Computer-implemented method for evaluating a three-dimensional angiography dataset, evaluation system, computer program and electronically readable storage medium
Inventors
DENZINGER, Felix • Wels, Michael • Suehling, Michael
Assignees
Publication Number
US-12354269-B2
Publication Date
2025-07-08
Expiration Date
2042-09-20
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
A computer-implemented method for evaluating a three-dimensional angiography dataset of a blood vessel tree of a patient, comprises determining a variant information describing a belonging to at least one anatomical variant class of a plurality of anatomical variant classes relating to anatomical variants of the blood vessel tree based on a comparison of angiography information of the angiography dataset to reference information describing at least one of the anatomical variant classes.
Core Innovation
The invention provides a computer-implemented method for evaluating a three-dimensional angiography dataset of a blood vessel tree of a patient, which determines a variant information describing a patient's belonging to at least one anatomical variant class from a plurality of anatomical variant classes relating to anatomical variants of the blood vessel tree. This determination is based on a comparison of angiography information extracted from the angiography dataset to reference information describing the anatomical variant classes.
The problem solved by the invention arises from the existence of non-pathological, normal anatomical variants in blood vessel trees, such as the coronary artery tree, which complicate diagnosis and treatment planning. These variants may influence diagnosis, patient risk stratification, management decisions, and complicate minimally invasive interventions, yet are currently assessed qualitatively by human readers, limiting robustness and automation.
The invention addresses this problem by providing a robust and automated approach that compares structural features of patient angiography data, such as centerlines or lumen segmentations, against reference information including rule sets or anatomical atlas datasets. This comparison allows for automatic detection and classification of anatomical variants, facilitating clinically relevant reporting and improving the applicability and reliability of downstream evaluation algorithms.
Claims Coverage
The patent contains two independent claims covering a computer-implemented method for evaluation of angiography datasets and an evaluation system configured for this purpose. The main inventive features relate to determining anatomical variant information by comparing angiography data to reference information with various structural and procedural details.
Determining anatomical variant information based on dataset comparison
Determining variant information describing a belonging to at least one anatomical variant class from plurality of classes by comparing angiography information of the dataset to reference information describing these anatomical variant classes.
Two-step comparison approach for detecting anatomical variants
Performing an initial comparison to indicate presence of anatomical variants and, if detected, performing further comparisons to determine belonging to specific anatomical variant classes; if no variants are detected, determining belonging to a no anatomical variant class.
Use of structural information as angiography information
Determining at least one structural information of the blood vessel tree, such as centerlines or lumen, by structural evaluation algorithms from the angiography dataset to serve as angiography information for comparison.
Centerline tree and rule set based comparison
Determining centerlines to form a centerline tree and providing rule sets for anatomical variant classes comprising conditions on features of the centerline tree; determining membership to a variant class when all conditions of a rule set are fulfilled.
Evaluation of additional anatomical features in conditions
Including conditions in rule sets that evaluate anatomy information of additional features not part of the blood vessel tree, derived from the angiography dataset or registered additional image data.
Atlas-based comparison for anatomical variant detection
Providing anatomical atlas datasets relating to variant classes as reference information and comparing these atlases to the angiography dataset or derived comparison datasets.
Definition and use of no variant anatomical atlas dataset
Using a no variant anatomical atlas dataset derived statistically from multiple base datasets of patients without anatomical variants, serving as a reference for normal anatomy.
Localized subarea comparison with atlas in further comparison steps
If variants are detected, performing further comparisons by restricting comparison with no variant atlas datasets to subareas associated with particular anatomical variant classes to determine specific variant class membership.
Use of similarity or correlation metrics and lumen derivation
Calculating similarity or correlation metrics in comparisons for variant detection and optionally deriving lumens from angiography datasets as comparison datasets.
Utilization of variant information in downstream processes
Using the variant information for automatic report generation or to determine applicability of downstream evaluation algorithms.
Inclusion of anatomical variant classes corresponding to imaging artifacts
Using at least one anatomical variant class relating to image features caused by imaging artifacts such as stacking artifacts.
Diverse types of anatomical variant classes covered
Anatomical variant classes may relate to dominance variants, region supply variants, additional or fewer blood vessel variants, course variants, or ostium anomaly variants.
Evaluation system configured for variant determination
An evaluation system comprising an interface to receive angiography datasets, a determination unit for variant information based on angiography information comparison to reference information, and an output interface for variant information provision.
The claims cover a method and system for automated detection and classification of anatomical variants in angiography datasets by structurally comparing patient data to reference information including rule sets and anatomical atlases, incorporating multi-step comparisons, structural and anatomical feature analysis, and enabling downstream clinical and algorithmic use.
Stated Advantages
Provides a robust and fully automated detection of anatomical variants in blood vessel trees.
Improves clinical workflow by enabling automatic report generation including anatomical variant information relevant for diagnosis and treatment planning.
Allows variant information to act as a gatekeeper to determine applicability and reliability of subsequent evaluation algorithms, especially deep learning-based approaches.
Enables detection and distinction of anatomical variants from imaging artifacts, improving data quality assessment.
Applicable to various blood vessel trees and imaging modalities, with preferred application to coronary and cerebral angiography.
Documented Applications
Automatic assessment and classification of anatomical variants in three-dimensional angiography datasets of blood vessel trees, particularly the coronary artery tree.
Support for clinical diagnosis and treatment planning by providing variant information relevant for risk stratification and management decisions.
Facilitation of minimally invasive interventions through identification of anatomical variants.
Automatic report generation incorporating information about anatomical variants.
Determination of applicability and reliability for downstream evaluation algorithms, including deep learning-based artificial intelligence algorithms.
Detection of imaging artifacts that resemble anatomical variants, aiding image quality assurance.
Interested in licensing this patent?