What is the accuracy of HER2 testing results in breast tumors?

Updated: Mar 20, 2019
  • Author: Oudai Hassan, MD; Chief Editor: Chandandeep Nagi, MD  more...
  • Print
Answer

Answer

Quality concerns [13]

Both false-negative and false-positive results are an important concern in HER2 testing. [16, 17] HER2 testing should aim to identify all patients who would benefit from trastuzumab therapy. However, false-positive results are a significant concern because of the small risk of serious cardiac toxicity and the cost associated with trastuzumab therapy (>$70,000/year). Also, incorrect classification of cancer impedes efforts to evaluate the effectiveness of HER2-targeted therapy in clinical trials. False-negative results can occur from fixation times that are too short or too long, decalcification and mercury-containing fixatives (B5), storage of unstained sections for more than 1 week, insufficient antigen retrieval, and nonoptimal enzymatic digestion.

Excessive antigen retrieval and accidental assessment of in situ rather than invasive lesions can lead to false-positive results. Other causes of false-positive results includes edge artifact (lobular carcinoma can appear falsely positive on the edge), cytoplasmic positivity, and overinterpretation. ASCO and CAP developed a joint guideline to improve the accuracy of HER2 testing. [13]

Considerations concerning pre-analytic/analytic variables are as follows:

  • Appropriate fixation: Time from tissue acquisition to fixation should be as short as possible; specimens should be fixed in neutral buffered formalin for 6-72 hours; they should be sliced at 5- to 10-mm intervals and placed in sufficient volume of neutral buffered formalin; time to fixation and duration of fixation should be recorded for each sample

  • Strict adherence to test protocols is essential

  • Appropriate controls include positive controls and negative controls (eg, cell line controls and/or tissue-based references) and an internal control (normal breast epithelium)

Considerations during pathologist interpretation of immunohistochemistry are as follows:

  • Care should be taken to make sure only the invasive component is assessed and reported, as HER2 is often overexpressed and/or amplified in in situ breast lesions.

  • When possible, the test should be performed on a block with an internal control, because this internal control will be an indicator of excessive antigen retrieval; if normal ducts and lobules show strong membrane staining, excessive antigen retrieval may have occurred, and repeat testing or testing via FISH should be performed.

  • If cytoplasmic staining obscures membrane staining or if obscuring artifacts (edge, retraction, crush artifact) are present, the assay should be repeated or FISH should be performed.

  • HER2 overexpression correlates with high tumor grade and is less common in pure tubular carcinomas, colloid carcinomas, and classic lobular carcinoma [18, 19] ; results that are unexpected should be questioned with repeat testing or testing with an alternative method or a different sample, if indicated.

  • If the results on core biopsy are questioned, repeat testing on an excision is warranted.

  • Negative results on unstained sections stored for prolonged periods should be questioned; sections ideally should not be used if they were cut more than 6 weeks earlier.

  • Negative results in bone biopsies should be questioned because of the possible effects of decalcification or B5 fixative. (Fine-needle aspiration or touch preparation samples are a good alternative.)

  • Equivocal results (2+) must be subjected to FISH testing.

  • If samples were known to be fixed for periods shorter than 6 hours or longer than 72 hours, the report should qualify any negative result with a statement about fixation.

Regarding FISH, according to ASCO/CAP guidelines, counting can be performed by a trained technologist, but the pathologist must confirm that the result (count) is correct, that invasive tumor was counted, and that the sample is surveyed for genomic heterogeneity. [13] Other considerations are as follows:

  • Care should be taken to make sure only the invasive component is assessed; HER2 is often overexpressed and/or amplified in in situ breast lesions; corresponding hematoxylin and eosin (H&E) and/or IHC slides should be reviewed to localize the invasive carcinoma.

  • Control samples should be reviewed; if the findings are not as expected, the test should be repeated.

  • At least 20 nonoverlapping cells in two separate areas of invasive carcinoma should be counted; counting of additional cells and/or by a different observer may be appropriate.

  • Samples with nonuniform signals (weak signals in >25% of cells), high autofluorescence, poor nuclear resolution (indistinct nuclei), or high background-obscuring signal resolution (>10% of signals over cytoplasm) should be rejected.

  • HER2 overexpression correlates with high tumor grade and is less common in pure tubular carcinomas, colloid carcinomas, and classic lobular carcinoma [18, 19] ; results that are unexpected should be questioned with repeat testing or testing with an alternative method or a different sample, if indicated.

  • If the results on core biopsy are questioned, repeat testing on an excision is warranted.

  • Negative results on unstained sections stored for prolonged periods should be questioned; sections ideally should not be used if they were cut more than 6 weeks earlier.

  • Negative results in bone biopsies should be questioned because of the possible effects of decalcification or B5 fixative. (Fine-needle aspiration or touch preparation samples are a good alternative.)

  • Equivocal FISH results are to be confirmed by counting additional cells or repeating the FISH test; if the FISH result remains equivocal, IHC is recommended.

  • If samples were known to be fixed for periods shorter than 6 hours or longer than 72 hours, the report should qualify any negative result with a statement about fixation.


Did this answer your question?
Additional feedback? (Optional)
Thank you for your feedback!