Methods and systems for generating dose estimation models for radiotherapy treatment planning

Patent metadata

Field	Value
Publication number	US-10342994-B2
Application number	US-201615377962-A
Country	US
Kind code	B2
Filing date	Dec 13, 2016
Priority date	Dec 13, 2016
Publication date	Jul 9, 2019
Grant date	Jul 9, 2019

Abstract

One example method for generating a dose estimation model for radiotherapy treatment planning may include obtaining training data that includes multiple treatment plans associated with respective multiple past patients. The method may also include processing the training data to determine, from each of the multiple treatment plans, first data that includes one or more features associated with a particular past patient, second data associated with treatment planning trade-off selected for the particular past patient and third data associated with radiation dose for delivery to the particular past patient. The method may further include generating the dose estimation model by training, based on the first data, second data and third data from the multiple treatment plans, the dose estimation model to estimate a relationship that transforms the first data and second data to the third data.

First claim

We claim:

1. A method for a computer system to generate a dose estimation model for radiotherapy treatment planning, the method comprising:
obtaining training data that includes multiple treatment plans associated with respective multiple past patients;
processing the training data to determine, from each of the multiple treatment plans, first data that includes one or more features associated with a particular past patient, second data associated with a treatment planning trade-off between a parameterized first objective and a parameterized second object selected for the particular past patient and third data associated with radiation dose for delivery to the particular past patient;
generating the dose estimation model by training, based on the first data, the second data and the third data from the multiple treatment plans, the dose estimation model to estimate a relationship that transforms the first data and the second data to the third data;
obtaining first input data associated with patient geometry of a new patient and second input data associated with a particular treatment planning trade-off selected by a clinician for the new patient; and
generating a treatment plan for the new patient using the dose estimation model based on the first input data and second input data.

2. The method of claim 1 , wherein processing the training data comprises:
determining, from each of the multiple treatment plans, the second data that is associated with one of the following treatment planning trade-offs:
trade-off between a first objective associated with an organ-at-risk (OAR) and a second objective associated with a target;
trade-off between a first objective associated with a first OAR and a second objective associated with a second OAR;
trade-off between a first objective associated with a target and a second objective associated with multiple OARs;
trade-off between a first objective associated with a first feature that is non-dosimetrical and a second objective associated with one or more second features; and
trade-off between a first objective associated with a first group of features and a second objective associated with a second group of features.

3. The method of claim 2 , wherein processing the training data comprises:
determining, from each of the multiple treatment plans, the second data that includes one or more dosimetrical features associated with the first objective or the second objective.

4. The method of claim 2 , wherein processing the training data to determine the second data comprises:
determining, from each of the multiple treatment plans, the second data that includes one or more non-dosimetrical features associated with the first objective or the second objective.

5. The method of claim 1 , wherein processing the training data comprises:
determining, from each of the multiple treatment plans, the first data that includes one or more of the following features associated with patient geometry: target volume, organ-at-risk (OAR) volume, relative overlap volume and relative out-of-field volume; and
determining, from each of the multiple treatment plans, the third data that includes one or more of the following features associated with radiation dose: dose volume histogram (DVH) and dose distribution.

6. The method of claim 1 , wherein training the dose estimation model comprises:
estimating a first relationship that transforms the first data to the second data; and
based on the first relationship, estimating the relationship, being a second relationship, that transforms the first data and the second data to the third data.

7. The method of claim 1 , wherein obtaining the second input data comprises:
receiving an instruction associated with the particular treatment planning trade-off; and
based on the instruction, determining one or more values of the second input data.

8. A non-transitory computer-readable storage medium that includes a set of instructions which, in response to execution by a processor of a computer system, cause the processor to perform a method of generating a dose estimation model for radiotherapy treatment planning, the method comprising:
obtaining training data that includes multiple treatment plans associated with respective multiple past patients;
processing the training data to determine, from each of the multiple treatment plans, first data that includes one or more features associated with a particular past patient, second data associated with a treatment planning trade-off between a parameterized first objective and a parameterized second object selected for the particular past patient and third data associated with radiation dose for delivery to the particular past patient;
generating the dose estimation model by training, based on the first data, the second data and the third data from the multiple treatment plans, the dose estimation model to estimate a relationship that transforms the first data and the second data to the third data;
obtaining first input data associated with patient geometry of a new patient and second input data associated with a particular treatment planning trade-off selected by a clinician for the new patient; and
generating a treatment plan for the new patient using the dose estimation model based on the first input data and second input data.

9. The non-transitory computer-readable storage medium of claim 8 , wherein processing the training data comprises:
determining, from each of the multiple treatment plans, the second data that is associated with one of the following treatment planning trade-offs:
trade-off between a first objective associated with an organ-at-risk (OAR) and a second objective associated with a target;
trade-off between a first objective associated with a first OAR and a second objective associated with a second OAR;
trade-off between a first objective associated with a target and a second objective associated with multiple OARs;
trade-off between a first objective associated with a first feature that is non-dosimetrical and a second objective associated with one or more second features; and
trade-off between a first objective associated with a first group of features and a second objective associated with a second group of features.

10. The non-transitory computer-readable storage medium of claim 9 , wherein processing the training data comprises:
determining, from each of the multiple treatment plans, the second data that includes one or more dosimetrical features associated with the first objective or the second objective.

11. The non-transitory computer-readable storage medium of claim 9 , wherein processing the training data to determine the second data comprises:
determining, from each of the multiple treatment plans, the second data that includes one or more non-dosimetrical features associated with the first objective or the second objective.

12. The non-transitory computer-readable storage medium of claim 8 , wherein processing the training data comprises:
determining, from each of the multiple treatment plans, the first data that includes one or more of the following features associated with patient geometry: target volume, organ-at-risk (OAR) volume, relative overlap volume and relative out-of-field volume; and
determining, from each of the multiple treatment plans, the third data that includes one or more of the following features associated with radiation dose: dose volume histogram (DVH) and dose distribution.

13. The non-transitory computer-readable storage medium of claim 8 , wherein training the dose estimation model comprises:
estimating a first relationship that transforms the first data to the

Assignees

• Varian Medical Systems Int Ag

Inventors

• Kuusela Esa
• Halko Lauri
• Cordero Marcos María

Classifications

• G16H30/20

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

• G16H30/40

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

• G16H20/40

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

• G06N20/00

  Machine learning · CPC title

• G16H50/50

  for simulation or modelling of medical disorders · CPC title