Systems and methods to process electronic images with automatic protocol revisions

Patent metadata

Field	Value
Publication number	US-12580071-B2
Application number	US-202318316711-A
Country	US
Kind code	B2
Filing date	May 12, 2023
Priority date	May 13, 2022
Publication date	Mar 17, 2026
Grant date	Mar 17, 2026

Abstract

Systems and methods are described herein for processing electronic medical images. The method may include determining, using an automated routine, whether a pathology protocol is accessible; determining a first set of one or more training images, the first set of one or more training images comprising digital medical images annotated utilizing the pathology protocol; and providing the training images to a machine learning model capable of analyzing digital medical images according to the pathology protocol or guideline. The providing may further include determining a starting model, splitting the first set of one or more training images into a training set A and an evaluation set B of digital medical images, fine tuning the starting model with the training set A to determine the machine learning model, evaluating the machine learning model with the training set B, and upon receiving a passing evaluation, saving the determined machine learning model.

First claim

What is claimed is:

1 . A computer-implemented method for processing electronic medical images comprising:
determining, using an automated routine, whether a pathology protocol is accessible; determining a first set of one or more training images, the first set of one or more training images comprising digital medical images annotated utilizing the pathology protocol; and providing the training images to a machine learning model capable of analyzing digital medical images according to the pathology protocol or guideline, the providing further including:
determining a starting model;
splitting the first set of one or more training images into a training set A and an evaluation set B of digital medical images;
fine tuning the starting model with the training set A to determine the machine learning model;
evaluating the machine learning model with the training set B; and
upon receiving a passing evaluation, saving the determined machine learning model to digital storage.

2 . The method of claim 1 , wherein upon determining that a pathology protocol is accessible further comprises:
parsing data of the pathology protocol; determining a new synoptic report based on the parsed data; and providing the new synoptic report to the machine learning model.

3 . The method of claim 1 , wherein the training images include annotations of digital medical images, the annotations being performed according to the pathology protocol, wherein the annotations comprise measurements, designations, and/or diagnosis.

4 . The method of claim 1 , wherein determining whether a pathology protocol is available is performed iteratively at predetermined time intervals.

5 . The method of claim 1 , wherein the pathology protocol is a new cancer protocol template.

6 . The method of claim 1 , wherein the pathology protocol is a PDF, word document, or CSV document.

7 . The method of claim 1 , wherein the starting model in a machine learning model trained on a previous version of the pathology protocol.

8 . The method of claim 1 , further including:
determining a new synoptic report based on the pathology guideline, the machine learning model being trained to fill out the new synoptic report when analyzing new digital medical images.

9 . The method of claim 8 , wherein determining a new synoptic report includes automatically creating, by a machine learning system and/or a rules-based artificial intelligence algorithm, a new synoptic report.

10 . The method of claim 8 , wherein determining a new synoptic reports includes receiving a synoptic report corresponding to the pathology protocol or guideline from an external user or system.

11 . The method of claim 1 , wherein the machine learning model is not determined until the determined first set of one or more training images exceed a threshold value of training images.

12 . The method of claim 1 , further including:
determining the machine learning model has been applied to a predetermined number of slides to meet a study requirement.

13 . A system for processing electronic medical images, the system comprising:
at least one memory storing instructions; and at least one processor configured to execute the instructions to perform operations comprising: determining, using an automated routine, whether a pathology protocol is accessible; determining a first set of one or more training images, the first set of one or more training images comprising digital medical images annotated utilizing the pathology protocol; and providing the training images to a machine learning model capable of analyzing digital medical images according to the pathology protocol or guideline, the providing further including:
determining a starting model;
splitting the first set of one or more training images into a training set A and an evaluation set B of digital medical images;
fine tuning the starting model with the training set A to determine the machine learning model;
evaluating the machine learning model with the training set B; and
upon receiving a passing evaluation, saving the determined machine learning model to digital storage.

14 . The system of claim 13 , wherein upon determining that a pathology protocol is available further comprises:
parsing data of the pathology protocol; determining a new synoptic report based on the parsed data; and providing the new synoptic report to the machine learning model.

15 . The system of claim 13 , wherein the training images include annotations of digital medical images, the annotations being performed according to the pathology protocol, wherein the annotations comprise measurements, designations, and/or diagnosis.

16 . The system of claim 13 , wherein determining whether a pathology protocol is available is performed iteratively at predetermined time intervals.

17 . The system of claim 13 , wherein the pathology protocol is a new cancer protocol template.

18 . The system of claim 13 , wherein the pathology protocol is a PDF, word document, or CSV document.

19 . The system of claim 13 , wherein the starting model in a machine learning model trained on a previous version of the pathology protocol.

20 . A non-transitory computer-readable medium storing instructions that, when executed by a processor, perform operations processing electronic medical images, the operations comprising:
determining, using an automated routine, whether a pathology protocol is accessible; determining a first set of one or more training images, the first set of one or more training images comprising digital medical images annotated utilizing the pathology protocol; and providing the training images to a machine learning model capable of analyzing digital medical images according to the pathology protocol or guideline, the providing further including:
determining a starting model;
splitting the first set of one or more training images into a training set A and an evaluation set B of digital medical images;
fine tuning the starting model with the training set A to determine the machine learning model;
evaluating the machine learning model with the training set B; and

upon receiving a passing evaluation, saving the determined machine learning model to digital storage.

Assignees

• Paige Ai Inc

Inventors

• Kunz Jeremy Daniel
• Lee Donghun

Classifications

• G06V10/70

  using pattern recognition or machine learning (optical pattern recognition or electronic computations therefor G06V10/88) · CPC title

• G16H30/20

  for handling medical images, e.g. DICOM, HL7 or PACS · CPC title

• G16H15/00

  ICT specially adapted for medical reports, e.g. generation or transmission thereof · CPC title

• G06V2201/03

  Recognition of patterns in medical or anatomical images · CPC title

• G06V10/776

  Validation; Performance evaluation · CPC title