External quality assurance of circulating tumor cell enumeration using the CellSearch® system: A feasibility study

• Published in: Cytometry. Part B, Clinical cytometry Vol. 80B; no. 2; pp. 112 - 118
• Main Authors: Kraan, Jaco, Sleijfer, Stefan, Strijbos, Michiel H., Ignatiadis, Michail, Peeters, Dieter, Pierga, Jean‐Yves, Farace, Francoise, Riethdorf, Sabine, Fehm, Tanja, Zorzino, Laura, Tibbe, Arjan G. J., Maestro, Marisa, Gisbert‐Criado, Rafael, Denton, Graeme, de Bono, Johann S., Dive, Caroline, Foekens, John A., Gratama, Jan W.
• Format: Journal Article
• Language: English
• Published: Hoboken Wiley Subscription Services, Inc., A Wiley Company 01.03.2011
• Subjects: Cell Count; circulating tumor cells; external quality assessment; Feasibility Studies; Flow Cytometry - standards; Humans; immunomagnetic enrichment; multicenter study; Neoplasms - blood; Neoplasms - pathology; Neoplastic Cells, Circulating - pathology; Quality Control; rare events; Reproducibility of Results
• ISSN: 1552-4949; 1552-4957
• DOI: 10.1002/cyto.b.20573

Background: Circulating tumor cells (CTCs) are cells that have detached from solid tumors and entered the blood. CTCs can be detected, among others, by semi‐automated immunomagnetic enrichment and image cytometry using CellSearch® (Veridex, Raritan, NJ). We studied the feasibility of external quality assurance (EQA) of the entire CellSearch procedure from blood draw to interpretation of results in multiple laboratories. Methods: Blood samples from six cancer patients and controls were distributed to 14 independent laboratories to test between‐laboratory, between‐assay, and between‐instrument variation. Additionally, between‐operator variability was assessed through the interpretation of blinded images of all blood samples on a website. Results: Shipment and storage of samples had no influence on CTC values. Between‐instrument (coefficient of variation (CV) < 12%) and between‐assay variation was low (CV ≤ 20%), indicating high reproducibility. However, between‐laboratory CV ranged from 45 to 64%. Although inter‐operator agreement on image interpretation (Fleiss' κ statistics) ranged from “substantial” to “almost perfect,” image interpretation, particularly of samples containing high numbers of apoptotic cells, was the main contributor to between‐laboratory variation. Conclusions: This multicenter study shows the feasibility of an EQA program for CTC detection in patient samples, and the importance of continuation of such a program for the harmonization of CTC enumeration. © 2010 International International Clinical Cytometry Society