FIELD: medicine.
SUBSTANCE: group of inventions relates to medical engineering, specifically to means for training and/or forecasting data when developing a treatment plan for radiation therapy. Radiation therapy system for treating a target patient, using a radiation therapy device, configured to perform radiation therapy in accordance with a treatment plan, comprises a data processing device for creating a treatment plan, including memory in which computer-executable instructions are stored, 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 obtaining training data associated with past treatment plans used to treat selected patients, wherein the training data includes observations related to the condition of selected patients, obtained from medical image data, and at least one plan result indicating the result obtained from the past treatment plan, or a plan parameter indicating the estimated parameter of the past treatment plan, determining a joint probability density indicating the probability that at least one particular observation, and at least one particular plan result or plan parameter are present in the training data, calculating conditional probability based on the determined joint probability density, wherein the conditional probability indicates the probability that a particular plan result or plan parameter is present in the training data, obtaining patient-specific test data associated with the target patient, including at least one patient-specific observation associated with the target patient and obtained from medical image data, predicting the likelihood of a patient-specific plan result or plan parameter based on conditional probability and patient-specific observation, creating a treatment plan based on prediction and prescribing a radiation therapy device to perform radiation therapy in accordance with the treatment plan created. Method of operation is carried out through the radiation therapy system. In the second embodiment of the system, the processor device is configured to obtain training data associated with past treatment plans, wherein the training data includes training samples, each of the training samples includes a feature vector and an output vector corresponding to the feature vector, wherein the feature vector includes at least one observation related to a previous treatment plan, and the output vector includes at least one plan result plan or plan parameter, determining the joint probability density associated with the feature vector and the corresponding output vector, wherein the joint probability density indicates the probability that both at least one observation of the feature vector and at least one plan result or plan parameter of the output vector are present in the training data, determining the conditional probability density associated with the output vector, taking into account the probability density of the feature vector in the training data, wherein the conditional probability density indicates the probability that at least one plan result or plan parameter of the feature vector is present in the training data, provided that at least one observation of the feature vector is present, or creating at least one predictive model based on at least one of the joint probability density or conditional probability density; wherein each predictive model includes the probability that at least one particular plan result or plan parameter is present in the training data, taking into account the probability that at least one particular observation is present in the training data, the probability that both at least one particular observation and at least one specific plan result or plan parameter are present in the training data, storing at least one predictive model in memory, obtaining patient-specific test data associated with the patient, wherein the patient-specific test data includes a patient-specific feature vector, wherein the patient-specific feature vector includes at least one patient-specific observation related to the patient, determining the probability density of a patient-specific feature vector in the patient-specific test data, predicting the probability density of a patient-specific output vector based on the probability density of the patient-specific feature vector and at least one predictive model; and creating a prediction-based treatment plan, and prescribing the radiation therapy device to perform radiation therapy in accordance with the treatment plan.
EFFECT: use of inventions improves the effectiveness of creating a treatment plan in radiation therapy.
21 cl, 5 dwg
