	US 11,869,664 B2
	Method and system for assessing drug efficacy using multiple graph kernel fusion
Ophir Frieder, Chevy Chase, MD (US); Hao-Ren Yao, McLean, VA (US); and Der-Chen Chang, Burtonsville, MD (US)
Assigned to Georgetown University, Washington, DC (US)
Filed by Georgetown University, Washington, DC (US)
Filed on Jun. 29, 2022, as Appl. No. 17/852,521.
Application 17/852,521 is a continuation of application No. 17/555,675, filed on Dec. 20, 2021, granted, now 11,410,763.
Application 17/555,675 is a continuation of application No. 17/088,172, filed on Nov. 3, 2020, granted, now 11,238,966, issued on Feb. 1, 2022.
Claims priority of provisional application 63/042,676, filed on Jun. 23, 2020.
Claims priority of provisional application 62/930,072, filed on Nov. 4, 2019.
Prior Publication US 2022/0344022 A1, Oct. 27, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 7/01 (2023.01); G16H 50/20 (2018.01); G16H 20/10 (2018.01); G16H 70/40 (2018.01); G06F 9/54 (2006.01); G16H 50/80 (2018.01); G06F 18/24 (2023.01)

CPC G16H 50/20 (2018.01) [G06F 9/545 (2013.01); G06F 18/24 (2023.01); G06N 7/01 (2023.01); G16H 20/10 (2018.01); G16H 50/80 (2018.01); G16H 70/40 (2018.01)]

27 Claims

1. A method of computing a probable drug efficacy implemented in a computer system comprising a processor, memory accessible by the processor and storing computer program instructions and data, and computer program instructions to perform:

for each of a plurality of patients, generating and storing in the memory a graph, each graph comprising:

a plurality of nodes, each node representing a medical event of the patient,

a plurality of edges connecting nodes representing two consecutive medical events, each edge of the plurality of edges having a weight based on a time difference between the two consecutive medical events;

capturing, using the processor, a plurality of features from each graph stored in the memory by:

transforming each graph to a shortest path graph;

generating a temporal topological kernel by recursively calculating similarity among temporal ordering on a plurality of groups of nodes; and

generating a temporal substructure kernel on edges connecting nodes in each group of nodes;

training, using a multiple graph kernel fusion (MGKF) framework, a classifier model using a plurality of the captured features of each graph;

applying the classifier model to determine a probability that a drug or treatment will be effective for a particular patient; and

determining, using the processor, a drug or treatment to be prescribed to the particular patient based on the determined probability.