[Evaluation of an automatic image-based screening technique for neonatal hyperbilirubinemia]

Area of Science:

• Neonatal Medicine
• Medical Imaging
• Biochemistry

Background:

• Neonatal jaundice is common, requiring accurate monitoring of total bilirubin (TB).
• Current methods may be invasive or lack objective assessment.
• Smartphone-based automated image-based bilirubin (AIB) testing offers a potential non-invasive solution.

Purpose of the Study:

• To evaluate the accuracy of a smartphone application for automated image-based bilirubin (AIB) testing.
• To compare AIB results with peripheral blood total bilirubin (TB) measurements in term and late preterm neonates.

Main Methods:

• AIB testing using smartphone-captured images and cloud-based analysis.
• Simultaneous measurement of AIB, transcutaneous bilirubin (TcB), and peripheral blood TB.
• Statistical analysis including t-tests, Bland-Altman analysis, and ROC curves.

Main Results:

• No significant difference was found between AIB and peripheral blood TB (P=0.109).
• Strong agreement was observed between AIB and peripheral blood TB (96% within limits of agreement).
• AIB demonstrated an AUC of 0.743 with 82% sensitivity and 60% specificity.

Conclusions:

• Automated image-based bilirubin (AIB) testing shows good agreement with peripheral blood TB.
• AIB is a promising new technique for objective jaundice monitoring in neonates.
• The sensitivity and specificity suggest AIB's utility for tracking jaundice progression and regression.