Single-Cell HER2 ADC Scoring for Objective Therapy Response Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for assessing a cancer patient's response to antibody-drug conjugate therapy are prone to variability and subjectivity due to visual assessment by pathologists, necessitating a more repeatable and objective scoring method.
Innovation Solution
A method involving immunohistochemical staining, digital image acquisition, and convolutional neural network analysis to compute single-cell ADC scores, followed by statistical aggregation to generate a response score for predicting patient response to anti-HER2 antibody-drug conjugate therapy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual assessment by pathologists is used to score cancer patient response, then the method is simple and quick, but the scoring is prone to variability and subjectivity
Solution Approach 1:
The patent replaces the manual visual assessment mechanism with an automated image analysis system using convolutional neural networks. The system processes digital images of tissue samples to generate ADC scores, eliminating human subjectivity while maintaining operational simplicity through automated workflows.
Solution Approach 2:
The patent introduces digital image processing and AI algorithms as intermediaries between the tissue sample and the final scoring result. The convolutional neural network acts as a mediator that objectively translates visual staining patterns into quantitative ADC scores, reducing variability while preserving the essential assessment function.
2Reliability
If single-cell ADC scoring with convolutional neural network is implemented, then scoring objectivity and repeatability are improved, but computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary actions by pre-training convolutional neural network models on labeled datasets before actual scoring. This pre-training phase enables the system to rapidly process new images with high reliability, reducing processing time for clinical samples while maintaining consistent and repeatable scoring results.
Solution Approach 2:
The patent uses digital copies of tissue samples in the form of images for analysis, allowing multiple processing attempts and validations without consuming additional physical samples. The convolutional neural network analyzes these digital representations to generate reliable ADC scores efficiently.
3Measurement precision
If aggregation of all single-cell ADC scores is performed using statistical operations, then a comprehensive response score is generated, but computational requirements increase
Solution Approach 1:
The patent transforms the complex multi-cell scoring problem into a simplified statistical aggregation of individual cell scores. By changing the approach from analyzing complex spatial relationships to aggregating standardized single-cell parameters, the system achieves comprehensive response scoring with reduced computational requirements.
Data Source
Figure 1
Figure 2~4
Figure 5~9
AI summary
A method for predicting how a cancer patient will respond to an antibody drug conjugate (ADC) therapy involves computing a predictive response score based on single-cell ADC scores for each cancer cell. The ADC includes an ADC payload and an ADC antibody that targets a protein on each cancer cell, wherein the protein is human epidermal growth factor receptor 2 (HER2). A tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to the protein on cancer cells in the tissue sample. Cancer cells in a digital image of the tissue are detected. For each cancer cell, a single-cell ADC score is computed based on the staining intensities of the dye in the membrane and/or cytoplasm of the cancer cell and/or in the membranes and cytoplasms of neighboring cancer cells. The response of the cancer patient to the ADC therapy is predicted by aggregating all single-cell ADC scores of the tissue sample using a statistical operation.