Method and system for assessing drug efficacy using multiple graph kernel fusion
Inventors
Frieder, Ophir • YAO, Hao-Ren • CHANG, Der-Chen
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12482568-B2
Publication Date
2025-11-25
Expiration Date
Abstract
Embodiments of the present systems and methods may provide techniques to predict the success or failure of a drug used for disease treatment. For example, a method of determining drug efficacy may include, for a plurality of patients, generating a directed acyclic graph from health related information of each patient comprising nodes representing a medical event of the patient, at least one first edge connecting the first node to an additional node, each additional edge connecting nodes representing two consecutive medical events, the edge having a weight based on a time difference between the two consecutive medical events, capturing a plurality of features from each directed acyclic graph, generating a binary graph classification model on captured features of each directed acyclic graph, determining a probability that a drug or treatment will be effective using the binary graph classification model, and determining a drug to be prescribed to a patient based on the determined probability.
Core Innovation
The problem is to compute a probable drug efficacy for a particular patient so that a drug or treatment can be determined to be prescribed based on an effectiveness probability. To address this, for each of a plurality of patients, the method generates and stores in memory a graph representing medical events, where nodes represent medical events and edges connect nodes representing two consecutive medical events. The patient-specific probability is produced by training and applying a classifier model using graph-derived features.
The captured features include generating a topological ordering of each graph based on an order of occurrence of a label associated with one or more nodes, and generating a topological sequence comprising a plurality of levels indicating an order of occurrence of the same node label in the topological sequence. The method also generates a temporal signature for each graph, generates a temporal proximity kernel between pairs of temporal signatures, and generates a shortest path kernel by calculating an edge walk similarity on at least one shortest path graph for pairs of graphs.
In parallel, it generates a node kernel by comparing node labels of pairs of graphs. The method fuses the temporal proximity kernel, the shortest path kernel, and the node kernel, then trains a classifier model using the plurality of captured features of each graph. The trained classifier model is applied to determine a probability that a drug or treatment will be effective for a particular patient, and a drug or treatment to be prescribed is determined based on the determined probability. This computer-implemented framework is implemented as a system and also as a computer program product.
Claims Coverage
The independent claims are clm-00001, clm-00010, and clm-00019. Across these claims, the same core inventive feature set appears: generating patient event graphs, capturing topological and temporal features via temporal proximity, shortest path similarity, and node-label comparisons, fusing kernels, training a classifier model, and determining a prescribing decision based on an effectiveness probability.
Patient graph generation of medical events and consecutive events
Generating and storing in memory a graph for each of a plurality of patients, the graph comprising a plurality of nodes, each node representing a medical event of the patient, and a plurality of edges connecting nodes representing two consecutive medical events.
Topological-temporal feature extraction from patient graphs
Capturing features from each graph by generating a topological ordering based on an order of occurrence of a label associated with one or more nodes in the graph and generating a topological sequence comprising a plurality of levels indicating an order of occurrence of a same node label in the topological sequence; generating a temporal signature for each graph.
Kernel construction from temporal proximity, shortest-path similarity, and node label comparison
Generating a temporal proximity kernel between pairs of temporal signatures; generating a shortest path kernel by calculating an edge walk similarity on at least one shortest path graph for pairs of graphs; and generating a node kernel by comparing node labels of pairs of graphs.
Fusing kernels and classifying for probable drug efficacy
Fusing the temporal proximity kernel, the shortest path kernel, and the node kernel; training a classifier model using captured features; applying the classifier model to determine a probability that a drug or treatment will be effective for a particular patient.
Prescription determination based on effectiveness probability
Determining, using the processor, a drug or treatment to be prescribed to the particular patient based on the determined probability.
The independent claims cover a computer-implemented framework that turns each patient’s labeled medical events into a graph, computes topological and temporal graph features, fuses the kernels, trains a classifier model, and outputs a probability of drug or treatment effectiveness that is used to determine a prescribing decision.
Stated Advantages
Documented Applications
Not explicitly described in partial content.
Interested in licensing this patent?