DEEP LEARNING AUTOMATED DERMATOPATHOLOGY

• US 20190286880A1
• Filed: 03/16/2018
• Published: 09/19/2019
• Est. Priority Date: 03/16/2018
• Status: Active Grant

• Current Assignee: Proscia Inc.
• Original Assignee: Proscia Inc.
• Inventors: Jackson, Brian H., Stavish, Coleman C.
• Granted Patent: US 10,460,150 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 18/2413 based on distances to training or reference patterns
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 2207/30088 Skin; Dermal
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/0014 using an image reference approach
• CPC Class Codes: G06V 10/454 Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation
• CPC Class Codes: G06V 20/698 Matching; Classification
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

Techniques for classifying a human cutaneous tissue specimen are presented. The techniques may include obtaining a computer readable image of the human tissue sample and preprocessing the image. The techniques may include applying a trained deep learning model to the image to label each of a plurality of image pixels with at least one probability representing a particular diagnosis, such that a labeled plurality of image pixels is obtained. The techniques can also include applying a trained discriminative classifier to contiguous regions of pixels defined at least in part by the labeled plurality of image pixels to obtain a specimen level diagnosis, where the specimen level diagnosis includes at least one of: basal cell carcinoma, dermal nevus, or seborrheic keratosis. The techniques can include outputting the specimen level diagnosis.

Claims

1. An at least partially computer implemented method of classifying a human cutaneous tissue specimen, the method comprising:

obtaining a computer readable image of the human tissue sample;

preprocessing the image;

applying a trained deep learning model to the image to label each of a plurality of image pixels with at least one probability representing a particular diagnosis, such that a labeled plurality of image pixels is obtained;

applying a trained discriminative classifier to contiguous regions of pixels defined at least in part by the labeled plurality of image pixels to obtain a specimen level diagnosis, wherein the specimen level diagnosis comprises at least one of;

basal cell carcinoma, dermal nevus, or seborrheic keratosis; and

outputting the specimen level diagnosis.

2. The method of claim 1, further comprising:

receiving a confirming diagnosis from a human operator; and

providing, automatically, the specimen level diagnosis to an electronic medical record system in response to the receiving the confirming diagnosis.

3. The method of claim 2, wherein the outputting the specimen level diagnosis further comprises causing the image of the human tissue sample to be displayed to the human operator contemporaneously with the specimen level diagnosis.

4. The method of claim 1, wherein the applying the trained discriminative classifier to the contiguous regions of pixels defined at least in part by the labeled plurality of image pixels further comprises obtaining a confidence level in the specimen level diagnosis, the method further comprising:

determining that the confidence level exceeds a predetermined threshold; and

providing the specimen level diagnosis to an electronic medical record system in response to the determining.

5. The method of claim 1, further comprising causing a heat map representing the labeled plurality of image pixels to be displayed.

6. The method of claim 1, wherein the trained deep learning model comprises a convolutional neural network.

7. The method of claim 6, further comprising training the trained deep learning model using stochastic graded descent with momentum.

8. The method of claim 1, wherein the trained discriminative classifier comprises a support vector machine.

9. The method of claim 8, further comprising training the trained discriminative classifier using a set of vectors derived from contiguous regions identified in images by the trained deep learning model.

10. The method of claim 1, wherein the preprocessing comprises detecting and isolating image regions that represent tissue.

11. A computer system for classifying a human cutaneous tissue specimen, the computer system comprising at least one electronic processor that executes instructions to perform operations comprising:

obtaining a computer readable image of the human tissue sample;

preprocessing the image;

applying a trained deep learning model to the image to label each of a plurality of image pixels with at least one probability representing a particular diagnosis, such that a labeled plurality of image pixels is obtained;

applying a trained discriminative classifier to contiguous regions of pixels defined at least in part by the labeled plurality of image pixels to obtain a specimen level diagnosis, wherein the specimen level diagnosis comprises at least one of;

basal cell carcinoma, dermal nevus, or seborrheic keratosis; and

outputting the specimen level diagnosis.

12. The system of claim 11, wherein the operations further comprise:

receiving a confirming diagnosis from a human operator; and

providing, automatically, the specimen level diagnosis to an electronic medical record system in response to the receiving the confirming diagnosis.

13. The system of claim 12, wherein the outputting the specimen level diagnosis further comprises causing the image of the human tissue sample to be displayed to the human operator contemporaneously with the specimen level diagnosis.

14. The system of claim 11, wherein the applying the trained discriminative classifier to the contiguous regions of pixels defined at least in part by the labeled plurality of image pixels further comprises obtaining a confidence level in the specimen level diagnosis, where the operations further comprise:

determining that the confidence level exceeds a predetermined threshold; and

providing the specimen level diagnosis to an electronic medical record system in response to the determining.

15. The system of claim 11, wherein the operations further comprise causing a heat map representing the labeled plurality of image pixels to be displayed.

16. The system of claim 12, wherein the trained deep learning model comprises a convolutional neural network.

17. The system of claim 16, wherein the operations further comprise training the trained deep learning model using stochastic graded descent with momentum.

18. The system of claim 11, wherein the trained discriminative classifier comprises a support vector machine.

19. The system of claim 18, wherein the operations further comprise training the trained discriminative classifier using a set of vectors derived from contiguous regions identified in images by the trained deep learning model.

20. The system of claim 11, wherein the preprocessing comprises detecting and isolating image regions that represent tissue.