Biases in the Parsortix system observed with pancreatic cancer cell lines

Pancreatic cancer has a 5-year survival rate of merely 12%. The high rate of late-stage diagnoses underscores 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 due to their rarity and phenotypic diversity. This study evaluates the Parsortix(R) system, a microfluidic device designed to enrich CTCs based on size and deformability, using pancreatic cancer cell lines. As increasing evidence indicates that during epithelial to mesenchymal transition (EMT) a cells deformability increases, we evaluated to what extent the Parsortix(R) system was biased towards epithelial cancer cells. First, the EMT stage of three pancreatic cancer cell lines, CAPAN-1, PANC-1 and MIA PaCa-2, was assessed using immunocytochemistry, flow cytometry and proteomics. CAPAN-1 cells were classified as epithelial, MIA PaCa-2 cells exhibited a mesenchymal-like phenotype, and PANC-1 cells demonstrated a hybrid phenotype. Then, by spiking these cells into blood samples, we determined the Parsortix(R) systems ability to recover the cancer cells. Our results indicated that epithelial and hybrid phenotypes are more efficiently captured (62.6 {+/-} 18.5% and 65.4 {+/-} 11.1%) than mesenchymal cancer cells (32.8 {+/-} 10.2%). To confirm these findings, spike-in experiments were repeated using an EMT inducible cell line. Again, significantly lower recovery rates were found for the cells in a mesenchymal-like state (31.5 {+/-} 6.4%) compared to those in an epithelial state (47.56 {+/-} 7.2%). In conclusion, the Parsortix(R) system may underestimate the presence of mesenchymal CTCs.