In Vivo Analyte Prediction Using NAS Algorithm Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing algorithms for predicting in vivo analyte concentrations, such as PLS and NAS, face performance deterioration due to distortion in the intrinsic spectrum of the analyte, particularly when the in vitro and in vivo spectra differ, leading to inaccurate predictions.
Innovation Solution
The method corrects the distortion in the in vivo intrinsic spectrum of the analyte using correction terms to improve the performance of the NAS algorithm by estimating the in vivo spectrum as a linear combination of the in vitro spectrum and correction terms, thereby enhancing prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the NAS algorithm uses an in vitro intrinsic spectrum for prediction, then the algorithm can be implemented, but the prediction accuracy deteriorates due to spectral distortion in vivo
Solution Approach 1:
The patent transforms the intrinsic spectrum from a fixed parameter obtained in vitro to a dynamic parameter that is corrected using in vivo measured spectra. By introducing correction terms that account for physiological variations (temperature, pH, protein binding), the spectrum parameters are adjusted to reflect actual in vivo conditions, thereby resolving the distortion issue and improving prediction accuracy
Solution Approach 2:
The patent introduces correction terms as intermediary elements between the in vitro intrinsic spectrum and the actual in vivo spectrum. These correction terms, derived from in vivo fasting spectra and spectral variation analysis, act as mediators that compensate for physiological distortions without requiring direct measurement of the true in vivo intrinsic spectrum
2Reliability
If the PLS algorithm periodically relearns spectral changes, then predictability is maintained, but the complexity and time consumption increase
Solution Approach 1:
The patent extracts the concentration prediction function from the complex PLS modeling process by using the NAS algorithm with corrected intrinsic spectra. Instead of periodically relearning the entire spectral model, the method separates the intrinsic spectrum correction (done once or occasionally) from the concentration prediction (done continuously), reducing computational complexity while maintaining reliability
Solution Approach 2:
The patent performs preliminary correction of the intrinsic spectrum using in vivo fasting spectra before the actual concentration prediction. This preliminary action accounts for physiological variations in advance, allowing subsequent predictions to be made quickly without requiring periodic relearning of the entire model
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for accurate prediction of in vivo analyte concentrations, as demonstrated by improved correlation between predicted and measured blood sugar levels, significantly enhancing the reliability of the NAS algorithm.
Implementation Method 1
an in vivo spectrum obtained as a result of an interaction between an analyte and the electromagnetic wave may be used... obtained by using an optical method such as infra-red spectroscopy
Implementation Method 2
the in vivo spectrum is one of an absorption spectrum and a reflection spectrum of infra-red light
Implementation Method 3
the in vivo spectrum is a dispersion spectrum of a single wavelength electromagnetic wave
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is a method for predicting an in vivo concentration of an analyte, including: estimating an in vivo intrinsic spectrum of the analyte; and predicting the in vivo concentration of the analyte by using a concentration predicting algorithm based on the estimated in vivo intrinsic spectrum and an in vivo spectrum obtained during a section in which the in vivo concentration of the analyte is not substantially changed.