Portal venous circulating tumor cells as a biomarker for relapse prediction in resected pancreatic cancer

Key findings

• Circulating tumor cells were detected far more often in portal than in peripheral venous blood. Among participants with epithelial cancer, the workflow found CTCs in 75% of intraoperative portal venous samples versus 40% of peripheral venous samples (median 6 versus 0 CTCs per 7.5 mL), at 89% detection specificity against 9 non-epithelial cancer controls.
• A positive portal venous CTC count predicted relapse in resected pancreatic cancer. Only portal venous detection was predictive: every patient above the five-CTC threshold relapsed (5 of 5) versus 33% (3 of 9) below it, a 6.67-fold relative risk for that threshold (odds ratio 20.43, p=0.031; relapse-prediction sensitivity 1.00, specificity 0.625), while none of the other clinicopathologic factors predicted relapse.
• In spike-in samples the workflow detected more circulating tumor cells than the microfiltration comparator. AccuCyte isolation recovered CTCs from 81% of surrogate samples versus 36% for the CellSieve comparator (p=0.009), while on the low-yield clinical samples the two methods detected CTCs equally (37.5% each).

The AccuCyte–CyteFinder workflow in the methods

“Blood samples (7.5 mL) were processed using the AccuCyte-CyteFinder platform, as previously described [ 21 ]. Briefly, blood nucleated cells were isolated using the AccuCyte system (RareCyte) and spread onto 8 Superfrost ® Plus slides (Thermo Scientific) using a plastic CyteSpreader ® device (RareCyte). Two slides (representing approximately 6 million nucleated cells) were immunofluorescently stained on a Autostainer Link 48 (Dako–Agilent Technologies) using the RarePlex 0700-MA staining kit (RareCyte) comprising a nuclear DAPI dye, anti-pan-cytokeratin (CK) antibody conjugated to CF ® 488, anti-EpCAM antibody conjugated to CF ® 647, and anti-CD45 antibody conjugated to R-phycoerythrin (PE). Stained slides were scanned on the CyteFinder digital immunofluorescent microscope (10X magnification; RareCyte)”

— Yeo et al., Cellular and Molecular Life Sciences (2025), Methods, “CTC isolation and enumeration using AccuCyte-CyteFinder platform”

Disclosure: RareCyte is named in the competing-interests statement of the publication cited above.

Why it matters for The AccuCyte–CyteFinder workflow users

If you are weighing circulating tumor cells as a prognostic readout in pancreatic cancer, look at what this study asked of its detection step. The signal rests on counting rare tumor cells reliably enough to set a threshold — here, five cells per 7.5 mL — and then showing that the threshold separates the patients who relapse from those who do not. That only holds if you keep every candidate cell in view. The AccuCyte–CyteFinder workflow isolates nucleated cells by density instead of selecting for an epithelial antigen first, spreads them across slides as a monolayer, and images the whole slide, so cells that carry little EpCAM are still present when scoring begins. The authors scored each cell on a fixed four-marker readout and called it a circulating tumor cell when it was nucleated, cytokeratin- or EpCAM-positive, and CD45-negative. The same imaging step also read the comparator's filters, standardizing analysis across two methods. For your own work the trade is explicit: you image a large nucleated-cell population rather than a pre-enriched subset, and that is what let this study detect portal venous tumor cells and hold a prognostic signal after adjustment.