Systems and methods for training a model to predict survival time for a patient
Inventors
Beck, Andrew H. • Khosla, Aditya
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12073948-B1
Publication Date
2024-08-27
Expiration Date
Abstract
In some aspects, the described systems and methods provide for a method for training a model to predict survival time for a patient. The method includes accessing annotated pathology images associated with a first group of patients in a clinical trial. Each of the annotated pathology images is associated with survival data for a respective patient. Each of the annotated pathology images includes an annotation describing a tissue characteristic category for a portion of the image. Values for one or more features are extracted from each of the annotated pathology images. A model is trained based on the survival data and the extracted values for the features. The trained model is stored on a storage device.
Core Innovation
The invention determines whether a trained model accurately identifies a patient subset that is likely to benefit from an experimental treatment compared to a standard treatment or a control treatment. The determining is performed by processing, using the trained model, values for a plurality of features extracted from annotated pathology images associated with a first group of patients and a second group of patients in a randomized controlled clinical trial.
The first group belongs to an experimental treatment group and the second group belongs to a control treatment group. The determination is based on predicted survival data for the patients in the first group and the predicted survival data for the patients in the second group, and in related implementations includes subset survival analyses and identifies a patient subset and/or responders using predicted survival data.
The plurality of features includes pathology-derived quantitative measurements such as tissue and cell areas, mitotic figures, nuclear grade and variability, and distances between cell or tissue types. The trained model can comprise one or more machine learning model types including a generalized linear model, a random forest, a support vector machine, and a gradient boosted tree.
Claims Coverage
The document includes three independent claim types: a method, a system, and a non-transitory computer-readable storage medium. Across these independent claims, there are three main inventive features: predicting survival data for experimental and control groups using a trained model on features extracted from annotated pathology images, comparing the predicted survival data to determine whether the model accurately identifies a patient subset likely to benefit, and implementing the determining as a method, system, or computer-readable medium.
Predicted survival from annotated pathology features in an experimental trial group
Processing, using the trained model, values for features extracted from annotated pathology images associated with a first group of patients to predict survival data for patients in the first group of patients, wherein the first group of patients belongs to an experimental treatment group of a randomized controlled clinical trial.
Predicted survival from annotated pathology features in a control trial group
Processing, using the trained model, values for features extracted from annotated pathology images associated with a second group of patients to predict survival data for patients in the second group of patients, wherein the second group of patients belongs to a control treatment group of the randomized controlled clinical trial.
Accuracy-based identification of a likely-benefit patient subset
Determining whether the trained model accurately identifies a patient subset that is likely to benefit from the experimental treatment compared to the standard treatment or the control treatment based on the predicted survival data for the patients in the first group of patients and the predicted survival data for the patients in the second group of patients.
System implementation for experimental/control survival determination using a trained model
A system, comprising at least one computer hardware processor and at least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed, cause the processor to perform determining whether a trained model accurately identifies a patient subset that is likely to benefit from an experimental treatment compared to a standard treatment or a control treatment.
Non-transitory computer-readable medium implementing the survival determination workflow
A non-transitory computer-readable storage medium storing processor-executable instructions that, when executed, cause a processor to perform determining whether a trained model accurately identifies a patient subset that is likely to benefit from an experimental treatment compared to a standard treatment or a control treatment.
Across the independent claims, the coverage centers on using a trained model to process pathology-derived feature values from annotated pathology images to predict survival data for patients in an experimental treatment group and in a control treatment group within a randomized controlled clinical trial, and then determining whether the model accurately identifies a patient subset likely to benefit from the experimental treatment based on the predicted survival data.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?