Artificial intelligence modeling for radiation therapy dose distribution analysis

Patent metadata

Field	Value
Publication number	US-2023310890-A1
Application number	US-202318329129-A
Country	US
Kind code	A1
Filing date	Jun 5, 2023
Priority date	Mar 22, 2021
Publication date	Oct 5, 2023
Grant date	—

Abstract

Disclosed herein are methods and systems to optimize a radiation therapy treatment plan using dose distribution values predicted via a trained artificial intelligence model. A server trains the AI model using a training dataset comprising data associated with a plurality of previously implemented radiation therapy treatments on a plurality of previous patients and dose distributions associated with one or more organs of each previous patient. The server then executes the trained AI model to predict dose distribution for a patient. The server then displays a heat map illustrating the predicted values, transmits the predicted values to a plan optimizer to generate an optimized treatment plan for the patient, and/or transmits an alert when a treatment plan generated by a plan optimizer deviates from rules and thresholds indicated within the patient's plan objectives.

First claim

What we claim is:

1 . A method comprising:
calculating, by a processor, a dose distribution value for an anatomical region of a patient; and displaying, by the processor, a heat map having a set of segments where each segment corresponds to a first coordinate and a second coordinate of the anatomical region of the patient,
wherein a visual attribute of each segment corresponds to the calculated dose distribution value, and
wherein at least one segment corresponding to a first region exceeding a first threshold or a second region below a second threshold is visually distinct from other segments within the heat map.

2 . The method of claim 1 , wherein the dose distribution value is calculated using an artificial intelligence model, wherein the artificial intelligence model is trained using a training dataset comprising data associated with a plurality of previously implemented radiation therapy treatments on a plurality of previous patients and dose distributions associated with one or more organs of each previous patient.

3 . The method of claim 2 , wherein the artificial intelligence model uses a treatment plan associated with the patient to identify the dose distribution value associated with the anatomical region of the patient.

4 . The method of claim 1 , wherein at least one of the first threshold or the second threshold is retrieved from a plan objective associated with the patient.

5 . The method of claim 1 , further comprising:
displaying, by the processor, an input element configured to receive an acceptance or rejection of at least one of the first region or the second region.

6 . The method of claim 1 , further comprising:
transmitting, by the processor, data associated with at least one of first region or the second region to a plan optimizer application.

7 . The method of claim 1 , wherein the visual attribute of each segment corresponds to a color, a shading, or a visual pattern.

8 . A computer system comprising:
a server comprising at least one processor and a non-transitory computer-readable medium containing instructions that when executed by the at least one processor causes the processor to perform operations comprising:
calculate a dose distribution value for an anatomical region of a patient; and
display a heat map having a set of segments where each segment corresponds to a first coordinate and a second coordinate of the anatomical region of the patient, wherein a visual attribute of each segment corresponds to the calculated dose distribution value, and wherein at least one segment corresponding to a first region exceeding a first threshold or a second region below a second threshold is visually distinct from other segments within the heat map.

9 . The computer system of claim 8 , wherein the dose distribution value is calculated using an artificial intelligence model, wherein the artificial intelligence model is trained using a training dataset comprising data associated with a plurality of previously implemented radiation therapy treatments on a plurality of previous patients and dose distributions associated with one or more organs of each previous patient.

10 . The computer system of claim 9 , wherein the artificial intelligence model uses a treatment plan associated with the patient to identify the dose distribution value associated with the anatomical region of the patient.

11 . The computer system of claim 8 , wherein at least one of the first threshold or the second threshold is retrieved from a plan objective associated with the patient.

12 . The computer system of claim 8 , wherein the instructions further cause the processor to:
display an input element configured to receive an acceptance or rejection of at least one of the first region or the second region.

13 . The computer system of claim 8 , wherein the instructions further cause the processor to:
transmit data associated with at least one of first region or the second region to a plan optimizer application.

14 . The computer system of claim 8 , wherein the visual attribute of each segment corresponds to a color, a shading, or a visual pattern.

15 . A non-transitory machine-readable storage medium having computer-executable instructions stored thereon that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
calculate a dose distribution value for an anatomical region of a patient; and display a heat map having a set of segments where each segment corresponds to a first coordinate and a second coordinate of the anatomical region of the patient, wherein a visual attribute of each segment corresponds to the calculated dose distribution value, and wherein at least one segment corresponding to a first region exceeding a first threshold or a second region below a second threshold is visually distinct from other segments within the heat map.

16 . The non-transitory machine-readable storage medium of claim 15 , wherein the dose distribution value is calculated using an artificial intelligence model, wherein the artificial intelligence model is trained using a training dataset comprising data associated with a plurality of previously implemented radiation therapy treatments on a plurality of previous patients and dose distributions associated with one or more organs of each previous patient.

17 . The non-transitory machine-readable storage medium of claim 16 , wherein the artificial intelligence model uses a treatment plan associated with the patient to identify the dose distribution value associated with the anatomical region of the patient.

18 . The non-transitory machine-readable storage medium of claim 15 , wherein at least one of the first threshold or the second threshold is retrieved from a plan objective associated with the patient.

19 . The non-transitory machine-readable storage medium of claim 15 , wherein the instructions are further configured to:
display an input element configured to receive an acceptance or rejection of at least one of the first region or the second region.

20 . The non-transitory machine-readable storage medium of claim 15 , wherein the instructions are further configured to:
transmit data associated with at least one of first region or the second region to a plan optimizer application.

Assignees

• Siemens Healthineers Int Ag

Inventors

• Peltola Jarkko
• Rusanen Marko
• Pietila Ville

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title

• G06N3/082

  modifying the architecture, e.g. adding, deleting or silencing nodes or connections · CPC title