Abstract: A candidate generator generates a set of candidate three-dimensional image patches from an input volume. A candidate classifier classifies the set of candidate three-dimensional image patches as containing or not containing disease. Classifying the set of candidate three-dimensional image patches comprises generating an attention mask for each given candidate three-dimensional image patch within the set of candidate three-dimensional image patches to form a set of attention masks, applying the set of attention masks to the set of candidate three-dimensional image patches to form a set of masked image patches, and classifying the set of masked image patches as containing or not containing the disease. The candidate classifier applies soft attention and hard attention to the three-dimensional image patches such that distinctive image regions are highlighted proportionally to their contribution to classification while completely removing image regions that may cause confusion.

Abstract: Methods and systems for training computer-aided condition detection systems. One method includes receiving a plurality of images for a plurality of patients, some of the images including an annotation associated with a condition; iteratively applying a first deep learning network to each of the images to produce an attention map, a feature map, and an image-level probability of the condition for each of the images; iteratively applying a second deep learning network to each feature map produced by the first network to produce a plurality of outputs; training the first network based on the attention map produced for each image; and training the second network based on the output produced for each of the patients. The second network includes a plurality of convolution layers and a plurality of convolutional long short-term memory (LSTM) layers. Each of the outputs includes a patient-level probability of the condition for one of the patients.

Abstract: A system for medical device deployment includes an optical shape sensing (OSS) system (104) associated with a deployable medical device (102) or a deployment instrument (107). The OSS system is configured to measure shape, position or orientation of the deployable medical device and/or deployment instrument. A registration module (128) is configured to register OSS data with imaging data to permit placement of the deployable medical device. An image processing module (142) is configured to create a visual representation (102?) of the deployable medical device and to jointly display the deployable medical device with the imaging data.

Abstract: Methods and systems for training computer-aided condition detection systems. One method includes receiving a plurality of images for a plurality of patients, some of the images including an annotation associated with a condition; iteratively applying a first deep learning network to each of the images to produce a segmentation map, a feature map, and an image-level probability of the condition for each of the images; iteratively applying a second deep learning network to each feature map produced by the first network to produce a plurality of outputs; training the first network based on the segmentation map produced for each image; and training the second network based on the output produced for each of the patients. The second network includes a plurality of convolution layers and a plurality of convolutional long short-term memory (LSTM) layers. Each of the outputs includes a patient-level probability of the condition for one of the patients.

Abstract: Mechanisms are provided to implement an automated echocardiograph measurement extraction system. The automated echocardiograph measurement extraction system receives medical imaging data comprising one or more medical images and inputs the one or more medical images into a deep learning network. The deep learning network automatically processes the one or more medical images to generate an extracted echocardiograph measurement vector output comprising one or more values for echocardiograph measurements extracted from the one or more medical images. The deep learning network outputs the extracted echocardiograph measurement vector output to a medical image viewer.

Abstract: Mechanisms are provided to implement an automated medical image processing pipeline selection (MIPPS) system. The MIPPS system receives medical image data associated with a patient electronic medical record and analyzes the medical image data to extract evidence data comprising characteristics of one or more medical images in the medical image data indicative of a medical image processing pipeline to select for processing the one or more medical images. The evidence data is provided to a machine learning model of the MIPPS system which selects a medical image processing pipeline based on a machine learning based analysis of the evidence data. The selected medical image processing pipeline processes the medical image data to generate a results output.

Abstract: An interventional tool stepper (30) employing a frame (31), a carriage (33), an optional gear assembly (32), and an optional grid template(34). The frame (31) is structurally configured to be positioned relative to an anatomical region for holding an interventional tool (40) relative to the anatomical region. The carriage (33) is structurally configured to hold the interventional tool (40) relative to the anatomical region. The gear assembly (32) is structurally configured to translate and/or rotate the carriage (33) relative to the frame (31). The grid template (34) is structurally configured to guide one or more additional interventional tools (41) relative to the anatomical region. The frame (31), the carriage (33), the optional gear assembly (32) and the optional grid template (34) have an electromagnetic-compatible material composition for minimizing any distortion by the interventional tool stepper (30) of an electromagnetic field.

Abstract: An interventional therapy system may include at least one catheter configured for insertion within an object of interest (OOI); and at least one controller which configured to: obtain a reference image dataset including a plurality of image slices which form a three-dimensional image of the OOI; define restricted areas (RAs) within the reference image dataset; determine location constraints for the at least one catheter in accordance with at least one of planned catheter intersection points, a peripheral boundary of the OOI and the RAs defined in the reference dataset; determine at least one of a position and an orientation of the distal end of the at least one catheter; and/or determine a planned trajectory for the at least one catheter in accordance with the determined at least one position and orientation for the at least one catheter and the location constraints.

Abstract: Methods and systems for training computer-aided condition detection systems. One method includes receiving a plurality of images for a plurality of patients, some of the images including an annotation associated with a condition; iteratively applying a first deep learning network to each of the images to produce a segmentation map, a feature map, and an image-level probability of the condition for each of the images; iteratively applying a second deep learning network to each feature map produced by the first network to produce a plurality of outputs; training the first network based on the segmentation map produced for each image; and training the second network based on the output produced for each of the patients. The second network includes a plurality of convolution layers and a plurality of convolutional long short-term memory (LSTM) layers. Each of the outputs includes a patient-level probability of the condition for one of the patients.

Abstract: Methods and systems for training computer-aided condition detection systems. One method includes receiving a plurality of images for a plurality of patients, some of the images including an annotation associated with a condition; iteratively applying a first deep learning network to each of the images to produce an attention map, a feature map, and an image-level probability of the condition for each of the images; iteratively applying a second deep learning network to each feature map produced by the first network to produce a plurality of outputs; training the first network based on the attention map produced for each image; and training the second network based on the output produced for each of the patients. The second network includes a plurality of convolution layers and a plurality of convolutional long short-term memory (LSTM) layers. Each of the outputs includes a patient-level probability of the condition for one of the patients.

Abstract: A candidate generator generates a set of candidate three-dimensional image patches from an input volume. A candidate classifier classifies the set of candidate three-dimensional image patches as containing or not containing disease. Classifying the set of candidate three-dimensional image patches comprises generating an attention mask for each given candidate three-dimensional image patch within the set of candidate three-dimensional image patches to form a set of attention masks, applying the set of attention masks to the set of candidate three-dimensional image patches to form a set of masked image patches, and classifying the set of masked image patches as containing or not containing the disease. The candidate classifier applies soft attention and hard attention to the three-dimensional image patches such that distinctive image regions are highlighted proportionally to their contribution to classification while completely removing image regions that may cause confusion.

Abstract: A system for dynamic localization of medical instruments includes an ultrasound imaging system (110) configured to image a volume where one or more medical instruments are deployed. A registration module (136) registers two images of the one or more medical instruments to compute a transform between the two images, the two images being separated in time. A planning module (142) is configured to have positions of the volume and the one or more medical instruments updated based on the transform and, in turn, update a treatment plan in accordance with the updated positions such that changes in the volume and positions of the one or more medical instruments are accounted for in the updated plan.

Abstract: Systems and methods for developing a disease detection model. One method includes training the model using an image study and an associated disease label mined from a radiology report. The image study including a sequence of a plurality of two-dimensional slices of a three-dimensional image volume, and the model including a convolutional neural network layer and a convolutional long short-term memory layer. Training the model includes individually extracting a set of features from each of the plurality of two-dimensional slices using the convolutional neural network layer, sequentially processing the features extracted by the convolutional neural network layer for each of the plurality of two-dimensional slices using the convolutional long short-term memory layer, processing output from the convolutional long short-term memory layer for each of the plurality of two-dimensional slices to generate a probability of the disease, and updating the model based on comparing the probability to the label.

Abstract: An intervention system employs an optical shape sensing tool (32) (e.g., a brachytherapy needle having embedded optical fiber(s)) and a grid (50, 90) for guiding an insertion of the optical shape sensing tool (32) into an anatomical region relative to a grid coordinate system. The intervention system further employs a registration controller (74) for reconstructing a segment or an entirety of a shape of the optical shape sensing tool (32) relative to a needle coordinate system, and for registering the needle coordinate system to the grid coordinate system as a function of a reconstructed segment/entire shape of the optical shape sensing tool (32) relative to the grid (50, 90) (i.e., reconstruction of a segment/entire shape of the OSS needle inserted into/through the grid serving as a basis for the grid/needle coordinate system registration).

Abstract: Systems and methods for developing a disease detection model. One method includes training the model using an image study and an associated disease label mined from a radiology report. The image study including a sequence of a plurality of two-dimensional slices of a three-dimensional image volume, and the model including a convolutional neural network layer and a convolutional long short-term memory layer.

Abstract: Mechanisms are provided to implement an automated medical image processing pipeline selection (MIPPS) system. The MIPPS system receives medical image data associated with a patient electronic medical record and analyzes the medical image data to extract evidence data comprising characteristics of one or more medical images in the medical image data indicative of a medical image processing pipeline to select for processing the one or more medical images. The evidence data is provided to a machine learning model of the MIPPS system which selects a medical image processing pipeline based on a machine learning based analysis of the evidence data. The selected medical image processing pipeline processes the medical image data to generate a results output.

Abstract: An ultrasound device (18) is arranged to acquire an ultrasound image (19) of a thoracic diaphragm. A current location (24) of a tumor is determined using the ultrasound image (19) of the thoracic diaphragm and a predetermined relationship (14) assigning tumor locations to a set of simulation phase ultrasound images of the thoracic diaphragm in different geometries, or using a predetermined relationship (114) assigning tumor locations to a set of meshes representing the thoracic diaphragm in different geometries. The tumor may, for example, be a lung tumor.

Abstract: Mechanisms are provided to implement an automated medical image processing pipeline selection (MIPPS) system. The MIPPS system receives medical image data associated with a patient electronic medical record and analyzes the medical image data to extract evidence data comprising characteristics of one or more medical images in the medical image data indicative of a medical image processing pipeline to select for processing the one or more medical images. The evidence data is provided to a machine learning model of the MIPPS system which selects a medical image processing pipeline based on a machine learning based analysis of the evidence data. The selected medical image processing pipeline processes the medical image data to generate a results output.

Abstract: A mechanism is provided in a data processing system comprising a processor and a memory, the memory comprising instructions that are executed by the processor to specifically configure the processor to implement an intelligent medical image viewing engine. The intelligent medical image viewing engine receives a medical imaging study data structure comprising a plurality of electronic medical images from a medical image database. An image processing component executing within the intelligent medical image viewing engine analyzes the medical imaging study data structure to identify, for each electronic medical image in the plurality of electronic medical images, a corresponding set of image attributes.

Abstract: An imaging apparatus (24) images an introduction element (17) like a needle or a catheter for performing a biopsy or a brachytherapy. The introduction element (17) includes at least one ultrasound receiver (21) arranged at a known location. An ultrasound probe (12) for being inserted into a living being (2) emits ultrasound signals for acquiring ultrasound data of an inner part (19) of the living being (2). A first tracking unit (3) tracks the location of the introduction element (17) based on a reception of the emitted ultrasound signals by the at least one ultrasound receiver (21). An imaging unit (4) generates an indicator image showing the inner part (19) and an indicator of the introduction element (17) based on the tracked location. A display (5) displays the indicator image providing feedback about the location of the introduction element (17).