Counting invasive breast cancer cells in the HER2 silver in‐situ hybridization test: how many cells are enough?. Issue 2 (26th April 2017)

• Record Type: Journal Article
• Authors: Polónia, António; Eloy, Catarina; Pinto, João; Braga, Ana Costa; Oliveira, Guilherme; Schmitt, Fernando
• Is Part Of: Histopathology. Volume 71:Issue 2(2017)
• Page Start: 247
• Page End: 257
• Publication Date: 2017-04-26
• Subjects: ASCO/CAP -- breast cancer -- HER2 -- SISH; Histology, Pathological -- Periodicals; 611.018
• DOI: 10.1111/his.13208
• Languages: English
• ISSNs: 0309-0167

Abstract

Abstract : Aim: To evaluate the intraobserver and interobserver reproducibility of the HER2 in‐situ hybridization (ISH) test in breast cancer by measuring the impact of counting different numbers of invasive cancer cells. Methods and results: A cohort of 101 primary invasive breast cancer cases were evaluated for HER2 gene amplification by silver ISH, and the concordance among four observers with different levels of experience, counting different numbers of invasive cancer cells, was determined. The evaluation of the samples included scoring 20 nuclei, in three different areas. The cases were scored twice, with a washout interval of at least 2 weeks. We observed an increase in the intraobserver concordance rate between the first and second evaluations with an increase in cell count. A count of 60 invasive cells was needed to obtain a concordance rate near 95% and an agreement rate greater than 0.80 by all observers. The interobserver concordance rate of the HER2 test also increased with the increase in cell count, reaching at least a 90% concordance rate with a count of 60 invasive cells. The median variability of both the HER2 /CEP17 ratio and the average HER2 copy number between different evaluations decreased with the increase in cell count, being statistically higher in HER2 ‐positive cases. Conclusions: The minimal cell number recommended in current guidelines should be raised to at least 40, and preferably 60, invasive cells. Moreover, cases with amplification levels close to the threshold should be subjected to a dual count from an experienced observer.