Light-guided dynamic phantom to mimic microvasculature for biomedical applications: an exploration for pulse oximeter

Abstract

Significance: Dynamic phantoms capable of changing optical properties by control are essential for standardizing and calibrating spectroscopy systems such as the pulse oximeter. However, current liquid dynamic phantoms containing human blood have a short shelf life and require complex experimental setups. Some solid dynamic phantoms are influenced by the angular-dependent performance of the liquid crystal display (LCD), some have a low spatial resolution, and some have slow control of optical properties.

Aim: We aimed to develop a solid dynamic phantom, which can overcome these obstacles by changing the optical properties rapidly and generating dynamic biological signals.

Approach: The absorption properties of the phantom can be controlled in real time by modulating an LCD. A light guide was employed to avoid the angular-dependent performance of the LCD by isolating the scattering top-layer tissue-mimicking silicone phantom from the LCD.

Results: The dynamic phantom was characterized at 940, 660, 530, and 455 nm to create a lookup table. Photoplethysmography signals of different heart rates from 80 to 120 beats per minute were synthesized, and oxygen saturation levels at 86%, 90%, 95%, and 100% were generated at multiple wavelengths.

Conclusions: The design, characterization, and potential applications of the dynamic phantom have been presented. This dynamic phantom can simulate various biological signals by applying corresponding modulation signals and has the potential to calibrate and validate pulse oximeter, imaging, and spectroscopy systems.

Keywords: diffuse optics; dynamic phantom; near-infrared spectroscopy; photoplethysmography; pulse oximeter; standardization.

Fig. 1

(a) Photo of the dynamic phantom. (b) Scheme of the dynamic phantom. (c) Scheme of the validation system.

Fig. 2

(a) Flowchart of dynamic phantom characterization. (b) Flowchart of modulation signal synthesis.

Fig. 3

(a) Characterization results of the dynamic phantom. (b) Characterization results of the dynamic phantom in the functional range. (c) Lookup table for the dynamic phantom.

Fig. 4

(a) Target PPG signal from the literature. (b) Normalized diffuse reflectance response at 660 nm. (c) Modulation signals for LCD at 660 nm to generate PPG signals. (d) PPG signals generated by the dynamic phantom and measured at 660 nm. (e) Percentage error between target and measured signal after max normalization of the target signal and the average percentage error. (f) PPG signals generated by the dynamic phantom and measured at 940, 660, 530, and 455 nm.

Fig. 5

(a) Offset plot of measured signals at 940 nm. Expected versus measured heart rate at (b) 940 nm, (c) 660 nm, (d) 530 nm, and (e) 455 nm.

Fig. 6

(a) Target signals at different oxygen saturation levels. (b) Measurement results at different oxygen saturation levels. (c) Expected ${\mathrm{SpO}}_2$ versus measured $R$ value. (d) Expected ${\mathrm{SpO}}_2$ versus measured ${\mathrm{SpO}}_2$.