Machine learning-based generation of 3D dose distributions for volumes not included in a training corpus

Patent metadata

Field	Value
Publication number	US-12138476-B2
Application number	US-202117485794-A
Country	US
Kind code	B2
Filing date	Sep 27, 2021
Priority date	Sep 27, 2021
Publication date	Nov 12, 2024
Grant date	Nov 12, 2024

Abstract

A radiation treatment plan three-dimensional dose prediction machine learning model is trained using a training corpus that includes a plurality of radiation treatment plans that are not specific to a particular patient and wherein the training corpus includes some, but not all, possible patient volumes of interest. Information regarding the patient (including information regarding at least one volume of interest for the patient that was not represented in the training corpus) is input to the radiation treatment plan three-dimensional dose prediction machine model. The latter generates predicted three-dimensional dose distributions that include a predicted three-dimensional dose distribution for the at least one volume of interest that was not represented in the training corpus.

First claim

What is claimed is:

1. A method to facilitate generating a radiation treatment plan for a patient, the method comprising:
providing a radiation treatment plan three-dimensional dose prediction machine learning model that has been trained using a training corpus that includes a plurality of radiation treatment plans that are not specific to the patient;
providing information regarding the patient, wherein the patient information includes information regarding at least one organ-at-risk for the patient that was not represented in the training corpus;
inputting the information regarding the patient to the radiation treatment plan three-dimensional dose prediction machine learning model and generating predicted three-dimensional dose distributions that include a predicted three-dimensional dose distribution for at least one organ-at-risk for the patient that was not represented in the training corpus.

2. The method of claim 1 wherein the radiation treatment plan three-dimensional dose prediction machine learning model comprises a neural network machine learning model.

3. The method of claim 2 wherein the neural network machine learning model comprises a convolutional neural network.

4. The method of claim 1 wherein the information regarding the patient comprises:
at least one computed tomography image;
at least one patient treatment volume contour;
at least one organ-at-risk contour; and
field geometry information.

5. The method of claim 1 further comprising:
generating at least one dosimetric parameter for the at least one organ-at-risk for the patient that was not represented in the training corpus as a function of the predicted three-dimensional dose distribution.

6. The method of claim 5 wherein the at least one dosimetric parameter comprises at least one of:
a dose volume histogram;
maximum dose levels within a given contour;
a minimum dose level for a target structure;
mean dose levels within a given contour;
dose-volume points;
a homogeneity index for target coverage;
a generalized equivalent uniform dose;
a conformity index.

7. The method of claim 1 wherein:
the information regarding the patient comprises information regarding a non-biological structure;
the training corpus does not include the non-biological structure;

and wherein the method further comprises:
inputting the information regarding the non-biological structure to the radiation treatment plan three-dimensional dose prediction machine learning model and generating predicted three-dimensional dose distributions that include a predicted three-dimensional dose distribution for the non-biological structure.

8. The method of claim 1 further comprising:
automatically generating an optimization objective for the at least one organ-at-risk for the patient that was not represented in the training corpus as a function of the predicted three-dimensional dose distribution for the at least one organ-at-risk for the patient that was not represented in the training corpus.

9. The method of claim 8 further comprising:
optimizing a radiation treatment plan for the patient as a function of the optimization objective.

10. The method of claim 9 further comprising:
administering therapeutic radiation to the patient as a function of the radiation treatment plan.

11. An apparatus to facilitate generating a radiation treatment plan for a patient, the apparatus comprising:
a memory having a radiation treatment plan three-dimensional dose prediction machine learning model stored therein, wherein the radiation treatment plan three-dimensional dose prediction machine learning model has been trained using a training corpus that includes a plurality of radiation treatment plans that are not specific to the patient;
a control circuit operably coupled to the memory and configured to:
receive information regarding the patient, wherein the patient information includes information regarding at least one organ-at-risk for the patient that was not represented in the training corpus;
input the information regarding the patient to the radiation treatment plan three-dimensional dose prediction machine learning model and generate predicted three-dimensional dose distributions that include a predicted three-dimensional dose distribution for the at least one organ-at-risk for the patient that was not represented in the training corpus.

12. The apparatus of claim 11 wherein the radiation treatment plan three-dimensional dose prediction machine learning model comprises a neural network machine learning model.

13. The apparatus of claim 12 wherein the neural network machine learning model comprises a convolutional neural network.

14. The apparatus of claim 11 wherein the information regarding the patient comprises:
at least one computed tomography image;
at least one patient treatment volume contour;
at least one organ-at-risk contour; and
field geometry information.

15. The apparatus of claim 11 wherein the control circuit is further configured to:
generating at least one dosimetric parameter for the at least one organ-at-risk for the patient that was not represented in the training corpus as a function of the predicted three-dimensional dose distribution.

16. The apparatus of claim 15 wherein the at least one dosimetric parameter comprises at least one of:
a dose volume histogram;
maximum dose levels within a given contour;
a minimum dose level for a target structure;
mean dose levels within a given contour;
dose-volume points;
a homogeneity index for target coverage;
a generalized equivalent uniform dose;
a conformity index.

17. The apparatus of claim 11 wherein:
the information regarding the patient comprises information regarding a non-biological structure;
the training corpus does not include the non-biological structure;
the control circuit is further configured to input the information regarding the non-biological structure to the radiation treatment plan three-dimensional dose prediction machine learning model and generate predicted three-dimensional dose distributions that include a predicted three-dimensional dose distribution for the non-biological structure.

18. The apparatus of claim 11 wherein the control circuit is further configured to:
automatically generate an optimization objective for the at least one organ-at-risk for the patient that was not represented in the training corpus as a function of the predicted three-dimensional dose distribution for the at least one organ-at-risk for the patient that was not represented in the training corpus.

19. The apparatus of claim 18 wherein the control circuit is further configured to:
optimize a radiation treatment plan for the patient as a function of the optimization objective.

20. The apparatus of claim 19 further comprising:
a radiation treatment platform operably coupled to at least receive the radiation treatment plan and configured to administer therapeutic radiation to the patient as a function of the radiation treatment plan.

Assignees

• Siemens Healthineers Int Ag

Inventors

• Czeizler Elena
• Hakala Mikko
• Basiri Shahab
• Laaksonen Hannu
• Cordero Marcos Maria
• Boylan Christopher
• Peltola Jarkko
• Pietila Ville
• Kuusela Esa

Classifications

• G06N3/08

  Learning methods · CPC title

• A61N5/1039

  using functional images, e.g. PET or MRI · CPC title

• G16H50/20

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

• G16H20/40

  relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture · CPC title

• A61N2005/1041

  using a library of previously administered radiation treatment applied to other patients · CPC title