Lung cancer intravasation-on-a-chip: Visualization and machine learning-assisted automatic quantification.

During lung cancer metastasis, tumor cells undergo epithelial-to-mesenchymal transition (EMT), enabling them to intravasate through the vascular barrier and enter the circulation before colonizing secondary sites. Here, a human in vitro microphysiological model of EMT-driven lung cancer intravasation-on-a-chip was developed and coupled with machine learning (ML)-assisted automatic identification and quantification of intravasation events. A robust EMT-inducing cocktail (EMT-IC) was formulated by augmenting macrophage-conditioned medium with transforming growth factor-β1. When introduced into microvascular networks (MVNs) in microfluidic devices, EMT-IC did not affect MVN stability and physiologically relevant barrier functions. To model lung cancer intravasation on-a-chip, EMT-IC was supplemented into co-cultures of lung tumor micromasses and MVNs. Wihin 24 h of exposure, EMT-IC facilitated the insertion of membrane protrusions of migratory A549 cells into microvascular structures, followed by successful intravasation. EMT-IC reduced key basement membrane and vascular junction proteins - laminin and VE-Cadherin - rendering vessel walls more permissive to intravasating cells. ML-assisted vessel segmentation combined with co-localization analysis to detect intravasation events confirmed that EMT induction significantly increased the number of intravasation events. Introducing metastatic (NCI-H1975) and non-metastatic (BEAS-2B) cell lines demonstrated that both, baseline intravasation potential and responsiveness to EMT-IC, are reflected in the metastatic predisposition of lung cancer cell lines, highlighting the model's universal applicability and cell-specific sensitivity. The reproducible detection of intravasation events in the established model provides a physiologically relevant platform to study processes of cancer metastasis with high spatio-temporal resolution and short timeframe. This approach holds promise for improved drug development and informed personalized patient treatment plans.