	US 12,243,628 B2
	Data processing systems and methods for identifying new indications for drugs
Paul Bryce, West Newton, MA (US); Emanuele de Rinaldis, Cambridge, MA (US); Ramon Antonio Hernandez Vecino, Paris (FR); Francisco Javier Jimenez Jimenez, Brookline, MA (US); Cliona Marie Molony, Boston, MA (US); and Anne-Gaëlle Radenac, Cambridge, MA (US)
Assigned to Sanofi, Paris (FR)
Filed by Sanofi, Paris (FR)
Filed on Dec. 9, 2020, as Appl. No. 17/116,995.
Claims priority of provisional application 62/945,814, filed on Dec. 9, 2019.
Claims priority of application No. 20315299 (EP), filed on Jun. 9, 2020.
Prior Publication US 2021/0174912 A1, Jun. 10, 2021
Int. Cl. G16H 40/20 (2018.01); G06N 20/00 (2019.01); G16H 10/20 (2018.01); G16H 50/70 (2018.01)

CPC G16H 10/20 (2018.01) [G06N 20/00 (2019.01); G16H 50/70 (2018.01)]

16 Claims

1. A computer-implemented method for repurposing a drug by identifying one or more new indications for the drug, comprising:

receiving, by a computer system, data representing medical records of a plurality of patients;

selecting, based on the medical records, a set of patients from the plurality of patients by operations comprising:

determining at least one target signaling pathway associated with the drug; and

determining one or more indicators based on one or more factors corresponding to a diagnosis linked to the target signaling pathway;

determining a plurality of patient characteristics of the set of patients, each patient of the set of patients exhibiting at least one of the plurality of patient characteristics;

grouping, by the computer system, in accordance with the plurality of patient characteristics, the set of patients to generate a plurality of distinct groups, each of the distinct groups including multiple patients of the set of patients, wherein the grouping comprises:

executing a machine learning system configured to perform one or more unsupervised clustering techniques, wherein the one or more unsupervised clustering techniques comprise an iterative clustering operation, comprising:

applying the iterative clustering operation to feature vectors characterizing the patients in the set of patients to optimize a clustering function, wherein the iterative clustering operation comprises a bisecting k-means clustering technique; and

identifying the plurality of distinct groups based on a result of applying the iterative clustering operation to the feature vectors characterizing the patients in the set of patients; and

processing data defining the plurality of distinct groups to identify one or more new indications for the drug, comprising:

selecting, based on one or more group selection criteria, a set of distinct groups of the plurality of distinct groups for use in identifying new indications for the drug, comprising:

determining, for each of the plurality of distinct groups, a stability of the distinct group under perturbations of parameters of the iterative clustering operation;

determining, for each of the plurality of distinct groups, a purity of the distinct group based on a measure of variance between feature vectors of patients included in the distinct group; and

determining, for each of the plurality of distinct groups, whether to select the distinct group based at least in part on whether: (i) the stability of the distinct group under perturbations of parameters of the iterative clustering operation satisfies a first threshold, and (ii) the purity of the distinct group as determined based on the measure of variance between feature vectors of patients included in the distinct group satisfies a second threshold;

wherein fewer than all of the plurality of distinct groups are selected for use in identifying new indications for the drug;

identifying one or more relevant patient characteristics by processing data characterizing only the set of distinct groups that have been selected for use in identifying new indications for the drug; and

identifying at least one of the one or more relevant patient characteristics as a new indication for the drug; and

administering the druq to a patient as a treatment for the new indication, wherein the druq comprises an anti-interleukin-4 receptor alpha (anti-IL-4Rα) antibody, wherein the druq comprises Dupilumab.