Method and system for image-based operational decision support

Patent metadata

Field	Value
Publication number	US-12380993-B2
Application number	US-202217939163-A
Country	US
Kind code	B2
Filing date	Sep 7, 2022
Priority date	Sep 7, 2021
Publication date	Aug 5, 2025
Grant date	Aug 5, 2025

Abstract

Techniques are described for providing image-based operational decision support. The technique includes providing initial imaging data from a preceding examination of a patient, determining clinical findings by automated processing of the initial imaging data, generating decision data at least comprising a decision whether a further recording of a number of images is necessary, and generating a suggested set of imaging parameter values for recording this number of images. The decision data may be based on the determined clinical findings, and the technique may further include outputting the suggested set of imaging parameter values for recording a number of images of the patient. Also described are a related clinical decision system, a decision module, and a related medical imaging system.

First claim

What is claimed is:

1. A method for image-based operational decision support, comprising:
generating initial imaging data from an initial examination of a patient;
determining, via automated processing of the initial imaging data, clinical findings;
generating decision data comprising a decision of whether a further recording of a number of medical images of the patient is required via a subsequent examination of the patient using a respective imaging modality, the decision data being based on the determined clinical findings;
when the decision data indicates that the further recording of the number of images is required, generating a suggested set of imaging parameter values for the further recording of the number of medical images of the patient,
wherein the suggested set of imaging parameter values define a manner in which the subsequent examination is to be performed via the imaging modality based upon the determined clinical findings; and
outputting the suggested set of imaging parameter values for the further recording of the number of images of the patient.

2. The method according to claim 1 , wherein in the set of generating the initial imaging data comprises:
generating an initial set of imaging parameter values; and
recording an image based on the initial set of imaging parameter values as part of the initial examination of the patient,
wherein the generating the suggested set of imaging parameter values is based upon the initial imaging data, and
wherein the initial set of imaging parameter values is based on patient attributes and/or sensor data associated with the patient.

3. The method according to claim 1 , wherein the generating the decision data comprises:
generating the decision data further based on a position of the patient in a pathway of a clinical guideline.

4. The method according to claim 1 , wherein the generating the suggested set of imaging parameter values comprises:
generating the suggested set of imaging parameter values based on the generated decision data and further based on one or more of (i) patient attributes, (ii) an initial set of imaging parameter values, (iii) the clinical findings, or (iv) sensor data associated with the patient.

5. The method according to claim 1 , wherein the suggested set of imaging parameter values are identified with a control of an imaging modality for further recording the number of medical images of the patient.

6. The method according to claim 1 , wherein the generating the decision data comprises:
generating the decision data via a decision engine that has been trained on one or more of (i) clinical findings, (ii) parameter settings used for recording the initial imaging data, or (iii) patient attributes, to generate the decision data as output data,
wherein the output data further comprises the suggested set of imaging parameter values and/or information about a diagnostic goal of a follow-up scan.

7. The method according to claim 6 , wherein the generating the suggested set of imaging parameter values comprises:
generating the suggested set of imaging parameter values based on the decision data output via the decision engine and further based on one or more of (i) the clinical findings, (ii) parameter settings used for recording the initial imaging data, (iii) patient attributes, or (iv) sensor data associated with the patient.

8. The method according to claim 1 , wherein:
the clinical findings are assigned to respective input nodes of a neural network,
the decision data is assigned to respective output nodes of the neural network,
the clinical findings comprise numerical values,
a number of input nodes of the neural network correspond to a number of the numerical values, and
each of the input nodes are assigned to one of the numerical values.

9. The method according to claim 1 , wherein:
the clinical findings are assigned to respective input nodes of a neural network and the decision data is assigned to respective output nodes of the neural network,
the clinical findings comprise spatially-resolved localization data that is fed into a corresponding number of input nodes of the neural network,
the spatially-resolved localization data comprises a plurality of voxels, and
a number of input nodes correspond to a number of the plurality of voxels.

10. The method according to claim 1 , wherein the acts of generating the decision data and the suggested set of imaging parameter values comprise:
training a neural network on training data generated based on a monitoring of a performance of a plurality of imaging studies on different medical imaging machines, and
wherein the training of the neural network is further based on an analysis of previously performed imaging studies, parameters of follow-up scans, and data related to the recording the initial imaging data and a follow-up scan.

11. The method according to claim 1 , wherein the acts of generating the decision data and the suggested set of imaging parameter values comprise:
generating the decision data and the suggested set of imaging parameter values via a decision engine and a parameter setting engine, respectively, each of the decision engine and the parameter setting engine comprising a respective individual neural network, and
wherein input nodes of the neural network of the parameter setting engine are adapted to output nodes of the neural network of the decision engine.

12. The method according to claim 11 , wherein:
the decision engine and the parameter setting engine are combined as a single neural network,
parameters at the input nodes of the single neural network comprise the clinical findings, and
parameters at the output nodes of the single neural network comprise the suggested set of imaging parameter values.

13. The method according to claim 12 , wherein the single neural network is configured as a multitask network, which generates the decision data and the suggested set of imaging parameter values for a follow-up scan.

14. The method according to claim 13 , wherein:
information is shared across multiple tasks of the multitask network, and
the multitask network has task-specific layers dedicated for generating the decision data and task-specific layers dedicated for generating the suggested set of imaging parameter values.

15. A system for image-based operational decision support, comprising:
a data interface configured to receive initial imaging data from an initial a preceding examination of a patient; and
processing circuitry configured to:
determine clinical findings by automated processing of the initial imaging data;
generate decision data comprising a decision of whether a further recording of a number of medical images of the patient is required via a subsequent examination of the patient using a respective imaging modality; and
when the decision data indicates that the further recording of the number of images is required, generate a suggested set of imaging parameter values for the further recording of the number of medical images of the patient, the decision data being based on the determined clinical findings,

wherein the suggested set of imaging parameter values define a manner in which the subsequent examination is to be performed via the imaging modality based upon the determined clinical findings, and
wherein the data interface is further configured to output the suggested set of imaging parameter values for the further recording of the number of images of the patient.

16. A non-transitory computer-readable medium configured to store a computer program that, when executed by processing c

Assignees

• Siemens Healthcare Gmbh
• Siemens Healthineers Ag

Inventors

• Kartmann René
• Krüger Benedikt
• Gall Peter
• Lauer Lars
• Gühring Jens

Classifications

• G06T2207/20081

  Training; Learning · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• G16H50/20

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

• G16H30/20

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

• G06T2207/30096

  Tumor; Lesion · CPC title