Non-Invasive Glucose Monitoring Using mμSORS: A Population-Based Validation Study Across Diverse Patient Variability
Authors' conclusion
A single, overall model for mμSORS-based glucose monitoring achieves high accuracy across diverse patient subgroups, demonstrating its strong robustness.
Study context
- Population
- Patients with T2D; venous plasma glucose 3.8–31.6 mmol/L; depth-resolved Raman spectra by mμSORS4,768 paired samples
- Participants
- 200 patients
- Comparison
- Overall partial least squares (PLS) model vs stratified subgroup models for mμSORS glucose prediction
- Outcome
- Glucose prediction accuracy (MARD, root mean square error (RMSE), consensus error grid)
Methodological detailspopulation-based validation, subject-wise tenfold cross-validation, stratified analyses with size-matched comparisons
- Population-based validation
- Subject-wise tenfold cross-validation
- Stratified analyses with size-matched comparisons
Performance
As reported in the abstract
Glucose prediction accuracy
| Test or model | Mean absolute relative difference (MARD) | Consensus error grid (CEG) A + B |
|---|---|---|
| Overall PLS model | 14.5% | 99.4% |
Overall PLS model
Mean absolute relative difference (MARD)14.5%
Consensus error grid (CEG) A + B99.4%
- No significant differences across stratified subgroups
- P > 0.05
Further findings
| Outcome | P value |
|---|---|
| Stratified subgroup models vs overall model | P < 0.05 |
- Stratified subgroup models vs overall model: significantly worse; comparable after controlling for sample size
Original publication
Non-Invasive Glucose Monitoring Using mμSORS: A Population-Based Validation Study Across Diverse Patient Variability. J Diabetes Sci Technol. Published 18 August 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →