3 results about "Computed tomography angiography" patented technology

This invention discloses a preoperative risk prediction method for cerebral aneurysms based on multimodal deep learning. To address the problems of time-consuming computational fluid dynamics simulations and difficulty in quickly obtaining individualized hemodynamic indicators, this invention reconstructs vascular geometry and a tree diagram by aligning computed tomography angiography, magnetic resonance angiography, and blood pressure and heart rate time series. It employs a graph network that satisfies Kirchhoff conservation and connects differentiable Wendt-Kessel impedance learning boundary conditions. Combining physical constraints of cardiac phase conditional neural operators, Helmholtz projection, and signed distance function boundary embedding, it approximates the flow field, calculates and closes the loop to correct wall shear stress, oscillatory shear exponent, and pressure gradient. Then, it fuses the risk features of the imaging branch, clinical branch, and text branch through expert product to output lesion-level and patient-level risks and confidence levels. This achieves the technical effect of rapidly estimating hemodynamics and performing interpretable preoperative risk assessment under physically consistent constraints.

Embodiments herein rapidly transform computed tomography angiography (CTA) data into an interactive, holographic visualization for clinical use. In one embodiment, a system for visualizing computed tomography angiography (CTA) images comprises a data interface configured to receive CTA image data in a Digital Imaging and Communications in Medicine (DICOM) format. The system also comprises a processor configured to render a plurality of two-dimensional CTA slice images into a three-dimensional (3D) volumetric dataset from the CTA image data, to filter the 3D volumetric dataset to resolve vascular structures relative to surrounding tissue and bone, and to generate an extended reality (XR) holographic representation of the filtered 3D volumetric dataset. The system also comprises an XR display device configured to present the holographic representation to a user for the user to manipulate the holographic representation and identify a vascular abnormality.

Systems and methods are provided for automated image analysis of volumetric image data depicting blood vessels. Volumetric imaging data, such as computed tomography angiography data, are processed by an image segmenter implemented as a convolutional neural network to generate a lumen mask identifying vessel lumen within the volumetric image data. In implementations, the image segmenter predicts binary masks on multiple two-dimensional slices extracted from the volumetric image data along different axes and / or angles, and the predicted masks are combined by voting, averaging, and / or nonlinear filtering to form a three-dimensional lumen segmentation. A three-dimensional segmented vessel model is formed from the lumen segmentation, a narrowing region is detected, and dimensional parameters of the narrowing region are computed, including at least stenosis length, stenosis thickness, and minimum lumen cross-sectional area. The dimensional parameters are output in a machine-readable format. Optionally, computational flow parameters and classification outputs may be generated.