Prostate Cancer Recurrence Prediction Using Bivariate Prognostic Features

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

There is an unsolved problem in accurately identifying cancer patients who will experience cancer recurrence, particularly prostate cancer patients who will exhibit prostate-specific antigen (PSA) recurrence after radical prostatectomy.

Innovation Solution

A method combining immunohistochemical staining and gene expression profiling to generate a network of bivariate prognostic features using a minimal spanning tree, where structural and gene expression features are integrated to predict cancer recurrence.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If immunohistochemical staining and gene expression profiling are combined to create bivariate prognostic features, then prediction accuracy of cancer recurrence is improved, but device complexity and analysis complexity increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines immunohistochemical staining data and gene expression profiling data into integrated bivariate prognostic features. This merging of multiple data sources improves prediction accuracy by capturing both spatial tissue architecture information and molecular expression patterns simultaneously, resolving the contradiction between improved measurement precision and increased system complexity through unified multi-omics analysis.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent transitions from univariate to bivariate feature analysis, adding a new dimension to the prognostic model. By creating bivariate features that combine structural and gene expression data, the system achieves higher prediction accuracy while the minimal spanning tree methodology provides dimensionality reduction to manage the increased complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If multiple features (structural and gene expression) are integrated into a prognostic model, then prediction accuracy is improved, but ease of operation decreases

Engineering Contradiction:
Improveprediction accuracyVSAvoidease of use
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent extracts and integrates only the most relevant structural and gene expression features into the prognostic model using minimal spanning tree analysis. This selective extraction approach improves prediction accuracy by focusing on discriminatory features while simplifying the overall model compared to using all available features, thereby maintaining ease of operation.

Inventive Principle:
Principle #2Taking out (Extraction)

3Measurement precision

If a minimal spanning tree network is constructed to integrate multiple features, then prediction accuracy is improved, but device complexity increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidnetwork complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the complex multi-omics data into distinct structural features and gene expression features, then integrates them through a minimal spanning tree network. This segmentation approach improves prediction accuracy by preserving the unique information in each data type while the tree structure provides an organized, manageable representation that controls network complexity.

Inventive Principle:
Principle #1Segmentation

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach allows for improved prediction of cancer recurrence, enabling accurate stratification of patients into treatment groups and predicting PSA recurrence in prostate cancer patients.

Implementation Method 1

immunohistochemical staining and gene expression profiling to generate a network of bivariate prognostic features

Methodology Applied
Scientific EffectImmunohistochemistry:

Implementation Method 2

the gene expression level of the gene is a measurement of expression of mRNA transcribed from the gene in the tissue sample by using an mRNA-specific probe biomarker

Methodology Applied
Scientific EffectHybridization:

Data Source

PatentEP3533883B1Predicting cancer recurrence using a prognostic model that combines immunohistochemical staining and gene expression profiling
Publication Date: 2025.12.10 ASTRAZENECA COMPUTATIONAL PATHOLOGY GMBH
  • EP3533883B1 patent drawingFigure 1A
  • EP3533883B1 patent drawingFigure 1B
  • EP3533883B1 patent drawingFigure 1C

AI summary

Predicting cancer recurrence using a prognostic model that combines immunohistochemical staining and gene expression profiling Disclosed is an in vitro method for determining a risk of a cancer patient of a cancer recurrence. Further disclosed is a method of generating a network of prognostic features for cancer recurrence of a cancer patient. More particularly, disclosed is the prediction of prostate cancer recurrence using a prognostic model that combines immunohistochemical staining and gene expression profiling.