue using a medical imaging system, the steps of the method comprising:

(a) acquiring image data from a subject using a medical imaging system;

(b) reconstructing a plurality of images of the subject from the acquired image data;

(c) providing a trained model that correlates a plurality of different medical image contrasts with histological feature data; and

(d) estimating quantitative histological feature values of a tissue in the subject by applying the trained model to the reconstructed images.

2. The method as recited in claim 1, wherein the provided trained model correlates a plurality of different medical image contrasts acquired from at least one subject with histological feature data acquired from tissue samples obtained from the same at least one subject.

3. The method as recited in claim 1, further comprising:

(e) producing a histological feature map that indicates a quantitative histological feature by assigning the estimated quantitative histological feature values to pixels in the histological feature map.

4. The method as recited in claim 1, wherein the image data acquired in step (a) is representative of a plurality of different image contrasts and the images reconstructed in step (b) depict the plurality of different image contrasts.

5. The method as recited in claim 4, wherein the plurality of different image contrasts depicted in the reconstructed images are the same as the plurality of different medical image contrasts on which the trained model is based.

6. The method as recited in claim 1, wherein step (d) includes using a machine learning algorithm to correlate the reconstructed images with the provided trained model.

7. The method as recited in claim 6, wherein the machine learning algorithm is based on a partial least squares regression.

8. The method as recited in claim 1, wherein the quantitative histological feature values are based on microscopic tissue analysis.

9. The method as recited in claim 1, wherein the quantitative histological feature values are representative of cell density.

10. The method as recited in claim 1, wherein the quantitative histological feature values are representative of vascularity.

11. The method as recited in claim 1, wherein the quantitative histological feature values are representative of a voxel-wise percentage of necrosis.

12. The method as recited in claim 1, wherein the quantitative histological feature values are representative of a clinically determined score indicating a state of tissue pathology.

13. The method as recited in claim 12, wherein the clinically determined score is a Gleason score.

14. The method as recited in claim 1, wherein the medical imaging system is a magnetic resonance imaging (MRI) system and the images used in the provided trained model are magnetic resonance images.

15. The method as recited in claim 1, wherein the provided trained model is generated based on medical images and histological feature data obtained from a subject different from the subject imaged in step (a).

16. A method for generating a trained model of quantitative histological features that are correlated with medical images, the steps of the method comprising:

(a) acquiring image data from a subject with a medical imaging system, the image data representing a plurality of different image contrasts;

(b) reconstructing a plurality of images of the subject from the acquired image data;

(c) acquiring quantitative histological features from a tissue sample obtained from the subject; and

(d) generating the trained model by correlating the plurality of images of the subject with the acquired quantitative histological features using a machine learning algorithm.

17. The method as recited in claim 16, wherein the machine learning algorithm includes a partial least squares regression.

18. The method as recited in claim 16, wherein step (c) includes acquiring a histology image of the tissue sample and computing the quantitative histological features from the histology image.

19. The method as recited in claim 16, wherein step (d) includes resampling the quantitative histological features to a spatial resolution of at least one of the plurality of medical images.

20. The method as recited in claim 16, wherein step (d) includes co-registering the quantitative histological features with at least one of the plurality of medical images.

21. The method as recited in claim 16, wherein the medical imaging system is a magnetic resonance imaging (MRI) system.

22. The method as recited in claim 16, wherein the tissue sample is an ex vivo tissue sample.