Abstract: This disclosure relates generally to treatment management systems, which may include a clinical database for storing therapeutic protocols. The system may also include a treatment engine operatively connected to the clinical database. The treatment engine may obtain diagnostic information and select a first plurality of therapeutic protocols from the clinical database based on the obtained diagnostic information and reference protocol data. The treatment engine may calculate a treatment efficacy probability for each protocol using the reference protocol data. The treatment engine may develop a first treatment plan and evaluate intermediate data indicating an altered patient state due to the first treatment plan. The treatment engine may select, based on reference protocol data and adaptive protocol data, a second treatment plan using a second plurality of therapeutic protocols.

Abstract: Techniques for enhancing image segmentation with the integration of deep learning are disclosed herein. An example method for atlas-based segmentation using deep learning includes: applying a deep learning model to a subject image to identify an anatomical feature, registering an atlas image to the subject image, using the deep learning segmentation data to improve a registration result, generating a mapped atlas, and identifying the feature in the subject image using the mapped atlas. Another example method for training and use of a trained machine learning classifier, in an atlas-based segmentation process using deep learning, includes: applying a deep learning model to an atlas image, training a machine learning model classifier using data from applying the deep learning model, estimating structure labels of areas of the subject image, and defining structure labels by combining the estimated structure labels with labels produced from atlas-based segmentation on the subject image.

Abstract: Embodiments disclose a method and system for segmenting medical images. In certain embodiments, the system comprises a database configured to store a plurality of medical images acquired by an image acquisition device. The plurality of images include at least one first medical image of an object, and a second medical image of the object, each first medical image associated with a first structure label map. The system further comprises a processor that is configured to register the at least one first medical image to the second medical image, determine a classifier model using the registered first medical image and the corresponding first structure label map, and determine a second structure label map associated with the second medical image using the classifier model.

Abstract: Systems, computer-implemented methods, and computer readable media for generating a synthetic image of an anatomical portion based on an origin image of the anatomical portion acquired by an imaging device using a first imaging modality are disclosed. These systems may be configured to receive the origin image of the anatomical portion acquired by the imaging device using the first imaging modality, receive a convolutional neural network model trained for predicting the synthetic image based on the origin image, and convert the origin image to the synthetic image through the convolutional neural network model. The synthetic image may resemble an imaging of the anatomical portion using a second imaging modality differing from the first imaging modality.

Abstract: Systems and techniques may be used to estimate a real-time patient state during a radiotherapy treatment using a magnetic resonance linear accelerator (MR-Linac). For example, a method may include generating a dictionary of expanded potential patient measurements and corresponding potential patient states using a preliminary motion model. The method may include training, using a machine learning technique, a correspondence motion model relating an input patient measurement to an output patient state using the dictionary. The method may include estimating, using a processor, the patient state corresponding to a 2D MR image using the correspondence motion model. The method may include directing radiation therapy, using the MR-Linac, to a target according to the patient state.

Abstract: Described herein is a system and method of controlling real-time image-guided adaptive radiation treatment of at least a portion of a region of a patient. The computer-implemented method comprises obtaining a plurality of real-time image data corresponding to 2-dimensional (2D) magnetic resonance imaging (MRI) images including at least a portion of the region, performing 2D motion field estimation on the plurality of image data, approximating a 3-dimensional (3D) motion field estimation, including applying a conversion model to the 2D motion field estimation, determining at least one real-time change of at least a portion of the region based on the approximated 3D motion field estimation, and controlling the treatment of at least a portion of the region using the determined at least one change.

Abstract: Systems and techniques may be used to estimate a relative motion of patient anatomy using a deep learning network during a radiotherapy treatment. For example, a method may include using a first deep neural network to relate input real-time partial patient measurements and a patient model including a reference volume to output patient states. The method may include using a second deep neural network to relate the patient states and the reference volume to relative motion information between the patient states and the reference volume. The deep neural networks may be used in real time to estimate a relative motion corresponding to an input image.

Abstract: Apparatus and techniques are described herein for nuclear magnetic resonance (MR) projection imaging. Such projection imaging may be used to control radiation therapy delivery to a subject, such as including receiving reference imaging information, generating a two-dimensional (2D) projection image using imaging information obtained via nuclear magnetic resonance (MR) imaging, the 2D projection image corresponding to a specified projection direction, the specified projection direction including a path traversing at least a portion of an imaging subject, determining a change between the generated 2D projection image and the reference imaging information, and controlling delivery of the radiation therapy at least in part using the determined change between the obtained 2D projection image and the reference imaging information.

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.

Abstract: Features, such as anatomical features, may be automatically segmented from medical imaging information, using a computer-implemented method. In an example, three-dimensional (3D) medical imaging information may be received, such as defining a first volume. A first trained convolutional neural network (CNN) may be applied to the three-dimensional medical imaging information. An output from the first trained CNN may be used to determine a region-of-interest within the first volume, the region-of-interest defining a lesser, second volume. A different, second trained CNN may be applied to the region-of-interest, a segmented representation of the 3D medical imaging information may be provided using the outputs from the first and second CNNs, where the second CNN provides enhanced segmentation detail in the region-of-interest without requiring application of the second CNN to an entirety of the first volume. Techniques are also described from training one or more of the first and second CNNs.

Abstract: Embodiments of the disclosure may be directed to an image processing system configured to receive a medical image of a region of a subject's body taken at a first time and to receive a surface image of an exterior portion of the region of the subject's body taken at the first time. The image processing may also be configured to receive a medical image of the region of the subject's body taken at a second time and to register the medical image taken at the first time, the surface image taken at the first time, and the medical image taken at the second time.

Abstract: A deformable radiotherapy phantom can be produced using an additive manufacturing process, based on a medical image of the patient. The deformable phantom can include dosimeters for measuring radiation dose distribution. A smart material can allow deformation in response to an applied stimulus. Among other things, the phantom can be used to validate radiation dose warping, a radiotherapy treatment plan, to determine a maximum acceptable deformation of the patient, to validate a cumulative accuracy of dose warping and deformable image registration, or the like.

Abstract: An image segmentation method is disclosed. The method includes receiving a plurality of atlases and a subject image, each atlas including an atlas image showing a structure of interest and associated structure delineations, the subject image being acquired by an image acquisition device and showing the structure of interest. The method further includes calculating, by an image processor, mapped atlases by registering the respective atlases to the subject image, and determining, by the image processor, a first structure label map for the subject image based on the mapped atlases. The method also includes training, by the image processor, a structure classifier using a subset of the mapped atlases, and determining, by the image processor, a second structure label map for the subject image by applying the trained structure classifier to one or more subject image points in the subject image.

Abstract: Radiation treatment planning and administration can include a Monte Carlo computer simulation tool to simulate photo-generated electrons in tissue. In the simulation, electrons that have left tissue voxels and entered air voxels can be evaluated to identify electrons that are circling along a spiraling trajectory in the air voxels. After at least one full spiraling circumference or other specified distance has been traversed using a detailed electron transport model, a simpler linear ballistic motion model can be instituted. This speeds simulation while accurately accounting for spiraling electrons that re-enter tissue voxels.

Abstract: Systems and methods are provided for generating a pseudo-CT prediction model using multi-channel MR images. An exemplary system may include a processor configured to retrieve training data including multiple MR images and at least one CT image for each of a plurality of training subjects. For each training subject, the processor may determine at least one tissue parameter map based on the multiple MR images and obtain CT values based on the at least one CT image. The processor may also generate the pseudo-CT prediction model based on the tissue parameter maps and the CT values of the plurality of training subjects.

Abstract: Embodiments of the disclosure may be directed to a system for generating a motion target volume representative of shape changes of a target region in a patient. The system may comprise at least one computer system configured to receive a plurality of electronic medical images that include the target region, and each of the plurality of images may have been taken at a different time point. The computer system may be configured to define a three-dimensional volume containing the target region in each of the plurality of images, and the three-dimensional volume may be different in at least two of the plurality of images due to differences in shape of the target region in the at least two images. The computer system may also be configured to co-register the three-dimensional volumes and generate the motion target volume, wherein the motion target volume encompasses each of the three-dimensional volumes.

Abstract: Systems and methods include training a deep convolutional neural network (DCNN) to reduce one or more artifacts using a projection space or an image space approach. In a projection space approach, a method can include collecting at least one artifact contaminated cone beam computed tomography (CBCT) projection space image, and at least one corresponding artifact reduced, CBCT projection space image from each patient in a group of patients, and using the artifact contaminated and artifact reduced CBCT projection space images to train a DCNN to reduce artifacts in a projection space image. In an image space approach, a method can include collecting a plurality of CBCT patient anatomical images and corresponding registered computed tomography anatomical images from a group of patients, and using the plurality of CBCT anatomical images and corresponding artifact reduced computed tomography anatomical images to train a DCNN to remove artifacts from a CBCT anatomical image.

Abstract: The present disclosure relates to systems, methods, and computer-readable storage media for segmenting medical images. Embodiments of the present disclosure may relate to a method for segmenting medical images. The method may be implemented by a processor device executing a plurality of computer executable instructions. The method may comprise receiving an image from a memory, and identifying at least one landmark point within the image. The method may further comprise selecting an image point in the image, and determining at least one feature for the image point relative to the at least one landmark point. The method may also comprise associating the image point with an anatomical structure by using a classification model based on the at least one determined feature.

Abstract: Systems and methods for performing radiation treatment planning are provided. An exemplary system may include a processor device communicatively coupled to a memory device and configured to perform operations when executing instruction stored in the memory device. The operations may include receiving a reference treatment plan including one or more dose constraints and determining, based on the reference treatment plan, segment information of a plurality of radiation beams. The operations may also include determining a fluence map for each of the plurality of radiation beams based on the one or more dose constraints using a fluence map optimization algorithm. The operations may also include determining a dose distribution based on the fluence maps of the plurality of radiation beams. The operations may also include determining at least one beam modulation property of a new treatment plan using a warm-start optimization algorithm based on the segment information and the dose distribution.

Abstract: Systems and methods can include training a deep convolutional neural network model to provide a beam model for a radiation machine, such as to deliver a radiation treatment dose to a subject. A method can include determining a range of parameter values for at least one parameter of a beam model corresponding to the radiation machine, generating a plurality of sets of beam model parameter values, wherein one or more individual sets of beam model parameter values can include a parameter value selected from the determined range of parameter values, providing a plurality of corresponding dose profiles respectively corresponding to respective individual sets beam model parameter values in the plurality of sets of beam model parameter values, and training the neural network model using the plurality of beam models and the corresponding dose profiles.