Systems and methods for processing images to classify the processed images for digital pathology
Systems and methods are disclosed for receiving a target image corresponding to a target specimen, the target specimen comprising a tissue sample of a patient, applying a machine learning model to the target image to determine at least one characteristic of the target specimen and/or at least one characteristic of the target image, the machine learning model having been generated by processing a plurality of training images to predict at least one characteristic, the training images comprising images of human tissue and/or images that are algorithmically generated, and outputting the at least one characteristic of the target specimen and/or the at least one characteristic of the target image.
1. A computer-implemented method for analyzing a digital pathology image from a user, the method comprising:
receiving the digital pathology image from the user, wherein the digital pathology image is a human tissue image or an algorithmically generated image;
applying a first trained machine learning model to the digital pathology image to determine specimen property information, the first trained machine learning model having been generated by processing a plurality of digital pathology images to identify one or more parameters of the plurality of digital pathology images and predict the specimen property information associated with the plurality of digital pathology images, the training images comprising one or both of human tissue images or algorithmically generated images;
comparing predicted specimen property information to stored laboratory system information to determine if the predicted specimen property information matches the stored laboratory system information to a predetermined margin;
upon determining a mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, generating an alert; and
providing the alert for display via a graphical user interface of the device associated with the user.
2. The computer-implemented method of claim 1 , further comprising:
receiving a request for specimen property information determination from the user; and
based on the request, applying a machine learning model to the digital pathology image to determine the specimen property information.
3. The computer-implemented method of claim 1 , wherein the specimen property information includes at least one of a specimen type, a specimen classification, or a specimen characteristic.
4. The computer-implemented method of claim 1 , further comprising:
upon determining the mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, causing a modification in processing behavior.
5. The computer-implemented method of claim 1 , further comprising:
upon determining the mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, determining if the mismatch is due to the stored laboratory system information; and
upon determining the mismatch is due to the stored laboratory system information, correcting the stored laboratory system information based on the determined mismatch.
6. The computer-implemented method of claim 1 , further comprising:
upon determining a match to the predetermined margin between the predicted specimen property information and the stored laboratory system information, generating a verification message based on the match; and
providing the verification message for display via the graphical user interface of the device associated with the user.
7. The computer-implemented method of claim 1 , further comprising:
applying a second trained machine learning model to the predicted specimen property information to process a received image or related information from the user and perform an automated diagnosis, determine contextual information, or generate an alert,
wherein the second trained machine learning model has been generated by processing the plurality of digital pathology images and associated predicted specimen property information to identify one or more of a diagnosis, the contextual information, or a requirement for the alert, the training images comprising one or both of human tissue images or algorithmically generated images.
8. The computer-implemented method of claim 1 , further comprising:
determining if the digital pathology image requires a specimen-specific diagnosis or a diagnosis-aide tool; and
upon determining the digital pathology image requires the specimen-specific diagnosis or the diagnosis-aide tool, applying a machine learning model to the digital pathology image to determine the specimen property information.
9. The computer-implemented method of claim 1 , wherein the first trained machine learning model is trained by:
receiving, as training data, a plurality of training images, the plurality of training images comprising the plurality of digital pathology images;
receiving, as training data, one or more specimen property information labels corresponding to each training image of the plurality of training images; and
training a machine learning model, using the training data, to infer the specimen property information for each digital pathology image.
10. The computer-implemented method of claim 9 , further comprising:
receiving, as training data, one or more image property information labels corresponding to each training image of the plurality of training images.
11. The computer-implemented method of claim 1 , further comprising:
providing the determined specimen property information for one or both of storage or display via a graphical user interface of a device associated with the user.
12. A system for analyzing a digital pathology image from a user, the system comprising:
at least one memory storing instructions; and
at least one processor executing the instructions to perform operations comprising:
receiving the digital pathology image from the user, wherein the digital pathology image is a human tissue image or an algorithmically generated image;
applying a first trained machine learning model to the digital pathology image to determine specimen property information, the first trained machine learning model having been generated by processing a plurality of digital pathology images to identify one or more parameters of the plurality of digital pathology images and predict the specimen property information associated with the plurality of digital pathology images, the training images comprising one or both of human tissue images or algorithmically generated images;
comparing predicted specimen property information to stored laboratory system information to determine if the predicted specimen property information matches the stored laboratory system information to a predetermined margin;
upon determining a mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, generating an alert; and
providing the alert for display via a graphical user interface of the device associated with the user.
13. The system of claim 12 , the operations further comprising:
upon determining the mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, causing a modification in processing behavior.
14. The system of claim 12 , the operations further comprising:
upon determining the mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, determining if the mismatch is due to the stored laboratory system information; and
upon determining the mismatch is due to the stored laboratory system information, correcting the stored laboratory system information based on the determined mismatch.
15. The system of claim 12 , the operations further comprising:
upon determining a match to the predetermined margin between the predicted specimen property information and the stored laboratory system information, generating a verification message based on the match; and
providing the verification message for display via the graphical user interface of the device associated with the user.
16. The system of claim 12 , the operations further comprising:
applying a second trained machine learning model to the predicted specimen property information to process a received image or related information from the user and perform an automated diagnosis, determine contextual information, or generate an alert,
wherein the second trained machine learning model has been generated by processing the plurality of digital pathology images and associated predicted specimen property information to identify one or more of a diagnosis, the contextual information, or a requirement for the alert, the training images comprising one or both of human tissue images or algorithmically generated images.
17. The system of claim 12 , wherein the first trained machine learning model is trained by:
receiving, as training data, a plurality of training images, the plurality of training images comprising the plurality of digital pathology images;
receiving, as training data, one or more specimen property information labels corresponding to each training image of the plurality of training images; and
training a machine learning model, using the training data, to infer the specimen property information for each digital pathology image.
18. The system of claim 17 , the operations further comprising:
receiving, as training data, one or more image property information labels corresponding to each training image of the plurality of training images.
19. The system of claim 12 , the operations further comprising:
providing the determined specimen property information for one or both of storage or display via a graphical user interface of a device associated with the user.
20. A computer-implemented method for analyzing a digital pathology image from a user, the method comprising:
receiving a request for specimen property information determination from the user;
upon receiving the request, receiving the digital pathology image from the user, wherein the digital pathology image is a human tissue image or an algorithmically generated image;
applying a first trained machine learning model to the digital pathology image to determine specimen property information, the first trained machine learning model having been generated by processing a plurality of digital pathology images to identify one or more parameters of the plurality of digital pathology images and predict the specimen property information associated with the plurality of digital pathology images, the training images comprising one or both of human tissue images or algorithmically generated images;
comparing predicted specimen property information to stored laboratory system information to determine if the predicted specimen property information matches the stored laboratory system information to a predetermined margin;
upon determining a mismatch to the predetermined margin between the predicted specimen property information and the stored laboratory system information, causing one or more of: (i) generating an alert, (ii) determining a modification in processing behavior, or (iii) determining the mismatch is due to the stored laboratory system information; and
providing the determined specimen property information, the alert, the determined modification in the processing behavior, or the determined mismatch for one or both of storage or display via a graphical user interface of a device associated with the user.