CC BY
3The 30th International Conference on Advanced Laser Technologies ALT'23
B-I-23
Non-invasive assessment of hemoglobin using spatial frequency domain imaging and machine learning approaches
B. Yakimov1'2, K. Buiankin1, G. Denisenko2, I. Bardadin1, E. Shirshin1'2
1 - Faculty of Physics, M.V.Lomonosov Moscow State University, 1-2 Leninskie Gory, 119234 Moscow, Russia 2 - Laboratory of Clinical Biophotonics, Biomedical Science and Technology Park, Sechenov First Moscow State Medical University, Trubetskaya 8, Moscow 119048, Russia
Main author email address: bp.jakimov@physics.msu.ru
Hemoglobin (Hb) is the main protein in the human blood, responsible for oxygen transport in the body, thus, the assessment of the total Hb concentration in the patient's blood is of high importance. Especially, correct estimation of Hb blood concentration is important to detect the decreased level of total hemoglobin (anemia), as it affects approximately ~30% of woman and children [1], while undiagnosed or improperly treated anemia can lead to adverse outcomes and significant reduction of the life quality [1].
The gold standard of total Hb blood quantitation is an invasive blood sampling followed by the analysis using flow-cytometry-based or staining-based procedures. This method is time-consuming (on the order of several hours), requires qualified medical staff for blood sampling, and could be painful for the patient, so noninvasive methods of blood Hb level assessment are in high demand.
It was recently demonstrated [2], that the Hb level could be assessed with ~15 g/L error (the Hb normal range is 130-160 g/L for adult males and 120-150 g/L for adult females) using simple RGB-imaging with a smartphone camera by the analysis of the color of fingernails. Indeed, fingernails are a promising tissue site for the optical Hb non-invasive assessment, as it does not contain any other visible chromophores, such as melanin, and are easily accessible to any imaging technique. Yet, the estimation error in the non-invasive Hb assessment using RGB-imaging does not meet clinical standards to readily translate this method to a clinic. Here we explore whether advanced imaging techniques, such as spatial frequency domain imaging (SFDI), can superior simple RGB-imaging in the accuracy of Hb non-invasive assessment. As SFDI allows one not only to estimate relative changes in optical signal but quantitatively determine the values of the absorption and reduced scattering coefficients of tissues at the imaging wavelength, it surpasses the accuracy of other optical techniques and is a promising method for non-invasive Hb level assessment. The work was supported by the Russian Science Foundation (grant No. 22-25-00759).
[1] E. Gayat et al. "Non-invasive measurement of hemoglobin: assessment of two different point-of-care technologies." PloS one 7.1 (2012): e30065.
[2] R. Mannino et al. "Smartphone app for non-invasive detection of anemia using only patient-sourced photos." Nature communications 9.1 (2018): 4924.
