Abstract: Embodiments described herein provide for implementing a language model in a use-case for Tumor Board meetings for radiotherapy treatment planning (RTTP) efforts and treatment planning assistance, which requires high-levels of accuracy and/or precision in the outputs generated by the LLM and presented to members of the Tumor Board participating in a Tumor Board discussion, which may include live meetings or asynchronous online discussions. Tumor Board Application (TBA) software collects from discussions of a Tumor Board meeting to train the LLM on predicting outputs that contribute information about the patient, proposed RTTP, or aspects of the patient treatment. An AI agent participates in the Tumor Board discussion to ingest the inputs of the members of the Tumor Board and output the responsive text produced by the LLM, thereby allowing the LLM-powered AI-agent to interact with Tumor Board discussions.

Abstract: Systems and methods for the automatic generation and optimization of radiation therapy treatment plans, and systems and methods for the automatic generation and optimization of an adapted plan in an adaptive radiation therapy workflow.

Abstract: An apparatus for developing an intensity-modulated radiation therapy treatment plan includes a memory that stores machine instructions and a processor that executes the machine instructions to receive a clinical goal associated with the treatment plan as a user input. The processor further executes the machine instructions to determine a plan objective based on the clinical goal, generate a cost function comprising a term based on the plan objective, and assign an initial value to a parameter associated with the term. The processor also executes the machine instructions to identify a microstate that results in a reduced value associated with the cost function, evaluate a fulfillment level associated with the clinical goal, and adjust the value of the parameter to improve the fulfillment level.

Abstract: Systems, devices, and methods for increasing the dose spread in treatment systems delivering non-coplanar treatments to a patient using a couch with limited couch rotation angle ranges, the increasing including combining dose spreads generated by shifting the isocenter location and shifting the couch angles in such a way as to maximize the dose spread in both sides of the patient, and systems, devices and methods for generating non-coplanar treatment plans with different couch-kick ranges from which the non-coplanar treatment plan that takes into account the couch shifts needed for correcting patient positioning is selected for delivery.

Abstract: Disclosed herein are systems and methods for iteratively training artificial intelligence models using reinforcement learning techniques including a method comprising executing, using at least one patient attribute, a number of arcs, and at least one clinical goal attribute associated with a volumetric modulated arc therapy (VMAT) treatment of a patient, an artificial intelligence model configured to predict a number of control points and a number of dose calculation sectors for the VMAT treatment, the artificial intelligence model having been iteratively trained using a reinforcement learning method, where with each iteration, the artificial intelligence model transmits one or more test control point and one or more test dose calculation sector to a plan optimizer model configured to predict a treatment plan, and trains a policy in accordance with calculated rewards.

Abstract: A control circuit accesses patient information for a patient and optimizes a radiation treatment plan for that patient as a function of both the patient information and a dose distribution objective function, wherein a corresponding available solution space is limited as a function of at least one three-dimensional conformal solution.

Abstract: Disclosed herein are methods and systems to evaluate cost values for different radiotherapy treatment plans using external AI models. A method comprises presenting a user interface providing an interaction interface between a plurality of medical professionals communicating regarding a radiation therapy treatment of a patient during a tumor board meeting or a radiotherapy treatment planning process; receiving, from the interaction interface, a first input comprising a first patient attribute of the patient and a second input corresponding to the radiation therapy treatment of the patient; executing a machine learning language processing model using the first input and the second input to predict a task associated with generating a treatment plan for the patient; receiving, from the interaction interface, a third input; and when the third input does not correspond to the predicted task, presenting an indication of the predicted task.

Abstract: During a radiation treatment plan optimization loop, a control circuit can conduct a dosimetric optimization iteration to yield a dosimetric-based plan result and then conduct a non-dosimetric optimization iteration to yield a non-dosimetric-based plan result. The control circuit can then assess the dosimetric-based plan result and the non-dosimetric-based plan result to yield a convergence assessment result. The latter can then be taken into account when determining whether to conclude continued radiation treatment plan optimization loops.

Abstract: Systems and methods are disclosed for optimizing a treatment plan using all degrees of freedom including those related to beam geometry parameters, the optimization including a step for limiting the search space for the beam geometry parameters using a trained machine learning model, and systems and methods are disclosed for obtaining beam geometry parameters for treatment planning that do not require knowledge of the beam delivery device isocenter.

Abstract: Embodiments described herein provide for radiotherapy treatment plan generation using a machine learning language processing model. A processor can present a user interface providing an interaction interface between a user and a machine learning language processing model. The processor can receive a first input comprising a first patient attribute of a patient. The processor can execute the machine learning language processing model using the first patient attribute as an input to generate a response requesting a second patient attribute. The processor can present the response requesting the second patient attribute of the patient. The processor can receive a second input comprising the second patient attribute of the patient. The processor can transmit the first patient attribute of the patient and the second patient attribute of the patient to a radiotherapy plan optimizer. The radiotherapy plan optimizer can be configured to generate a radiotherapy treatment plan for the patient.

Abstract: Disclosed herein are methods and systems to evaluate cost values for different radiotherapy treatment plans using external AI models. A method comprises receiving a radiation therapy plan objective for a patient; executing a plan optimizer to generate one or more treatment attributes for a treatment plan complying with the radiation therapy plan objectives, the plan optimizer iteratively calculating the one or more attributes, where with each iteration, the plan optimizer revises the one or more attributes of the treatment plan in accordance with a cost value; executing an AI model to calculate a second cost value for the treatment plan, wherein the AI model is trained to calculate the second cost value in accordance with a likelihood of occurrence of a health-problem for the patient after being treated via the treatment plan having the one or more attributes; and outputting the treatment plan for the patient.

Abstract: In a radiation treatment plan that includes a plurality of treatment fields of multiple treatment modalities, such as IMRT modality and dynamic treatment path modality (e.g., VMAT and conformal arc therapy), an optimized spatial point sequence may be determined that optimizes the total treatment time, which includes both the beam-on time (i.e., during the delivery of radiation dose) and the beam-off time (i.e., during transitions between consecutive treatment fields). The result is a time-ordered field trajectory that intermixes and interleaves different treatment fields. In one embodiment, a dynamic treatment path may be cut into a plurality of sections, and one or more IMRT fields may be inserted between the plurality of sections.

Abstract: Systems, devices and methods for using similarity metrics to determine optimization strategy in adaptive radiation therapy, and systems and methods for an automated adaptive workflow to automatically adapt and optimize a treatment plan to a current treatment session using MLC leaf configurations selected based on calculated similarity metric values.

Abstract: A method of generating a treatment plan for treating a patient with radiotherapy, the method includes obtaining a plurality of sample plans, which are generated by use of a knowledge base comprising historical treatment plans and patient data. The method also includes performing a multi-criteria optimization based on the plurality of sample plans to construct a Pareto frontier, where the plurality of sample plans are evaluated with at least two objectives measuring qualities of the plurality of sample plans such that treatment plans on the constructed Pareto frontier are Pareto optimal with respect to the objectives. The method further includes identifying a treatment plan by use of the constructed Pareto frontier.

Abstract: A control circuit, while optimizing a radiation treatment plan for a particular patient, outsources an optimization calculation to an external resource and then receives from that external resource a resultant optimization calculation. By one approach, that optimization calculation comprises an optimization high-level utility function calculation. The external resource may comprise, for example, a third-party resource.

Abstract: A control circuit accesses a memory having stored therein a plurality of hierarchically-diversified radiation treatment planning templates. These templates include at least a first radiation treatment planning template that specifies radiation treatment planning information at a first hierarchical level. These templates also include at least a second radiation treatment planning template that specifies radiation treatment planning information at a second hierarchical level, wherein the second hierarchical level is more granular than the first hierarchical level. By one approach, the control circuit may access a plurality of differing ones of the second radiation treatment planning templates wherein each such template specifies radiation treatment planning information at the second hierarchical level.

Abstract: A control circuit accesses patient information including anatomical image information of the patient, segmentation information corresponding to the anatomical image information, and a dose map for the radiation treatment plan. The control circuit then generates at least one organ-specific three-dimensional risk map as a function of the patient information and presents that risk map to a user via a display.

Abstract: 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.

Abstract: Disclosed herein are systems and methods for identifying radiation therapy treatment data for patients. A processor accesses a neural network trained based on a first set of data generated from characteristic values of a first set of patients that received treatment at one or more first radiotherapy machines. The processor executes the neural network using a second set of data comprising characteristic values of a second set of patients receiving treatment at one or more second radiotherapy machines. The processor executes a calibration model using an output of the neural network based on the second set of data to output a calibration value. The processor executes the neural network using a set of characteristics of a first patient to output a first confidence score associated with a first treatment attribute. The processor then adjusts the first confidence score according to the calibration value to predict the first treatment attribute.

Abstract: Systems and methods for radiation treatment planning can include a computing system determining a first shell structure defined around a PTV and having a first thickness, and a second shell structure defined around the first shell structure and having a second thickness. The computing system can generate a first objective term of an objective function for optimizing a radiotherapy treatment plan to penalize dose values in the first shell structure exceeding a first radiation dose level, and generate a second objective term of the objective function to penalize dose values in the second shell structure exceeding a second radiation dose level. The second fraction can be smaller than the first fraction. The computing system can generate a third objective term of an objective function for penalizing radiation dose values in another region deviating from a predefined dose distribution, and optimize the objective function to determine the radiotherapy treatment plan.