A scalable workflow for sensitive detection and genomic profiling of tumour cells in lymph nodes from FFPE tissue.

Abstract

Accurate detection of lymph node (LN) metastases is critical for prognosis and surgical decision-making in head and neck squamous cell carcinoma (HNSCC) and other epithelial cancers. Standard histopathology examines only a small fraction of LN tissue, limiting sensitivity for micrometastases and isolated tumour cells. Scalable approaches are needed to improve staging accuracy and reduce overtreatment. We adapted the FDA-approved CELLSEARCH System, originally developed for circulating tumour cell detection, for use on dissociated 50 μm formalin-fixed paraffin-embedded (FFPE) LN sections. Tumour cells were enriched via EpCAM-based immunomagnetic capture, stained with cytokeratin (CK)/DAPI and quantified using the ACCEPT (Automated CTC Classification, Enumeration and Phenotyping Technology) image analysis tool. Specificity and reproducibility were tested in 50 metastasis-positive and 50 metastasis-free LNs from patients with HNSCC. Enriched cells underwent single-cell and pooled whole-genome amplification (WGA), quality-controlled by a genomic integrity index and profiled for copy number alterations (CNA) by low-pass whole-genome sequencing. Tumour cells were detected in 100% of LN metastases (range: 37-12,479), with counts strongly correlating with histological tumour area (p < 0.0001). Rare background events in control LNs were excluded by manual CELLSEARCH review. Based on their frequency, a threshold of > 12 cells was defined to classify samples as positive, yielding an assay specificity and sensitivity of 100%. Reproducibility across adjacent sections was high. Despite FFPE preservation, WGA succeeded in 27% of single cells and > 75% of pooled fractions. CNA profiling yielded 50 high-quality tumour genomes, revealing typical HNSCC alterations and marked inter-/intra-patient heterogeneity. This adapted CELLSEARCH workflow enables sensitive detection and genomic profiling of tumour cells from FFPE LN tissue. Demonstrated in HNSCC as a proof of concept, the approach is broadly applicable to epithelial cancers where LN staging is prognostically decisive. © 2026 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.