Virtual Hyperspectral Imaging of Eyelids - mHematology for Blood Hemoglobin Analysis

Background: Blood hemoglobin (Hgb) testing is a routine procedure in a variety of clinical situations. However, noninvasive, continuous, and real-time Hgb measurements are still challenging, deviating from clinical laboratory Hgb values. Photonics-based spectroscopy can provide a platform for reliable quantification of blood Hgb content, but requires bulky optical components that limit the development of compact, portable, and cost-efficient mHealth technologies. Methods: We introduce virtual hyperspectral imaging (VHI) that virtually transforms a built-in camera (i.e. RGB sensor) in the smartphone into a hyperspectral imager. As data-driven computational spectroscopy, VHI completely avoids the use of additional smartphone attachments. Our mobile version of VHI - mHematology combines imaging of peripheral tissue (i.e. conjunctiva) and spectroscopic quantification of total blood Hgb content without a priori personalized calibration. Findings: When compared with a wide range of blood Hgb values of human subjects, Bland-Altman analyses show excellent performance, comparable with laboratory blood Hg tests. Separate preliminary and testing datasets as well as intra-class correlations further strengthen the validation and utilization of VHI blood Hgb measurements. Interpretation: As mHematology of noninvasive and reliable blood hemoglobin measurements solely uses the smartphone camera, the key features include mobility, simplicity, and affordability (no additional cost for hardware) for rapid field adaptation in resource-limited or homecare settings. Funding: National Institutes of Health (NIH) NIH R21TW010620, United States Agency for International Development (USAID) 7200AA18CA00019, and Purdue University Shah Family Global Innovation Lab. The contents of this manuscript are the responsibility of the authors and do not necessarily reflect the views of USAID or the United States Government. Declaration of Interest: None. Ethical Approval: IRB and IREC were approved at Purdue University, USA and Moi University, Kenya.