System and method for automatic treatment planning

Patent metadata

Field	Value
Publication number	US-10046177-B2
Application number	US-201414308450-A
Country	US
Kind code	B2
Filing date	Jun 18, 2014
Priority date	Jun 18, 2014
Publication date	Aug 14, 2018
Grant date	Aug 14, 2018

Abstract

The present disclosure relates to systems, methods, and computer-readable storage media for radiotherapy. Embodiments of the present disclosure may receive a plurality of training data and determine one or more predictive models based on the training data. The one or more predictive models may be determined based on at least one of a conditional probability density associated with a selected output characteristic given one or more selected input variables or a joint probability density. Embodiments of the present disclosure may also receive patient specific testing data. In addition, embodiments of the present disclosure may predict a probability density associated with a characteristic output based on the one or more predictive models and the patient specific testing data. Moreover, embodiments of the present disclosure may generate a new treatment plan based on the prediction and may use the new treatment plan to validate a previous treatment plan.

First claim

What is claimed is:

1. A radiotherapy system for treating a target patient, the radiotherapy system including a radiotherapy device configured to perform a radiotherapy treatment to treat the target patient according to a treatment plan, the system comprising:
a data processing device configured to generate the treatment plan, the data processing device including:
a memory storing computer executable instructions; and
a processor device communicatively coupled to the memory, wherein the computer executable instructions, when executed by the processor device, cause the processor device to perform operations including:
receiving training data associated with past treatment plans used to treat sample patients, the training data including:
a plurality of observations associated with conditions of the sample patients, wherein the plurality of observations are derived from medical image data; and
one or more plan outcomes reflecting outcomes resulting from the past treatment plans, or plan parameters reflecting design parameters of the past treatment plans;

determining a joint probability density indicating a likelihood that both at least one particular observation and at least one particular plan outcome or plan parameter are present in the training data;
calculating a conditional probability based upon the determined joint probability density, wherein the conditional probability indicates a likelihood that the particular plan outcome or plan parameter is present in the training data;
receiving patient specific testing data associated with the target patient, the patient specific testing data including at least one patient specific observation associated with the target patient, wherein the patient specific observation is derived from medical image data;
predicting a probability of a patient specific plan outcome or plan parameter based on the conditional probability and the patient specific observation;
generating the treatment plan based on the prediction; and
controlling the radiotherapy device to perform the radiotherapy treatment according to the generated treatment plan.

2. The radiotherapy system of claim 1 , wherein the training data comprise a plurality of training samples.

3. The radiotherapy system of claim 1 , wherein the training data comprise a plurality of images.

4. The radiotherapy system of claim 3 , wherein the training data comprise a training sample and the training sample includes characteristics of a voxel in an image.

5. The radiotherapy system of claim 3 , wherein the plurality of images comprise at least one of a Magnetic Resonance Imaging (MRI) image, a 30 MRI image, a 2D streaming MRI image, a 4D volumetric MRI image, a Computed Tomography (CT) image, a Cone-Beam CT image, a Positron Emission Tomography (PET) image, a functional MRI (fMRI) image, an X-ray image, a fluoroscopic image, an ultrasound image, a radiotherapy portal image, or a single-photo emission computed tomography (SPECT) image.

6. The radiotherapy system of claim 1 , wherein the past treatment plans are from a current patient, a plurality of other patients, or a combination thereof.

7. The radiotherapy system of claim 1 , wherein the past treatment plans are from at least one of a single patient or a plurality of patients.

8. The radiotherapy system of claim 1 , wherein the computer executable instructions additionally cause the processor device to calculate a probability that the particular observation is present in the training data: and
wherein the conditional probability indicates a likelihood that the particular plan outcome or plan parameter is present in the training data given the probability that the particular observation is present in the training data.

9. The radiotherapy system of claim 8 , wherein determining the joint probability density or calculating the conditional probability comprises using at least one of a non-parametric method, a parametric method, a Monte Carlo based method, a regression method, a machine learning method, or combinations thereof.

10. A method for operating a radiotherapy system to perform a radiotherapy treatment for treating a target patient, comprising:
receiving training data associated with past treatment plans used to treat sample patients, the training data including a plurality of training samples, each of the training samples including a feature vector and an output vector corresponding to the feature vector, wherein:
the feature vector includes one or more observations associated with conditions of the sample patients, wherein the one or more observations are derived from medical image data; and
the output vector includes one or more plan outcomes reflecting outcomes resulting from the past treatment plans, or plan parameters reflecting design parameters of the past treatment plans;

determining, by a processor device, a joint probability density associated with the feature vector and the corresponding output vector, wherein the joint probability density reflects a likelihood that both the one or more observations of the feature vector and the one or more plan outcomes or plan parameters of the output vector are present in the training data;
generating, by the processor device, one or more predictive models for predicting future plan outcomes or plan parameters for the target patient based on the joint probability density, each predictive model including a conditional probability based upon the determined joint probability density, wherein the conditional probability indicates a likelihood that one or more particular plan outcomes or plan parameters are present in the training data given a probability that one or more particular observations are present in the training data;
storing the one or more predictive models in a memory;
receiving patient specific testing data associated with the target patient, the patient specific testing data including a patient specific feature vector, wherein the patient specific feature vector includes at least one patient specific observation associated with the target patient, wherein the at least one patient specific observation is derived from medical image data;
determining, by the processor device, a probability density of the patient specific observation in the patient specific testing data;
predicting, by the processor device, a probability density of a patient specific output vector based on the probability density of the patient specific observation and the conditional probability;
generating a treatment plan based on the prediction; and
performing, by a radiotherapy device, the radiotherapy treatment to treat the target patient according to the treatment plan.

11. The method of claim 10 wherein determining the conditional probability comprises using at least one of a non-parametric method, a parametric method, a Monte Carlo based method, a regression method, a machine learning method, or combinations thereof.

12. The method of claim 10 , wherein determining the joint probability density comprises using at least one of a non-parametric method, a parametric method, a Monte Carlo based method, a regression method, a machine learning method, or combinations thereof.

13. The method of claim 10 , wherein the patient specific testing data comprises at least one of imaging data, organ or volume of interest segmentation data, functional organ modeling data, radiation dosage, laboratory data, genomic data, demographics, other diseases affecting the patient, medications and drug reactions, diet and lifestyle, environmental risk factors, tumor characteristics, genetic/protein biomarkers, or previous medical treatments of the patient.

14. The method

Assignees

• Elekta Ab
• Elekta Inc
• Elekta Inc

Inventors

• Sjölund Jens Olof
• Han Xiao

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• A61N5/1039

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

• A61N2005/1041

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

• G06N5/04 Primary

  Inference or reasoning models · CPC title

• A61N5/103 Primary

  Treatment planning systems · CPC title