Machine-learning based contrast agent administration
AssigneeSiemens Healthcare GmbH
InventorsPooyan Sahbaee Bagherzadeh, Saikiran Rapaka
Life of this patent
Estimating the term…
Abstract
A method comprises: inputting contrast enhancement data for at least one organ of a patient, at least one patient attribute of the patient, and a test bolus data or bolus tracking data to a regressor; receiving a contrast agent administration protocol from the regressor, based on the contrast enhancement data, the at least one patient attribute and the test bolus or bolus tracking data; and injecting a contrast agent into the patient according to the received contrast agent administration protocol.
Claims25
A method comprising:
- inputting contrast enhancement data for at least one organ of a patient, at least one patient attribute of the patient, and a test bolus data or bolus tracking data to a regressor;
- receiving a contrast agent administration protocol from the regressor, based on the contrast enhancement data, the at least one patient attribute and the test bolus or bolus tracking data;
- and
- injecting a contrast agent into the patient according to the received contrast agent administration protocol.
A system comprising:
- a simulator for modeling flow of a contrast agent through a body and generating contrast enhancement data;
- a non-transitory, machine readable storage medium storing the contrast enhancement data, at least one patient attribute of a patient, and a test bolus or bolus tracking data of the patient;
- and
- at least one processor including a regressor configured to generate a contrast agent administration protocol, based on the contrast enhancement data and the at least one patient attribute.
A non-transitory, machine-readable storage medium encoded with instructions for controlling a processor, such that when the processor executes the instructions, the processor performs a method comprising:
- simulating flow of a contrast agent through a body and generating contrast enhancement data from the simulating;
- inputting at least one patient attribute of a patient to a regressor trained using the contrast enhancement data;
- and
- using the regressor to generate a scan delay between an injection of a contrast agent and a scan, such that performance of the scan after the injection and the scan delay captures an image of an organ with a peak contrast agent concentration in the organ.
Dependencies are inferred from the claim text.