EMT-associated bias in the Parsortix® system observed with pancreatic cancer cell lines.

Abstract

Pancreatic cancer has a 5-year survival rate of 12%, highlighting the need for reliable biomarkers for early detection and disease monitoring. Circulating tumor cells (CTCs) have emerged as a promising biomarker, yet their detection remains challenging. This study evaluates the Parsortix® system, a microfluidic device that enriches CTCs based on size and deformability, using pancreatic cancer cell lines. As increasing evidence indicates that during epithelial-to-mesenchymal transition (EMT) a cell's deformability increases, we evaluated possible biases by the device. The EMT stage of three pancreatic cancer cell lines, CAPAN-1, MIA PaCa-2, and PANC-1, was assessed to classify them as epithelial, mesenchymal-like, and hybrid, respectively. Spike-in experiments showed that epithelial and hybrid phenotypes were more efficiently captured (62.6 ± 18.5% and 65.4 ± 11.1%) than mesenchymal-like cancer cells (32.8 ± 10.2%). These results were confirmed using an EMT-inducible breast cancer cell line. Lower recovery rates were found for the cells in a mesenchymal-like state (31.5 ± 6.4%) than those in an epithelial state (47.56 ± 7.2%). In conclusion, the Parsortix® device may underestimate the number of mesenchymal CTCs.