Radiotherapy plan parameters with privacy guarantees
Inventors
Fay, Dominik • Sjolund, Jens Olof
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11975218-B2
Publication Date
2024-05-07
Expiration Date
Abstract
Techniques for producing segmentation with privacy are provided. The techniques include receiving a medical image; processing the medical image with a student machine learning model to estimate radiotherapy plan parameters, the student machine learning model being trained to establish a relationship between a plurality of public training medical images and corresponding radiotherapy plan parameters, the radiotherapy plan parameters of the plurality of public training medical images being generated by aggregating a plurality of radiotherapy plan parameter estimates produced by: processing the plurality of public training medical images with a plurality of teacher machine learning models to generate sets of radiotherapy plan parameter estimates; and reducing respective dimensions of the sets of radiotherapy plan parameter estimates or medical images, the radiotherapy plan parameters of the plurality of public training medical images being perturbed in accordance with privacy criteria; and generating a radiotherapy treatment plan based on the estimated radiotherapy plan parameters.
Core Innovation
A method for privacy based radiotherapy treatment planning receives a medical image of a patient and processes the medical image with a student machine learning model to estimate one or more radiotherapy plan parameters. The student machine learning model is trained to establish a relationship between a plurality of public training medical images and corresponding radiotherapy plan parameters of the public training medical images.
The radiotherapy plan parameters are generated by aggregating radiotherapy plan parameter estimates produced by processing the public training medical images with a plurality of teacher machine learning models. The sets of radiotherapy plan parameter estimates are processed to reducing respective dimensions of the sets of radiotherapy plan parameter estimates or the plurality of public training medical images, and the radiotherapy plan parameters of the plurality of public training medical images are perturbed in accordance with privacy criteria.
The perturbed, dimension-reduced results are then used to train and generate radiotherapy plan parameters for the patient. A radiotherapy treatment plan is generated for the patient based on the estimated one or more radiotherapy plan parameters of the medical image of the patient.
Claims Coverage
The document provides three independent claims, a method, a non-transitory computer-readable medium, and a system, all centered on the same inventive pipeline. The claims include the same core inventive features: student estimation of radiotherapy plan parameters for a patient, training on public training medical images, aggregation of teacher outputs, and dimension reduction with perturbation under privacy criteria.
Privacy based radiotherapy treatment planning with student estimation
Receiving a medical image of a patient; processing the medical image with a student machine learning model to estimate one or more radiotherapy plan parameters; and generating a radiotherapy treatment plan for the patient based on the estimated one or more radiotherapy plan parameters.
Public training images with student relationship training
Training the student machine learning model to establish a relationship between a plurality of public training medical images and corresponding radiotherapy plan parameters of the public training medical images.
Aggregated teacher outputs to generate public radiotherapy plan parameters
Generating the radiotherapy plan parameters of the plurality of public training medical images by aggregating a plurality of radiotherapy plan parameter estimates produced by processing the plurality of public training medical images with a plurality of teacher machine learning models.
Dimension reduction with privacy-criteria perturbation of radiotherapy plan parameters
Reducing respective dimensions of the sets of radiotherapy plan parameter estimates or the plurality of public training medical images, wherein the radiotherapy plan parameters of the plurality of public training medical images are perturbed in accordance with privacy criteria.
Computer-readable medium for privacy based radiotherapy planning
A non-transitory computer-readable medium comprising non-transitory computer-readable instructions for receiving a medical image of a patient, processing the medical image with a student machine learning model to estimate one or more radiotherapy plan parameters, and generating a radiotherapy treatment plan for the patient based on the estimated one or more radiotherapy plan parameters.
Computing system implementing privacy based radiotherapy planning pipeline
A system comprising memory for storing instructions and one or more processors executing the instructions for receiving a medical image of a patient, processing the medical image with a student machine learning model to estimate one or more radiotherapy plan parameters, and generating a radiotherapy treatment plan for the patient based on the estimated one or more radiotherapy plan parameters.
Across the independent claims, the core claim coverage is directed to a privacy based radiotherapy treatment planning pipeline in which a student machine learning model estimates radiotherapy plan parameters for a patient and is trained using public training medical images whose radiotherapy plan parameters are generated by aggregating teacher-produced estimates after reducing dimensions and perturbing the radiotherapy plan parameters according to privacy criteria.
Stated Advantages
Improves privacy/data availability and ML accuracy/reliability by enabling efficient low-noise privacy guarantees.
Reduces computation relative to privacy methods that are computationally prohibitive.
Documented Applications
Privacy based radiotherapy treatment planning using a student machine learning model trained from public training medical images to generate a radiotherapy treatment plan from a patient medical image.
Generating radiotherapy treatment plan parameters such as segmentation labels, dose distribution, synthetic CT, 3D volume, and radiotherapy device parameters for privacy based radiotherapy treatment planning.
Interested in licensing this patent?