Automatic Analyzer Abnormality Detection via Reaction Curve Shape Descriptors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional automatic analyzers struggle to detect abnormal reactions outside photometric measurement points and accurately identify causes of abnormalities, leading to unreliable reexamination and potential recurrence of errors, due to dependence on operator experience and skill.

Innovation Solution

An automatic analyzer with a controller that generates an approximation curve from reaction process data, calculates shape descriptors, and performs abnormality judgment based on these descriptors to detect and identify causes of abnormalities, ensuring accurate and reliable analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional techniques using reference temporal data or function approximation are used to detect abnormalities, then abnormal reactions can be detected to some extent, but the causes of abnormalities cannot be identified and reexamination reliability decreases

Engineering Contradiction:
Improveabnormality detection accuracyVSAvoidcause identification capability
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent segments the reaction process into multiple phases (lag phase, exponential phase, linear phase) and analyzes each phase separately using shape descriptors. This segmentation allows the system to detect abnormalities in specific reaction phases and identify their causes by comparing phase characteristics against reference data, thereby resolving the contradiction between detecting abnormalities and identifying their causes.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces shape descriptors that capture the temporal shape characteristics of reaction curves beyond simple absorbance values. By transforming the one-dimensional absorbance data into multi-dimensional shape descriptors (such as slope, curvature, and timing parameters), the system gains the ability to both detect abnormalities and identify their causes through comparative analysis with reference curves.

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

2Loss of information

If operators manually confirm examination results and reaction processes, then cause identification can be performed, but the workload becomes excessive and accuracy depends on operator experience

Engineering Contradiction:
Improvecause identification capabilityVSAvoidanalysis throughput
Core Design Contradiction:
Loss of informationVSProductivity

Solution Approach 1:

The system performs self-diagnosis by automatically comparing measured reaction curves against reference temporal data and generating shape descriptors. The abnormality detection and cause identification processes are automated through computational algorithms that independently analyze reaction patterns, eliminating the need for operator intervention while maintaining high accuracy in cause identification.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical process of manual operator review with an automated information processing system. The controller automatically processes reaction process data, generates shape descriptors, compares them with reference data, and identifies causes of abnormalities through computational algorithms, substituting human cognitive processing with automated digital processing.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Adaptability or versatility

If reagent properties change over time, then measurement range varies, but conventional techniques cannot sufficiently detect abnormalities from each measurement result

Engineering Contradiction:
Improvereagent property variabilityVSAvoidabnormality detection capability
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent dynamically adapts the reference temporal data and shape descriptors to account for reagent property changes over time. The system continuously updates reference curves and analysis parameters based on observed reagent behavior, allowing it to detect abnormalities even when reagent properties vary temporally. This dynamic adaptation maintains measurement precision despite reagent degradation or batch differences.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP2594944B1Automatic analysis device, analysis method and information processing device
Publication Date: 2020.06.03 HITACHI HIGH TECH CORP
  • EP2594944B1 patent drawingFigure 1~2
  • EP2594944B1 patent drawingFigure 3~4
  • EP2594944B1 patent drawingFigure 5~6

AI summary

An automatic analyzer includes sample vessels 16 each containing a sample to be measured; reaction vessels 21 in which to mix a sample and a reagent; a sample dispenser 5 for dispensing a sample from any of the sample vessels 16 to any of the reaction vessels 21; reagent vessels 18 each containing a reagent to be mixed with a sample; a reagent dispenser 6 for dispensing a reagent from any of the reagent vessels 18 to any of the reaction vessels 21; a stirrer 7 for stirring the sample-reagent mix contained in any reaction vessel 21; and a photometric measurement unit 8 for obtaining multiple measurement data points during the progress of reaction of a mixed solution. The analyzer performs the steps of: selecting at least one approximation formula; generating an approximation curve from the measurement data points; calculating a shape descriptor from the approximation curve; and judging abnormalities based on the shape descriptor. This not only allows abnormalities to be detected accurately from each measurement result, but also allows the causes of the abnormalities to be identified.