A System-on-Chip Assay for Bilirubin Levels Measurement in Whole Blood †

Bilirubin (BR) is clinically confirmed as a biomarker for liver health and is used to assess the prognosis of cirrhosis. Optical and chemical methods have been utilized for blood BR biosensing. While optical methods offer real-time monitoring and are handy and immune to infection, measurements may not be practical due to the instrument complexity and space requirements. This study investigated the dual-wavelength (DWL) technique for BR estimation using a system-on-chip (SoC). The SoC includes an optical module with blue (455 nm) and green (530 nm) LEDs which were used for DWL measurement. Porcine blood was used as a surrogate of human blood and BR levels were kept within the pathophysiological ranges projected from healthy individuals (<1.2 mg/dL) to a cirrhotic patient (up to 50 mg/dL). Our findings show a high BR sensitivity in blood and this lays the groundwork for point-of-care testing for BR levels primarily for hyperbilirubinemia infants and cirrhotic adults out in homes or in-community settings.


Introduction
Bilirubin is a by-product of heme catabolism [1], where less than 1.2 mg/dL is always present in the blood of healthy individuals [2]. It exists as conjugated and unconjugated bilirubin (UCB) in the bloodstream. Unconjugated bilirubin (UCB) is toxic and water-insoluble; hence it should be excreted. For excretion, UCB binds with albumin and is transported to the liver through the bloodstream. In the liver, UCB is enzymatically converted into conjugated bilirubin (CB). The latter is less toxic and water-soluble, allowing excretion into the urine or poop [3]. However, in the case of pathophysiological events, bilirubin levels are elevated, and the liver fails to live up to its conjugation ability leading to irreversible neurological damage or death [4]. High bilirubin levels have been correlated with hepatic and hemolytic disorders [5,6].
Several studies on BR levels estimation have been reported [7,8], but these techniques suffer from the instrument complexity and cost. This mini-paper investigated the dualwavelength (DWL) technique for BR estimation using a system-on-chip (SoC).
Blood sample preparation: Porcine blood was procured from the abattoir immediately after sacrifice and was preferred as a prominent replacement for human blood due to its biochemical resemblance [9] and low cost. Solutions were prepared by mixing BR at varied concentrations and blood with anticoagulant (4% w/v)-to-blood ratio of 1:9 (v/v)   [10]. BR-blood samples were stored at ~5 °C in the dark to prevent photodegradation. For extended preparation steps, refer to the study done by Ndabakuranye, et al. [11].
Optical measurement: A SoC platform was used for BR measurement. The setup consisted of a MAX86916 optical module, a MAX32630 host and an optical stage designed to operate in transflection mode (Figure 1). Major features of the system-on-chip are summarized in Table 1. Feature selection: Several features were optimally selected using theoretical investigations and simulations to ensure the accuracy, reliability, and safety of bilirubin measurements. These features are summarized in Table 2.

Results and Discussions
The analytical (470 nm) and reference (525 nm) wavelengths were obtained by analyzing the distinct optical signatures of blood and BR. It does not significantly absorb green light but strongly absorbs blue light (ε525 = 214, ε460 = 53869 [cm −1 M −1 ]) [12]. However, although our SoC has 455 and 530 nm as the maximum power outputs, they still include 470 and 525 nm respectively since LEDs are not ideal monochromatic sources.
The DWL measurement was performed using a revised 2-dimensional Beer's law as shown in Equation (1) which suggests that BR concentration can be correlated with its absorbance ratio at 470 and 525 nm (t: path length [cm], ε: extinction coefficient, C: concentration).
To investigate BR's sensitivity in blood, R-parameters were deduced by collecting the MAX86916 ADC counts data and calculated using Equation (2). Figure 1 shows the SoC platform used to measure the ADC data used to compute R-parameters. The plot of Rparameters vs. BR concentrations is shown in Figure 2, and results showed a strong linear relationship with R-square greater than 0.991.
Although this technique provides a robust and simple way of measuring BR, it may be susceptible to errors due to LEDs' spectral and spatial distribution inadequacies, residual bilirubin, other hemoglobin forms (COHb or MetHb), SaO2 and Hb levels variability.

Conclusions
The feasibility of BR monitoring by the DWL method was investigated. BR's sensitivity in blood was explored at pathophysiological ranges (1.2-50 mg/dL) using an Soc. The SoC includes a miniature MAX86916 optical module with integrated signal conditioning and processing capabilities. Results showed a strong correlation between R-parameters and BR concentration (R-squared > 0.99). Our findings lay the groundwork for pointof-care testing for BR levels, primarily for hyperbilirubinemia infants and cirrhotic adults out of clinical settings.

Institutional Review Board Statement:
Informed Consent Statement: Data Availability Statement: