Smart-phone-integrated, non-invasive, depth-resolved optical spectroscopy for the detection of neonatal jaundice

Newborns have immature liver function that is inefficient at metabolizing bilirubin. Consequently, nearly 80% of preterm and 60% of term babies develop hyperbilirubinemia resulting in neonatal jaundice within a week of birth. Severe hyperbilirubinemia can be fatal, making early, frequent, and accurate monitoring of bilirubin vital to avoid severe health issues and determine appropriate treatment. The gold standard for detecting hyperbilirubinemia is an invasive blood test to measure total serum bilirubin (TSB); however, frequent blood sampling in neonates is costly, painful, and increases the risk of infection. Existing non-invasive methods to monitor hyperbilirubinemia lack sufficient accuracy to replace blood tests. Commercial transcutaneous bilirubinometry (TcB), although clinically accepted for screening, has low correlation with TSB for clinical decision-making in dark-skinned neonates and in neonates undergoing phototherapy.
A main reason for TcB limitations is spectral cross-talk: the inability to reliably distinguish contributions between skin analytes (e.g., melanin) and blood. Our central hypothesis is that a non-invasive mobile phone-based bilirubin detector can be developed that provides accurate, point-of-care blood bilirubin measurements in dark-skinned neonates and neonates where TcB underperforms. We propose to use spectroscopic optical coherence tomography (sOCT), an imaging technique with depth-resolved capabilities that can overcome spectral cross-talk.
We will pursue three specific aims: (1) Develop a portable, mobile phone-integrated sOCT system for non-invasive, depth-resolved measurement of blood spectra. We will build and characterize a miniaturized sOCT device integrated with a smartphone application for data processing, analysis, display, and HIPAA-compliant transmission or storage. (2) Refine and test the sOCT algorithm in vivo. We will compare sOCT data to TSB blood tests and commercial TcB in 100 neonates, evaluating correlation, mean bias, and precision across a diverse range of skin tones. (3) Aim 3: Validate sOCT performance in neonates with poor TcB correlation. We will test sOCT in 78 neonates at Vanderbilt with darker skin tones to assess whether sOCT provides comparable results to lab-based TSB in populations where TcB underperforms. We will also perform usability and workflow assessments to optimize clinical integration.

Grant Number: 5R01EB032382-04
NIH Institute/Center: NIH
Principal Investigator: Audrey Bowden