A novel AI-coupled flow chamber method quantifying erythrocyte osmotic fragility

Osmotic fragility (OF) is widely used to evaluate red blood cell (RBC) membrane stability, water transport dynamics and hemoglobinopathies, traditionally via spectrophotometric, visual, or flow cytometric techniques. Here, we present a novel flow chamber-based platform integrated with a proprietary imaging software providing AI-driven RBC detection for automated and reproducible OF analysis (BioExP). To assess agreement, we compared MCF₅₀ values from four healthy donors using classical and flow chamber methods. Protocol optimization included determining the "satiation time" (optimal incubation required to induce maximal hemolysis without overexposure) and hemolysis kinetics. Biological sensitivity was tested using two modulators: HgCl₂ (40 µM) to inhibit aquaporin (AQP) channels and lipopolysaccharide (LPS, 1000 µg/mL) to increase membrane fragility. Both treatments caused significant shifts in MCF₅₀ when compared to control: AQP inhibition decreased MCF₅₀ to 0.37 ± 0.01% NaCl (flow chamber) and 0.40 ± 0.01% NaCl (classical), while LPS increased MCF₅₀ to 0.44 ± 0.01% NaCl and 0.47 ± 0.004% NaCl, respectively (p < 0.001 for all). The BioExP replicated classical OF measurements, captured donor-specific variability, and detected changes. Importantly, it demonstrated that LPS alone, in plasma-free conditions, can compromise RBC membrane integrity. The platform requires minimal sample volume and enables real-time imaging and multi-condition testing.