Mass-analysis data processing using optical microscope images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current imaging mass spectrometers struggle to accurately distinguish different types of tissue or lesions in biological samples due to low spatial resolution in mass-analysis images, relying heavily on optical microscope images for classification, which limits the effective utilization of mass-analysis data.
Innovation Solution
A method and system for processing mass-analysis data that involves specifying small areas with similar compositions or properties based on optical microscope images, extracting common peak information, and comparing these areas to identify specific expression information, allowing for accurate classification and characterization of tissue types or lesions without relying solely on optical microscope images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If mass-analysis imaging is performed to obtain molecular distribution information, then chemical composition data is acquired, but spatial resolution is insufficient to distinguish different tissue types
Solution Approach 1:
The patent merges mass-analysis data with optical microscope image data to create a combined dataset that leverages the chemical information from mass spectrometry and the high spatial resolution morphological information from optical microscopy. This integration allows tissue classification to achieve both molecular specificity and spatial precision.
Solution Approach 2:
The patent introduces an intermediary classification process where optical microscope images serve as a bridge between the low-resolution mass-analysis data and the desired high-resolution tissue classification. The optical images provide intermediate spatial information that guides the interpretation of mass-spectrometry data at higher resolution.
2Reliability
If optical microscope images are used to evaluate mass-analysis results, then tissue classification reliability is improved, but the potential of mass-analysis data is not fully utilized
Solution Approach 1:
Instead of using optical microscope images merely to evaluate mass-analysis results, the patent inverts the approach by using optical images as primary data for classification and employing mass-analysis data to enhance and validate the classification. This reversal allows full utilization of both data types with optical images taking the lead role.
Solution Approach 2:
The patent makes the optical microscope image data serve multiple functions: as primary classification input, as spatial reference for mass-analysis data interpretation, and as validation tool. This multi-functionality ensures that optical image data is fully utilized rather than being merely an evaluative add-on.
3Productivity
If mass-analysis data is processed without integrating optical microscope images, then processing speed is improved, but tissue distinction accuracy deteriorates
Solution Approach 1:
The patent performs preliminary classification using optical microscope images before applying mass-analysis data processing. This preliminary action quickly identifies candidate tissue regions, which then guides focused processing of mass-analysis data only for those regions, maintaining speed while improving accuracy.
Data Source
AI summary
Provided is a technique for using an optical microscope image of an area on a sample to collect area-specific information characterizing each kind of biological tissue from imaging mass analysis data. On an optical image of a two-dimensional target area on a sample, a difference is examined in the kind of tissue or other features and areas are specified, each regarded as the same kind of tissue. When data processing is initiated, peak information is extracted, for each specified area, from mass spectrum data of all the measurement points. A peak method is applied to each area to extract peak information. Then, when a command to compare a set of areas is given, the peak information of those areas is collected. By comparing the peak information of different areas by a machine learning algorithm or similar judging technique, area-specific peak information is obtained, and this information is stored in memory.