Automated Viability Estimation from Digital Holographic Microscopy: Validation on Heterogeneous Industrial Bioproduction Cultures

Cell viability is a critical parameter in bioproduction, yet most facilities still rely on manual, offline assays. This work introduces a new label-free Digital Holographic Microscopy (DHM)-based viability prediction pipeline using a simple optical design compatible with both on-line and in-line probe implementation. Unlike previous approaches validated under controled laboratory conditions, the proposed pipeline was designed to operate across diverse CHO bioprocess conditions without calibration or parameter tuning. It was validated on a large, heterogeneous dataset comprising 40 cell cultures collected from industrial and academic sites, spanning multiple cell lines, culture media, process modes and cell densities up to 100 million cells/mL. Beyond viability estimation, exploratory analyses suggest that DHM-based monitoring can provide additional process-relevant insights, including early detection of viability decline and correlation with recombinant protein titer. Together, these results indicate that DHM has the potential to enable a new generation of non-invasive, multiparametric monitoring tools for advanced bioproduction control.