Method and computer program product for dose map prediction for radiotherapy treatment planning
An estimated or predicted dose for radiotherapy treatment may be generated based on a partial dose map including dose information only for one or more regions of interest within a treatment site, by use of a properly trained machine learning system such as a U-Net or a V-Net. Said partial dose map typically set to fulfil clinical goals. A method of training such a machine learning system is also disclosed.
1. A method for training a machine learning system to generate an estimated spatial radiotherapy dose distribution for a treatment site including at least a portion of a patient to be treated comprising inputting to the machine learning system a plurality of training data sets, each training data set comprising an actual spatial radiotherapy dose distribution for a treatment site and a training partial radiotherapy dose distribution, which is a part of the actual spatial radiotherapy dose distribution including dose information only for one or more regions of interest within the treatment site, and evaluating the output based on the actual spatial radiotherapy dose distribution.
2. The method of claim 1 , wherein the training data set also includes image data and/or structure data related to the treatment site.
3. A computer implemented method for generating an estimated spatial radiotherapy dose distribution for a treatment site including at least a portion of a patient to be treated, the method comprising the steps of inputting a planning partial radiotherapy dose distribution including dose information for one or more regions of interest within the treatment site to a machine learning system that has been trained according to claim 1 , and generating, by the machine learning system, an estimated spatial radiotherapy dose distribution for the treatment site based on the planning partial radiotherapy dose distribution.
4. The method of claim 3 , wherein the planning partial radiotherapy dose distribution is set by a human operator such as a clinician for the one or more regions of interest.
5. The method of claim 3 , wherein the planning partial radiotherapy dose distribution is determined based on a previously calculated plan for the patient, after a part of the previously calculated plan has been delivered to the patient.
6. The method of claim 3 , wherein the step of generating an estimated spatial radiotherapy dose distribution for the treatment site includes maintaining the dose information from the planning partial dose information.
7. The method of claim 3 , wherein the step of generating an estimated spatial radiotherapy dose distribution for the treatment site includes allowing the dose distribution in the one or more regions of interest to deviate from the dose information from the planning partial dose information only within certain limits.
8. The method of claim 3 , further comprising the step of evaluating the estimated spatial radiotherapy dose distribution and, depending on the result of the estimation, modifying the planning partial radiotherapy dose and generating an updated estimated spatial radiotherapy dose distribution based on the modified planning partial radiotherapy dose and outputting the updated estimated spatial radiotherapy dose distribution as the total dose.
9. The method of claim 3 , further comprising the step of performing dose mimicking based on the estimated spatial radiotherapy dose distribution.
10. The method of claim 3 , wherein the input data set further includes information regarding machine parameters and/or beam setup for a delivery apparatus to be used when delivering the radiotherapy.
11. The method of claim 3 , wherein the input data set further includes image data and/or structure data related to the treatment site.
12. The method of claim 1 , wherein the one or more regions of interest includes a target.
13. The method of claim 12 , wherein the one or more regions of interest includes one or more organs at risk.
14. The method of claim 1 , wherein the machine learning system is a U-Net or a V-Net.
15. A computer program product comprising computer-readable code means which, when executed in a computer will cause the computer to perform the method according to claim 3 .
16. A computer comprising a processor and a program memory, the program memory being arranged to hold a computer program product according to claim 15 .