Method and system for predicting response to immune anticancer drugs
Inventors
Yoo, Donggeun • OCK, Chanyoung • Paeng, Kyunghyun
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
The present disclosure relates to a method, performed by at least one computing device, for predicting a response to an immune checkpoint inhibitor. The method includes receiving a first pathology slide image, detecting one or more target items in the first pathology slide image, determining at least one of an immune phenotype of at least some regions in the first pathology slide image or information associated with the immune phenotype based on the detection result for the one or more target items, and generating a prediction result as to whether or not a patient associated with the first pathology slide image responds to the immune checkpoint inhibitor, based on the immune phenotype of the at least some regions in the first pathology slide image or the information associated with the immune phenotype.
Core Innovation
The invention predicts response to an immune checkpoint inhibitor by receiving a first pathology slide image and detecting one or more target items comprising a cancer area, a stroma area, and immune cells in the first pathology slide image using an artificial neural network model. The artificial neural network model is trained to reduce errors between label information for reference target items corresponding to reference pathology slide images and detected output from the reference pathology slide images.
The invention identifies a plurality of patches of a predetermined size in the first pathology slide image as regions of interest, and determines immune phenotypes of the plurality of patches based on at least one of a density, a number, or a distribution of immune cells in the cancer area or the stroma area of each patch. The immune phenotypes are determined so that a specific class of immune phenotype is assigned to patches among classes including immune-inflamed, immune-excluded, and immune-desert.
The invention calculates a prediction score indicating whether a patient associated with the first pathology slide image responds to the immune checkpoint inhibitor based on a ratio or a number of patches having an immune phenotype of a specific class among the plurality of patches, and outputs the prediction score. The invention further describes computing immune phenotype information for patches as a basis for prediction, including outputting an immune phenotype map and determining a most common immune phenotype across a whole region.
Claims Coverage
The partial content includes two independent claims: a method for predicting patient response to an immune checkpoint inhibitor and an information processing system configured to perform the same. Each independent claim includes five main inventive features.
Artificial neural network target-item detection on pathology slide image
Detecting one or more target items comprising a cancer area, a stroma area, and immune cells in the first pathology slide image using an artificial neural network model trained to reduce errors between label information for reference target items and detected output.
Patch-based regions of interest on pathology slide
Identifying a plurality of patches of a predetermined size in the first pathology slide image as regions of interest.
Immune phenotype determination from immune-cell density/number/distribution
Determining immune phenotypes of the plurality of patches based on at least one of a density, a number, or a distribution of immune cells in the cancer area or the stroma area of each patch.
Prediction score from ratio or number of phenotype-class patches
Calculating a prediction score indicating whether or not a patient responds to an immune checkpoint inhibitor based on a ratio or a number of patches having an immune phenotype of a specific class among the plurality of patches.
Outputting the prediction score
Outputting the prediction score.
Both independent claims cover artificial-neural-network-based detection of cancer area, stroma area, and immune cells, followed by patch-based immune phenotype determination and prediction-score calculation from the ratio or number of patches in a specified immune-phenotype class, with output of the prediction score.
Stated Advantages
Objective quantification of immune phenotype information derived from pathology slide images.
Improved predictive accuracy for response prediction to immune checkpoint inhibitors.
Faster/automatic analysis.
Documented Applications
Predicting patient response to an immune checkpoint inhibitor using digital pathology slides including a first pathology slide image, with optional incorporation of biomarker expression from a second pathology slide image.
Producing immune phenotype information outputs such as a representative phenotype and an immune phenotype map/feature map for use in response prediction.
Supporting immune checkpoint inhibitor recommendation among multiple checkpoint inhibitors based on the generated immune phenotype and prediction output.
Interested in licensing this patent?