Skin Color Bias in Neonatal Transcutaneous Bilirubin Measurement: A Diffuse Reflectance Spectroscopy Approach With Melanin Correction.

Commercial jaundice meters rely on reflectance data from specific wavelengths for Transcutaneous bilirubin (TcB) measurements. However, skin color bias is introduced due to melanin contribution. In this study, we evaluated melanin concentration as an objective indicator of neonatal skin pigmentation and assessed the performance of melanin corrected TcB measurements (TcBcorr). A total of 333 neonates were categorized into four groups using the Neomar skin color scale. Melanin concentration estimated from neonatal reflectance spectra strongly correlated with Neomar classification (Spearman's r = 0.90), with group differences confirmed by Kruskal-Wallis and Dunn's tests. TcBcorr values showed good agreement with total serum bilirubin (TSB) across all color groups with minimal bias (-0.06 to 0.11 mg/dL). Compared with commercial TcBmeter readings, skin color bias is reduced, particularly in the darker groups (colors 3 and 4). Chromophore-decoupled diffuse reflectance spectroscopy offers a skin color unbiased, clinically reliable approach for TcB measurements across diverse populations.