Paper Number
PO36 

Session
Poster Session

Title
AI-powered single-cell analysis to probe in-vitro capillary blood flow

Presentation Date and Time
October 12, 2022 (Wednesday) 6:30

Track / Room
Poster Session / Riverwalk A

Authors (Click on name to view author profile)

  1. Recktenwald, Steffen M. (Saarland University, Department of Experimental Physics)
  2. Simionato, Greta (Saarland Campus University Hospital, Institute for Clinical and Experimental Surgery)
  3. Lopes, Marcelle (Cysmic GmbH)
  4. Kaestner, Lars (Saarland Campus University Hospital, Theoretical Medicine and Biosciences)
  5. Quint, Stephan (Cysmic GmbH)
  6. Wagner, Christian (Saarland University, Department of Experimental Physics)

Author and Affiliation Lines (in printed abstract book)
Steffen M. Recktenwald1, Greta Simionato2, Marcelle Lopes3, Lars Kaestner4, Stephan Quint3 and Christian Wagner1
1Department of Experimental Physics, Saarland University, Saarbruecken 66123, Germany; 2Institute for Clinical and Experimental Surgery, Saarland Campus University Hospital, Homburg 66421, Germany; 3Cysmic GmbH, Saarbruecken 66123, Germany; 4Theoretical Medicine and Biosciences, Saarland Campus University Hospital, Homburg 66421, Germany

Speaker / Presenter 
Recktenwald, Steffen M.

Keywords
experimental methods; AI based; bio-fluids; biomaterials; microscopy; ML based

Text of Abstract
Blood is mainly comprised of red blood cells (RBCs) that determine the unique flow properties of blood in the circulatory system. Healthy RBCs are characterized by high deformability, which enables them to dynamically adapt their shape to the flow conditions of the vessel. However, diseases, drugs, and medical treatments can affect the RBC deformability, thus leading to microvascular complications. In this study, we present a microfluidic-based approach that mimics the flow of RBCs under confined flow conditions as they are in the microcirculation. In narrow vessels, RBCs exhibit characteristic shapes that depend on their confinement, velocity, and intrinsic cell properties, such as cytosol viscosity and membrane viscoelasticity. Here, we use a high-precision pressure device to pump highly diluted RBC suspensions through microfluidic capillaries with a cross-section similar to the RBC size covering a broad pressure drop range. The single-file flow is recorded with a highspeed camera, and individual RBCs are analyzed with a customized artificial neural network to determine the cell morphology. Based on the cell morphology in the flow, we derive the shape phase diagram (fraction of cell shapes versus the cell velocity) for both healthy and diseased RBCs. Additionally, the cell size and distribution across the channel width are determined as a function of the cell velocity. We discuss how the microfluidic analysis of shape transitions and dynamics of RBCs in capillary flow can be used as a biomarker for diseases or treatment tests, to study the effect of drugs in-vitro, and for quality assessment of stored blood bags.