Direct medical treatment predictions using artificial intelligence
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12763179-B2
Publication Date
2026-10-13
Expiration Date
Abstract
A device is disclosed herein that receives image data corresponding to an anatomy of a patient. The device applies the image data to one or more feature models trained using training data that pairs anatomical images to an anatomical feature label, and receives, as output from the one or more feature models, scores for each of a plurality of anatomical features corresponding to the image data. The device applies the scores as input to a treatment model, the treatment model trained to output a prediction of a measure of efficacy of a particular treatment based on features of the patient's anatomy. The device receives, as output from the treatment model, data representative of the predicted measure of efficacy of the particular treatment.
Core Innovation
The invention describes AI systems that predict patient-specific treatment benefit directly from clinical outcomes, avoiding an intermediate diagnosis and subjective clinician interpretation. In the image-based architecture, a first machine learning model scores anatomical features reflected in patient image data, and a second machine learning model uses the scores to predict a measure of efficacy for a particular treatment.
In the sensor-based architecture, a machine learning model uses patient sensor data and labels describing treatment outcomes to predict whether a patient would benefit from a particular treatment. The disclosed multi-task model includes shared processing and treatment-specific branches. The stated problem is addressed by predicting treatment benefit from image or sensor data rather than relying on an intermediate diagnosis or subjective clinician interpretation.
Claims Coverage
The patent has four independent claims: claims 1 and 6 cover image-based methods and computer-readable-medium instructions, while claims 11 and 16 cover sensor-based methods and computer-readable-medium instructions. Together, they cover two main inventive features.
Image-based prediction of treatment efficacy
Claims 1 and 6 receive image data of a patient’s anatomy, obtain scores for anatomical features using a first machine learning model, and provide those scores to a second machine learning treatment model. The treatment model outputs data representing a predicted measure of efficacy for a particular treatment.
Sensor-based prediction of treatment benefit
Claims 11 and 16 receive sensor data from an electronic device that monitors a patient and apply it to a machine learning model trained using previously obtained patient sensor data paired with labels describing treatment outcomes. The model outputs data representing the likelihood that the patient would benefit from a particular treatment.
The independent claims cover image-based prediction of treatment efficacy and sensor-based prediction of treatment benefit, each claimed as both a method and computer-readable-medium instructions.
Stated Advantages
Outcome-based training may reduce validation cost and risk.
Shared layers can improve predictions for treatments with limited training data.
Documented Applications
Predicting a measure of treatment efficacy from patient anatomical image data.
Predicting whether a patient would benefit from a particular treatment using sensor data from an electronic device that monitors the patient.
Interested in licensing this patent?