US 12,205,684 B2
Systems and methods for patient-trial matching
Alexander Padmos, New York, NY (US); Angel Leung, Brooklyn, NY (US); Caroline Nightingale, Brooklyn, NY (US); Zexi Chen, Forest Hills, NY (US); Janet Donegan, Park City, UT (US); Peter Larson, New York, NY (US); and Lauren Sutton, New York, NY (US)
Assigned to Flatiron Health, Inc., New York, NY (US)
Filed by Flatiron Health, Inc., New York, NY (US)
Filed on May 22, 2020, as Appl. No. 16/881,317.
Claims priority of provisional application 62/851,870, filed on May 23, 2019.
Prior Publication US 2020/0372979 A1, Nov. 26, 2020
Int. Cl. G16H 10/20 (2018.01); G16H 10/60 (2018.01); G16H 50/20 (2018.01)
CPC G16H 10/20 (2018.01) [G16H 10/60 (2018.01); G16H 50/20 (2018.01)] 16 Claims
OG exemplary drawing
 
1. A system for determining trials using a metastatic condition of a patient, the system comprising:
at least one processor programmed to:
receive, via a first user interface, a selection of trial eligibility criteria associated with a plurality of trials;
generate, based on the selected trial eligibility criteria, code representing a plurality of expression tree algorithms associated with the plurality of trials;
receive a selection of the patient;
access, in response to the selection of the patient, a patient dataset associated with the patient;
receive at least one image of a medical record associated with the patient, wherein the at least one image is captured by a scanning device associated with the at least one processor and uploaded from the scanning device to the at least one processor;
extract, by a trained recurrent neural network model, unstructured metastatic condition data associated with the patient from the at least one image;
convert, by the trained recurrent neural network model, the unstructured metastatic condition data to structured metastatic condition data associated with the patient;
generate, based on the structured metastatic condition data, a predicted metastatic condition for the patient;
wherein determining the predicted metastatic condition includes application of a trained model configured to receive unstructured information and output the predicted metastatic condition based on the unstructured information;
cause display of at least a first portion of the patient dataset and the predicted metastatic condition via a second user interface, the second user interface including at least one element for verifying the predicted metastatic condition;
identify, based on at least a second portion of the patient dataset, the predicted metastatic condition, and the code representing the plurality of expression tree algorithms, an initial subset of trials for which the patient is eligible within the plurality of trials;
cause display of the initial subset of trials for the patient via the second user interface;
after causing display of the initial subset of trials for the patient via the second user interface:
receive, via the second user interface, an input indicating that a user has verified the predicted metastatic condition of the patient thereby providing a verified metastatic condition of the patient, the input received through an interaction of the user with the at least one element for verifying the predicted metastatic condition; and
in response to receiving the input indicating that the user verified the predicted metastatic condition of the patient, cause display of an updated subset of trials for which the patient is eligible within the plurality of trials based on the verified metastatic condition.