Vascular characteristic determination based on multiple images identifying a common vascular segment with correspondence modeling of a vascular tree

Inventors

Lavi, Guy • Merhav, Uri • Lavi, Ifat

Assignees

CathWorks Ltd

Abstract

Automated image analysis used in vascular state modeling. Coronary vasculature in particular is modeled in some embodiments. Methods of “virtual revascularization” of a presently stenotic vasculature are described; useful, for example, as a reference in disease state determinations. Structure and uses of a model which relates records comprising acquired images or other structured data to a vascular tree representation are described.

Core Innovation

The invention relates to vascular modeling that relates 2-D projection images of a vasculature to a vascular tree representation. It identifies vascular segments represented in common by different regions of the 2-D projection images by determining a group of back-projecting rays from identified vascular segments and by identifying respective regions in at least a subset of the 2-D projection images that commonly-represent a same identified vascular segment.

The common representation is based on analyzing back-projecting rays indicated as intersecting each other. Ray intersection is determined based on at least a portion of a first back-projecting ray being at a distance of at least twice a sampling interval of the 2-D projection images from a second back-projecting ray. The framework further supports coarse-grained homology and correspondence relationships that relate 2-D angiography-derived vascular region/segment data to a vascular tree representation without requiring a full-resolution 3-D reconstruction.

The model supports metric computation and functional estimation based on the vascular tree representation, including vascular width and resistance and functional estimation including FFR. It further supports virtual revascularization and procedure planning with virtual stenosis opening, virtual stent modeling, virtual shunt effects, and dynamically updatable or partially updatable vascular tree models using new images.

For dynamic updates, the approach relates additional data samples to the vascular tree representation by generating a first record of an image region of samples projected from the vasculature and generating a second record comprising differences compared to the first record. It updates automatically the first record to reflect the differences, where the differences are based on 3-D projected epipolar mapping between samples of the first record and the second record.

Claims Coverage

The document provides three independent claims covering (i) identifying commonly represented vascular segments across regions of multiple 2-D projection images, (ii) a corresponding vascular assessment apparatus, and (iii) automatically relating additional data samples to a vascular tree representation by updating records using 3-D projected epipolar mapping. Across these independent claims, the inventive features primarily include back-projection ray intersection logic with a sampling-interval constraint, apparatus implementation of the same identification workflow, and record update via 3-D projected epipolar mapping mapped to a vascular tree representation.

Across the independent claims, the coverage centers on determining groups of back-projecting rays from identified vascular segments and identifying commonly-representing image regions via intersecting-ray analysis under a constraint tied to twice a sampling interval, implementing that identification in a vascular assessment apparatus with a processor and memory, and updating a vascular tree-associated representation by generating records for projected sample regions and applying automatic updates based on 3-D projected epipolar mapping.

Stated Advantages

Documented Applications