Systems and methods for making medical decisions based on multimodal data

Patent metadata

Field	Value
Publication number	US-2025029720-A1
Application number	US-202318225009-A
Country	US
Kind code	A1
Filing date	Jul 21, 2023
Priority date	Jul 21, 2023
Publication date	Jan 23, 2025
Grant date	—

Abstract

Disclosed herein are deep-learning based systems, methods, and instrumentalities for medical decision-making. A system as described herein may implement an artificial neural network (ANN) that may include multiple encoder neural networks and a decoder neural network. The multiple encoder neural networks may be configured to receive multiple types of patient data (e.g., text and image based patient data) and generate respective encoded representations of the patient data. The decoder neural network (e.g., a transformer decoder) may be configured to receive the encoded representations and generate a medical decision, a medical summary, or a medical questionnaire based on the encoded representations. In examples, the decoder neural network may be configured to implement a large language model (LLM) that may be pre-trained for performing the aforementioned tasks.

First claim

What is claimed is:

1 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement an artificial neural network (ANN), wherein the ANN comprises:
multiple encoder neural networks configured to receive respective types of patient data and generate respective encoded representations of the types of patient data, wherein the types of patient data include at least a first type of patient data comprising a description of symptoms experienced by a patient, and a second type of patient data comprising one or more of a test result of the patient, a medical history of the patient, or demographic information about the patient; and a decoder neural network configured to receive the encoded representations and generate an output based on the encoded representations, wherein the output includes at least one of a medical decision, a medical summary, or a medical questionnaire associated with the patient.

2 . The system of claim 1 , wherein the types of patient data further include a third type of patient data comprising one or more medical images of the patient.

3 . The system of claim 2 , wherein the one or more medical images of the patient include multiple mammogram images of the patient that correspond to different views of a breast area of the patient, and wherein the output generated by the decoder neural network includes an indication of whether a medical abnormality exists in the breast area.

4 . The system of claim 1 , wherein the multiple encoder neural networks include a first encoder neural network and a second encoder neural network, the first encoder neural network configured to encode the first type of patient data into at least a first vector representing features of the first type of patient data, the second encoder neural network configured to encode the second type of patient data into at least a second vector representing features of the second type of patient data.

5 . The system of claim 4 , wherein the first encoder neural network and the second encoder neural network are trained jointly via contrastive learning and, during the training, first training data and second training data provided respectively to the first encoder neural network and the second encoder neural network are treated as a positive pair if the first training data and second training data belong to a same patient, and as a negative pair if the first training data and second training belong to different patients.

6 . The system of claim 1 , wherein the decoder neural network includes a transformer neural network.

7 . The system of claim 6 , wherein the respective encoded representations of the types of patient data are concatenated into a combined representation and provided to the decoder neural network, the combined representation including separators that distinguish the respective encoded representations of the types of patient data.

8 . The system of claim 1 , wherein the decoder neural network is configured to implement a large language model (LLM) pre-trained for predicting the output based on the respective encoded representations of the types of patient data.

9 . The system of claim 8 , wherein, when executed by the one or more computers, the instructions stored in the one or more storage devices further cause the one or more computers to receive a medical question and generate an answer to the medical question based on the LLM.

10 . The system of claim 1 , wherein the decoder neural network is configured to determine a distribution of actions associated with respective reward values and select, from the distribution, a first action associated with a highest reward value as the medical decision for the patient.

11 . The system of claim 10 , wherein the decoder neural network is trained to learn the distribution of actions via reinforcement learning.

12 . The system of claim 10 , wherein, when executed by the one or more computers, the instructions stored in the one or more storage devices further cause the one or more computers to obtain additional patient data based on the first action selected by the decoder neural network, encode the additional patient data into an additional representation using at least one of the multiple encoder neural networks, and determine, using the decoder neural network, a second action for the patient based at least on the additional representation.

13 . The system of claim 1 , wherein the medical decision associated with the patient includes a recommendation of a medical scan procedure for the patient.

14 . A method, comprising:
encoding, using respective encoder neural networks, multiple types of patient data into respective encoded representations, wherein the multiple types of patient data include at least at least a first type of patient data comprising a description of symptoms experienced by a patient, and a second type of patient data comprising one or more of a test result of the patient, a medical history of the patient, or demographic information about the patient; and generating an output based on the encoded representations, wherein the output is generated using a decoder neural network and comprises at least one of a medical decision, a medical summary, or a medical questionnaire associated with the patient.

15 . The method of claim 14 , wherein the multiple types of patient data further include a third type of patient data comprising one or more medical images of the patient.

16 . The method of claim 15 , wherein the one or more medical images of the patient include multiple mammogram images of the patient that correspond to different views of a breast area of the patient, and wherein the output generated by the decoder neural network includes an indication of whether a medical abnormality exists in the breast area.

17 . The method of claim 14 , wherein the decoder neural network includes a transformer neural network.

18 . The method of claim 17 , wherein the respective encoded representations of the multiple types of patient data are concatenated into a combined representation and provided to the decoder neural network, the combined representation including separators that distinguish the respective encoded representations of the multiple types of patient data.

19 . The method of claim 14 , wherein the decoder neural network is configured to implement a large language model (LLM) pre-trained for predicting the output based on the respective encoded representations of the multiple types of patient data, and wherein generating the output based on the encoded representations comprises receiving a medical question and generating an answer to the medical question based on the LLM.

20 . The method of claim 14 , wherein the output generated by the decoder neural network includes a recommendation of a medical scan procedure for the patient.

Assignees

• Shanghai United Imaging Intelligence Co Ltd

Inventors

• Sun Shanhui
• Chen Zhang
• Chen Xiao
• Liu Yikang
• Zhao Lin
• Chen Terrence
• Innanje Arun
• Sharma Abhishek
• Cui Wenzhe
• Fan Xiao

Classifications

• G16H15/00

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

• G16H10/20

  for electronic clinical trials or questionnaires · CPC title

• G16H30/40

  for processing medical images, e.g. editing · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H50/20 Primary

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title