# Development and Evaluation of a Smartphone App-Based Rapid 25-Hydroxy Vitamin D Test

## TL;DR

A smartphone app-based test for vitamin D levels was developed and shown to be accurate and reliable for point-of-care use.

## Contribution

A smartphone-integrated sandwich-type LFA with automated image analysis for semi-quantitative 25(OH)D measurement was developed and validated.

## Key findings

- The test achieved a detection range of 5–100 ng/mL with minimal interference.
- The system accurately classified vitamin D levels into clinical categories with high reproducibility.
- It showed strong agreement with a commercial analyzer and 100% classification agreement between sample types.

## Abstract

Objectives: The purpose of this study is to develop and verify a sandwich-type lateral flow immunoassay (LFA) integrated with a smartphone, enabling semi-quantitative 25-hydroxyvitamin D [25(OH)D] measurement including automated image analysis function, thereby establishing a reliable and accessible vitamin D evaluation system for point-of-care (POCT). Methods: A smartphone-based sandwich-type LFA was constructed, and 25(OH)D was measured semi-quantitatively. The system combined a customized test strip with an automatic image acquisition, calibration, and classification module integrated into an application dedicated to a smartphone. Analysis performance, reproducibility, and equivalence between sample types were comprehensively evaluated. Results: The developed analysis achieved a detection range of 5–100 ng/mL, and there were little interference and cross-reactivity for endogenous substances or structurally similar vitamin D derivatives. The image processing algorithm accurately classified the samples into three clinically important categories: deficiency (<20 ng/mL), insufficient (20–30 ng/mL), and sufficient (>30 ng/mL). Cross-platform testing between Android and iOS devices showed excellent reproducibility (r = 0.99, R2 = 0.9967). Comparative analysis with the Atellica IM 1600 analyzer showed a high degree of agreement between 97.0% category consensus and κ = 0.951 (r = 0.99, R2 ≥ 0.98). Comparative tests between serum and capillary samples also confirmed a 100% classification agreement rate and an overall diagnostic accuracy of 95.5%. Conclusions: This next-generation smartphone integration platform enables rapid, accurate, and semi-quantitative detection of 25(OH)D from fingerstick and serum specimens. By combining the sandwich-type LFA design with computational-based imaging analysis, the system effectively overcomes the major limitations of small-molecule immunoassay and has the potential to be applied to field diagnosis (POCT), decentralized diagnostics, and vitamin D screening in large populations.