Automated FOV Selection for Multiplex Tissue Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The traditional manual selection of fields of view (FOVs) in whole-slide imaging for immunoscore computations is subjective and biased, leading to reproducibility issues in diagnosing diseases like colorectal cancer.
Innovation Solution
An automated image processing method that uses spatial low pass filtering, local maximum filtering, and thresholding to identify FOVs in multi-channel images, allowing for the merging of FOVs from different marker images and reducing human error, while maintaining analysis at full resolution for accurate cell counting and tissue region segmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of fields of view is performed by pathologists, then diagnostic accuracy can be maintained through expert judgment, but reproducibility deteriorates due to subjective bias and inter-reader variability
Solution Approach 1:
The system enables automated FOV selection that performs the diagnostic task independently of human readers. The algorithm automatically identifies and selects FOVs based on quantitative image analysis criteria, eliminating the need for manual pathologist review and selection, thereby achieving both high reproducibility and maintained diagnostic accuracy
Solution Approach 2:
The patent replaces the mechanical/manual process of pathologist review with an automated computational system. The system uses image processing algorithms, heat map generation, and automated FOV selection to substitute the human expert review process, eliminating subjective bias while maintaining diagnostic quality through objective, reproducible criteria
2Measurement precision
If full-resolution images are analyzed for accurate cell counting, then measurement precision is improved, but computational time and processing speed deteriorate
Solution Approach 1:
The system segments the analysis process into two stages: first, automated FOV selection identifies candidate regions using processed images; second, full-resolution analysis is performed only on the selected FOVs. This segmentation allows the system to maintain measurement precision in the final cell counting stage while improving overall computational efficiency by limiting full-resolution processing to smaller, pre-selected regions
Solution Approach 2:
The system performs preliminary FOV selection and filtering operations before the final full-resolution analysis. By pre-processing images to identify and select candidate FOVs using automated algorithms and heat maps, the system prepares the data in advance, allowing subsequent full-resolution cell counting to be performed more efficiently on a reduced dataset
Data Source
AI summary
Systems and methods for automatic FOV selection in immunoscore computation that involve reading images for individual markers from an unmixed multiplex slide or single stain slides, and computing the tissue region mask from the individual marker image. The heat map of each marker is determined by applying the low pass filter on the individual marker image channel and selecting the top K highest intensity regions from the heat map as the candidate FOVs for each marker. The candidate FOVs from the individual marker images are merged together in the same coordinate system by either adding all of the FOVs together or by only adding the FOVs from the selected marker images depending on the user's choice, and registering all the individual marker images to a common coordinate system and transferring the FOVs back to the original images.