Rare Disease Onset Estimation Model With Privacy-Preserving Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
General doctors face difficulties in making appropriate judgments on rare diseases, leading to missed opportunities for early diagnosis and treatment.
Innovation Solution
A learning device and method that constructs an estimation model using secret calculation AI to estimate the onset probability of rare diseases by preprocessing patient data from multiple medical institutions and performing multi-party calculation to learn the relationship between inspection and medical care data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If general doctors make judgments on rare diseases using conventional methods, then they can maintain simple diagnostic procedures, but diagnostic accuracy and early detection capability deteriorate due to lack of expertise and data
Solution Approach 1:
The patent introduces an estimation model as an intermediary tool between doctors and rare disease diagnosis. The model processes inspection data and medical care data to provide onset probability estimates, enabling doctors to make accurate judgments on rare diseases without requiring specialized expertise. This resolves the contradiction by providing expert-level diagnostic support through an automated system while maintaining the simplicity of clinical workflow.
Solution Approach 2:
The system performs preliminary analysis of patient data using the estimation model before final diagnosis. By pre-processing inspection data and medical care data to generate onset probability estimates, the system prepares diagnostic information in advance, allowing doctors to make more accurate and timely judgments. This preliminary action improves diagnostic accuracy without adding significant complexity to the overall process.
2Reliability
If patient data from multiple medical institutions is collected and processed, then the quality and reliability of the estimation model improves, but data security and privacy protection become more difficult to ensure
Solution Approach 1:
The patent introduces a secure data processing architecture where an intermediary system collects, preprocesses, and analyzes data from multiple medical institutions without exposing raw patient information. The estimation model learns from aggregated data patterns while patient-specific information remains protected. This intermediary layer enables model reliability improvement through multi-institutional data while maintaining data security and privacy protection.
3Measurement precision
If preprocessing is performed on inspection data and medical care data, then the accuracy of onset probability estimation improves, but processing time and computational resources increase
Solution Approach 1:
The system performs data preprocessing in advance to prepare inspection data and medical care data for estimation. By pre-processing data before it is needed for diagnosis, the system improves estimation accuracy while the actual diagnostic process benefits from having prepared data ready. This preliminary action resolves the time-accuracy tradeoff by shifting processing burden to non-critical periods.
Solution Approach 2:
The patent applies selective preprocessing only to the most critical and relevant data features that have the greatest impact on estimation accuracy. Rather than processing all data equally, the system identifies and preprocesses only the essential inspection data and medical care data elements, achieving high estimation accuracy with reduced processing time and computational resources.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A server (20) includes: a first preprocessing unit (24) that performs predetermined preprocessing on inspection data and medical care data of a patient; and a first estimation unit (25) that uses at least the inspection data and medical care data of a patient who has developed a rare disease after the preprocessing as learning data, and that estimates an onset probability of an estimation target patient for each of a plurality of rare diseases, based on the inspection data and the medical care data of the estimation target patient after the preprocessing, using a first estimation model (26) that learns a relationship between the inspection data and the medical care data of the patient and a onset probability of a rare disease of the patient, and that estimates an onset probability for each of the plurality of rare diseases.