1. ¹Academic Depertment of Biochemistry, Royal Marsden Hospital, London, UK
2. ²Department of Anatomic Pathology, University of Toronto, Toronto, ON, Canada
3. ³Histogenex, Antwerp, Belgium
4. ⁴Département de Pathologie, Centre Jean-Perrin, Clermont-Ferrand, France
5. ⁵Institute für Pathologie, Klinikum Kassel, Kassel, Germany
6. ⁶F Hoffmann-La Roche Ltd, Basel, Switzerland
7. ⁷Roche Diagnostics GmbH, Penzberg, Germany
8. ⁸Department of Pathology, Netherlands Cancer Institute, Amsterdam, The Netherlands

Correspondence: Professor M Dowsett, PhD, Academic Department of Biochemistry, Royal Marsden Hospital, London SW3 6JJ, UK. E-mail: mitch.dowsett@icr.ac.uk

Received 12 October 2006; Revised 12 February 2007; Accepted 14 February 2007; Published online 30 March 2007.

Abstract

Human epidermal growth factor receptor 2 (HER2) positivity in breast cancer is a prognostic factor regarding tumor aggressiveness and a predictive factor for response to trastuzumab (Herceptin^®). Early and accurate HER2 testing of all breast cancer patients at primary diagnosis is essential for optimal disease management. Routine HER2 tests, such as immunohistochemistry and fluorescence in situ hybridization (FISH), are subject to interlaboratory variation, and validation by laboratory proficiency testing is important to improve standardization. This study compared immunohistochemistry and FISH testing between five international pathology reference centers. Each center evaluated 20 immunohistochemistry and 20 FISH breast cancer specimens in five testing rounds. In each round, one center selected two sets of four different invasive tumor specimens (set A for immunohistochemistry and set B for FISH) and sent samples to the other four centers in a blinded manner, while retaining samples for its own evaluation. Results were analyzed by an independent coordinator. With immunohistochemistry, there were no differences between the five centers for any of the specimens at the level of diagnostic decision (positive or negative HER2 status). However, differences between laboratories were observed in immunohistochemistry scoring. Of the 20 specimens, four were scored as negative (0/1+) and five as positive (3+) in all centers; eight were negative or equivocal (2+), and three positive or equivocal. After FISH retesting of nine of the 11 equivocal immunohistochemistry cases, consensus was achieved in 15 of 18 (83%) specimens. FISH analysis of set B specimens resulted in consensus between centers in 16 of 20 (80%) specimens (six negative and 10 positive). All four discordant FISH specimens were scored as having HER2:CEP17 ratios within the range 1.7–2.3 by at least one center. Equivocal immunohistochemistry and borderline FISH cases are difficult to interpret, even for highly experienced and validated laboratories, highlighting the need for quality-control procedures.