Systems and methods for combining clinical goals with knowledge based dose prediction in treatment planning

Patent metadata

Field	Value
Publication number	US-11602643-B2
Application number	US-201816237496-A
Country	US
Kind code	B2
Filing date	Dec 31, 2018
Priority date	Dec 31, 2018
Publication date	Mar 14, 2023
Grant date	Mar 14, 2023

Abstract

A treatment planning apparatus includes: a modeler configured to obtain a model definition, wherein the model definition comprises a first quality metric of a first clinical goal; and a treatment planner having: a model trainer configured to obtain a set of existing treatment plans following desired clinical practice, and to perform model training to obtain a trained model based on the existing treatment plans and the first quality metric of the first clinical goal; an objective generator configured to generate a cost function based on the trained model; and an optimizer configured to determine a treatment plan based on the cost function.

First claim

What is claimed:

1. A treatment planning apparatus, comprising:
at least one processor; and
a memory storing computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to
obtain a model definition, wherein the model definition includes a first quality metric of a first clinical goal,
obtain a set of existing treatment plans following desired clinical practice,
perform model training to obtain a trained model based on the set of existing treatment plans and the first quality metric of the first clinical goal,
generate a cost function based on the trained model, and
determine a treatment plan based on the cost function.

2. The apparatus of claim 1 , wherein the model definition does not have a goal value associated with the first quality metric.

3. The apparatus of claim 2 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine an estimate of the goal value for the first quality metric.

4. The apparatus of claim 3 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine a cost function term based on the estimate of the goal value for the first quality metric.

5. The apparatus of claim 1 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine a regression model for a principal component of a dose-volume-histogram (DVH) curve.

6. The apparatus of claim 5 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine the principal component with emphasis on the DVH curve.

7. The apparatus of claim 1 , wherein the model definition comprises a first goal value corresponding to the first clinical goal.

8. The apparatus of claim 7 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine a cost function term based on the first goal value.

9. The apparatus of claim 7 , wherein the model definition comprises a second goal value corresponding to the first clinical goal.

10. The apparatus of claim 9 , wherein the model definition comprises a first weight for the first goal value, and a second weight for the second goal value.

11. The apparatus of claim 10 , wherein the first weight for the first goal value and the second weight for the second goal value are for influencing a manner in which a dose distribution is improved during treatment plan optimization.

12. The apparatus of claim 1 , wherein the model definition comprises a second quality metric of a second clinical goal.

13. The apparatus of claim 12 , wherein the model definition comprises:
a first weight for the first clinical goal, and
a second weight for the second clinical goal, wherein
the first weight for the first clinical goal and the second weight for the second clinical goal are for prescribing an order in which the first clinical goal and the second clinical goal are to be satisfied during treatment plan optimization.

14. The apparatus of claim 1 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to use a machine learning technique to create a statistical model for transferring the desired clinical practice into a patient geometry, and
generate the cost function based on the patient geometry.

15. The apparatus of claim 1 , wherein the first quality metric comprises a mean dose, a maximum dose, target coverage, or a relative or absolute volume of an organ having a dose larger than a specified dose level.

16. The apparatus of claim 1 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine the cost function using a knowledge-based technique based on the set of existing treatment plans and the model definition.

17. The apparatus of claim 1 , wherein. the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to provide a first prediction model for cases where the first clinical goal is met, and a second prediction model for cases where the first clinical goal is not met.

18. The apparatus of claim 1 , wherein the memory stores computer-executable instructions that, when executed by the at least one processor, cause the treatment planning apparatus to determine whether a plan would satisfy the first clinical goal or not.

19. A treatment planning method, comprising:
obtaining a model definition by a modeler, wherein the model definition includes a first quality metric of a first clinical goal;
obtaining, by a model trainer, a set of existing treatment plans following desired clinical practice;
performing, by the model trainer, model training to obtain a trained model based on the set of existing treatment plans and the first quality metric of the first clinical goal;
generating, by an objective generator, a cost function based on the trained model; and
determining a treatment an based on the cost function.

20. A product having a non-transitory medium storing a set of instructions, an execution of which causes a treatment planning method to be performed, the treatment planning method comprising:
obtaining a model definition by a modeler, wherein the model definition comprises a first quality metric of a first clinical goal;
obtaining, by a model trainer, a set of existing treatment plans following desired clinical practice;
performing, by the model trainer, model training to obtain a trained model based on the set of existing treatment plans and the first quality metric of the first clinical goal;
generating, by an objective generator, a cost function based on the trained model; and
determining a treatment plan based on the cost function.

Assignees

• Varian Medical Systems Int Ag

Inventors

• Laaksonen Hannu Mikael
• Kuusela Esa
• Cordero Marcos Maria Isabel
• Peltola Jarkko

Classifications

• G06N20/00

  Machine learning · CPC title

• A61N5/1031 Primary

  using a specific method of dose optimization · CPC title

• G06N20/20

  Ensemble learning · CPC title

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title

• G06N3/042

  Knowledge-based neural networks; Logical representations of neural networks · CPC title