A predictive model for body water and fluid balance using 3D smartphone anthropometry

Area of Science:

• Biomedical Engineering
• Body Composition Analysis
• Health Technology Assessment

Background:

• Body fluid volumes (TBW, ECF, ICF) are key indicators of body composition and hydration status.
• Traditional bioelectrical impedance devices face application challenges for fluid volume assessment.
• 3D smartphone scanning offers a potential alternative for automated anthropometric and fluid volume estimations.

Purpose of the Study:

• To develop and validate prediction models for body fluid volumes using 3D smartphone scanning.
• To assess the diagnostic performance of smartphone-derived fluid overload and imbalance predictions.
• To explore the utility of smartphone scanning for body fluid assessment beyond clinical environments.

Main Methods:

• 338 participants underwent fluid volume assessment via bioelectrical impedance spectroscopy (BIS) and 3D smartphone scanning.
• LASSO regression developed TBW and ECF prediction models in a training subset (n=272), tested on a validation subset (n=66).
• Smartphone-derived ICF was calculated; fluid overload/imbalance were predicted using logistic regression and ROC analyses.

Main Results:

• Newly developed smartphone-based models showed no significant difference from BIS estimates (p ≥ 0.70).
• High agreement was observed for TBW (R²=0.91), ECF (R²=0.94), and ICF (R²=0.87) in the testing sample.
• Smartphone-predicted fluid overload (AUC=0.81) and imbalance (AUC=0.76) demonstrated acceptable diagnostic performance.

Conclusions:

• 3D smartphone scanning accurately assesses body fluid volumes and imbalances.
• This technology presents novel opportunities for remote health screening and monitoring.
• Smartphone-based fluid assessment can extend health monitoring capabilities beyond traditional clinical settings.