Bayesian State-Space Model for Atopic Dermatitis Severity Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assessing atopic dermatitis severity are complex, time-consuming, and lack precision, making it challenging to predict disease dynamics and tailor treatment strategies effectively for individual patients.
Innovation Solution
A computer-implemented method using time-series images of the skin, which involves segmenting atopic dermatitis lesions and predicting severity item evolution over time using a Bayesian state-space trained model, allowing for accurate and personalized treatment approaches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional scoring systems (EASI, SCORAD, SASSAD) are used to assess atopic dermatitis severity, then comprehensive severity evaluation is achieved, but the assessment becomes complex and time-consuming
Solution Approach 1:
The patent segments the complex severity assessment into distinct severity items (e.g., erythema, edema, excoriation, crusting) that can be independently evaluated and scored. Each severity item is assessed separately using standardized criteria, allowing for comprehensive evaluation while enabling parallel processing and reducing overall assessment time.
Solution Approach 2:
The patent introduces standardized severity item definitions and scoring criteria as intermediaries between the complex clinical presentation and the final severity score. These standardized intermediaries provide a systematic framework that simplifies the assessment process while maintaining measurement precision across different evaluators.
2Measurement precision
If traditional scoring systems are used to assess atopic dermatitis severity, then comprehensive severity evaluation is achieved, but the complexity of calculation increases
Solution Approach 1:
The patent divides the overall severity assessment into separate severity items, each with its own scoring criteria. This segmentation allows for independent calculation of each item's contribution to the total severity score, simplifying the overall calculation process while maintaining comprehensive evaluation.
Solution Approach 2:
The patent transforms the complex clinical assessment into standardized numerical parameters for each severity item. By converting qualitative clinical observations into quantifiable scores with defined weightings, the patent simplifies calculation while preserving measurement precision.
3Reliability
If static severity assessment is performed, then current disease state is captured, but prediction of future disease dynamics is not achieved
Solution Approach 1:
The patent performs preliminary assessment of multiple severity items at different time points to establish baseline patterns and trajectories. By analyzing the temporal evolution of each severity item, the system predicts future disease dynamics and flare-up risks before they fully manifest, enabling proactive treatment adjustments.
Solution Approach 2:
The patent implements a feedback mechanism where historical severity data from multiple time points is continuously analyzed to update predictions of future disease course. The system uses past severity patterns to inform future assessments, creating a dynamic prediction model that improves reliability over time.
Data Source
Figure 1~2
Figure 3a~4b
Figure 5
AI summary
The invention relates to a method implemented by computer means for predicting the atopic dermatitis severity dynamics of at least one area of a patient based on time-series images of the skin of said patient, said method including an inference phase comprising the following steps: (a) segmenting each image to delineate at least one atopic dermatitis lesion in said image, (b) predicting the evolution over time of at least one severity items of said segmented lesion, using a Bayesian state-space trained model for the prediction of each severity item.