Aim: Her2 protein is the key marker determining the choice of Herceptin therapy after a diagnosis of breast cancer. Its evaluation is made in most laboratories by immunohistochemistry, and interpreted by a pathologist using an optical microscope, a process subject to inter-observer variability, particularly for samples scored as equivocal (2+). Software analysis products have been introduced, seeking to reduce this variability. In this study, we compared the results of both traditional evaluation and a specific software package (VISIA Imaging) to those from fluorescent in situ hybridization (FISH).
Materials and methods: We selected 176 cases of invasive breast cancer sampled during 2012-2014 that were classified as equivocal after evaluation of Her2 immunohistochemistry, and that were also evaluated by FISH. Each tissue slide was scanned with a digital D-Sight Fluo 2.0 microscope and analysed with VISIA Imaging S.r.l. software. The final results were categorised as follows: negative (0-1+), equivocal (2+), or positive (3+). Then each result was compared to that obtained by FISH.
Result: The digital method confirmed 85 samples (48.3%) as equivocal (2+), while 23 (15.1%) were reclassified as negative (1+) and 44 (28.9%) as positive (3+). Of the 176 cases, 24 (13.6%) were not suitable for digital analysis (inadequate). Of 67 reclassified cases (1+ or 3+), 62 were in agreement with FISH results (concordance rate 92.5%). The sensitivity and specificity of the digital method were 100% and 82%, respectively.
Conclusion: The application of this analysis software led to an improvement in the interpretation of cases classified as equivocal, decreasing the need for FISH and increasing diagnostic certainty.
Keywords: Breast cancer; Digital analysis; Fluorescence; Her2; Immunohistochemistry.
Copyright © 2016. Published by Elsevier GmbH.