A radiation treatment plan three-dimensional dose prediction machine learning model is trained using a training corpus that includes a plurality of radiation treatment plans that are not specific to a particular patient and wherein the training corpus includes some, but not all, possible patient volumes of interest. Information regarding the patient (including information regarding at least one volume of interest for the patient that was not represented in the training corpus) is input to the radiation treatment plan three-dimensional dose prediction machine model. The latter generates predicted three-dimensional dose distributions that include a predicted three-dimensional dose distribution for the at least one volume of interest that was not represented in the training corpus.

Systems and methods are disclosed for monitoring anatomic position of a human subject and modifying a radiotherapy treatment based on anatomic position changes, as determined with a regression model trained to estimate movement of a region of interest. Example operations for movement monitoring and therapy control include: obtaining 3D image data for a subject, which provides a reference volume and at least one defined region of interest; obtaining real-time 2D image data corresponding to the subject, captured during the radiotherapy treatment session; extracting features from the 2D image data; producing a relative motion estimation of a region of interest with a machine learning regression model, the model trained to estimate a spatial transformation from the 2D image data based on training from the reference volume; and controlling a radiotherapy beam of a radiotherapy machine used in the radiotherapy session, based on the relative motion estimation.

Disclosed herein are systems and methods for iteratively training artificial intelligence models using reinforcement learning techniques. With each iteration, a training agent applies a random radiation therapy treatment attribute corresponding to the radiation therapy treatment attribute associated with previously performed radiation therapy treatments when an epsilon value indicative of a likelihood of exploration and exploitation training of the artificial intelligence model satisfies a threshold. When the epsilon value does not satisfy the threshold, the agent generates, using an existing policy, a first predicted radiation therapy treatment attribute, and generates, using a predefined model, a second predicted radiation therapy treatment attribute. The agent applies one of the first predicted radiation therapy treatment attribute or the second predicted radiation therapy treatment attribute that is associated with a higher reward. The agent iteratively repeats training the artificial intelligence model until the existing policy satisfies an accuracy threshold.

Systems and methods for shuttle mode radiation delivery are described herein. One method for radiation delivery comprises moving the patient platform through the patient treatment region multiple times during a treatment session. This may be referred to as patient platform or couch shuttling (i.e., couch shuttle mode). Another method for radiation delivery comprises moving the therapeutic radiation source jaw across a range of positions during a treatment session. The jaw may move across the same range of positions multiple times during a treatment session. This may be referred to as jaw shuttling (i.e., jaw shuttle mode). Some methods combine couch shuttle mode and jaw shuttle mode. Methods of dynamic or pipelined normalization are also described.

Techniques for solving a radiotherapy treatment plan optimization problem are provided. The techniques include receiving a radiotherapy treatment plan optimization problem; processing the radiotherapy treatment plan optimization problem with a machine learning model to estimate one or more optimization variables of the radiotherapy treatment plan optimization problem, wherein the machine learning model is trained to establish a relationship between the one or more optimization variables and parameters of a plurality of training radiotherapy treatment plan optimization problems; and generating a solution to the radiotherapy treatment plan optimization problem based on the estimated one or more optimization variables of the radiotherapy treatment plan optimization problem.

Nominal values of parameters, and perturbations of the nominal values, that are associated with previously defined radiation treatment plans are accessed. For each treatment field of the treatment plans, a field-specific planning target volume (fsPTV) is determined based on those perturbations. At least one clinical target volume (CTV) and at least one organ-at-risk (OAR) volume are also delineated. Each OAR includes at least one sub-volume that is delineated based on spatial relationships between each OAR and the CTV and the fsPTV for each treatment field. Dose distributions for the sub-volumes are determined based on the nominal values and the perturbations. One or more dose prediction models are generated for each sub-volume. The dose prediction model(s) are trained using the dose distributions.

The present disclosure discloses a standardized artificial intelligence automatic radiotherapy planning method and system, wherein the radiation therapy planning method includes: acquiring a medical image; automatically delineating an ROI area of the medical image to acquire a geometric anatomical structure; determining a prescription according to disease type information corresponding to the medical image, the geometric anatomical structure, and a preset disease-prescription template library, and determining a radiation angle of radiation therapy; obtaining a radiation therapy dose distribution result using a dose prediction model; performing optimization processing using a reverse optimization algorithm based on dose distribution or DVH guidance, with reference to the radiation dose distribution result, to generate executable radiation therapy plans. The technical solution of the present disclosure realizes fully automatic dose prediction, improves efficiency and effect of dose prediction, so that an executable radiation therapy plan can be generated quickly and with high quality, with good accuracy, stability and standardization, and can edit and adjust the dose distribution visually and directly, greatly improving efficiency of plan design.

A tool for radiation therapy simulation or planning is disclosed which aids in avoiding collisions during treatment. Configurations of components including at least a radiation delivery device (30) and a patient (32) are generated. Each configuration defines positions of the components in a common coordinate system. For each configuration, proximities of pairs of components of the configuration are computed using ray tracing between three-dimensional surface models (30m, 32m, 36m, 38m) representing the components of the pair. A collision is identified as any pair of components having a computed proximity that is less than a margin for the pair of components. Each identified collision is displayed on a display (12), e.g. as a rendering. The simulations or planning may be used to verify deliverability of arc, 4Pi, or static therapy, to determine safety margins for collisions, to calculate and display realizable trajectories, and so forth.

The present disclosure relates to systems and methods for radiotherapy planning. The systems may obtain a delineation of a region of interest (ROI) in an image of an object. The ROI may include at least one target region. The systems may obtain modified delineation of the ROI based on one or more modifications to the delineation of the ROI. The systems may determine a target radiotherapy plan of the object by performing a radiotherapy dose optimization on the ROI. The modifications of the delineation of the ROI and the radiotherapy dose optimization of the ROI may be performed at least partially overlap temporally.

A resource management system for better operation of a plurality of devices. The system comprises an input interface (IN) for receiving input data including one or more characteristics of at least one work object (P1-P3) and/or including context data. The at least one work object (P1-P3) can be processed by one or more processing devices (M.sub.ij). The said processing is specified in a respective work specification (S1-S3). A predictor component (PC) of the system is configured to predict, based on the input data, a change to the work specification. The system comprises an output interface (OUT) for providing output data that represents said predicted change.