The FDA-Approved Breast Cancer HER2 Evaluation kit (HercepTest; Dako) May Miss Some HER2-Positive Breast Cancers

Frank Schneider, MD; Yulan Jin, MD, PhD; Kevin Van Smaalen, MD; Evin H. Gulbahce, MD; Rachel E. Factor, MD; Xiaoxian Li, MD, PhD


Am J Clin Pathol. 2019;151(5):504-510. 

In This Article


The FDA-approved HercepTest is widely used for breast cancer HER2 IHC testing.[28,29]Our results show that five (around 22.7%) of 22 FISH-positive cases were not detected (false negative) by the FDA-approved method. The negative IHC results of these five cases were confirmed by a reference laboratory using the same tissues and the same FDA kit. All five patients received HER2-targeted therapy and benefited clinically. Compared with the FDA kit, our LDT showed higher sensitivity (100% vs 77.3%) by classifying all 22 FISH-positive cases as either positive (five) or equivocal (17). Solomon et al[30] showed that six of 61 HER2+ cases in their study were FISH positive but IHC negative using a rabbit monoclonal antibody (clone 4B5; Roche/Ventana). A study from Scandinavia compared IHC and FISH testing in 538 cases and found that 11 FISH HER2-positive cases were negative by the HercepTest kit.[29] Similar results were shown by other studies.[31]

We also found that a significant number (17/38 [44.7%]) of the LDT-equivocal cases were scored as negative by the FDA kit. Five of these cases were positive by FISH analysis. Such discrepancy has been reported by multiple studies.[18,25–27,32] The discrepancy between the FDA-approved kit and LDT could be caused by many factors, including preanalytical, analytical, and postanalytical factors. In our study, preanalytical factors were an unlikely cause of the discrepancy since all stains were performed on the same tissue blocks. Studies have shown that IHC staining interpretation could lead to false-positive or false-negative results and had great interobserver variation,[33–35] which was also unlikely in our study since the same pathologists scored IHC staining in both methods. Both the FDA kit and LDT used the same antibody clone. The two tests differed in using different stainers. The FDA kit used the Dako Link 48 automated stainer with the manufacturer's recommended protocol. Our LDT used a Leica Bond III automated stainer. Discrepancies between different IHC staining protocols have been reported. Layfield et al[36] reported an absolute agreement of 58% between the HercepTest antibody and the 4B5 clone antibody, which was also a rabbit monoclonal HER2 antibody. Powell et al[37]compared the 4B5 clone antibody and the CB11 clone antibody, a mouse monoclonal HER2 antibody, and reported a concordance rate of 93.3% between the two antibodies in a single institutional cohort and 84.7% in multi-institutional cohorts. The lower concordance in the multi-institutional cohorts might due to preanalytical variations, including tissue fixation and processing.

More IHC-equivocal cases have been identified since publication of the 2013 ASCO/CAP guidelines. To increase the sensitivity of detecting HER2+ cancers, some authors suggest to reflex not only IHC-equivocal but also IHC-negative cases to FISH testing, given the higher sensitivity and accuracy of FISH compared with IHC assays.[14] This strategy could potentially introduce economic burden and false positivity because of the possibility of unfunctional gene amplification without downstream protein overexpression.[38] Our results support that reflex FISH testing should be performed only on cases that are equivocal by HER2 IHC. In this study, all FISH-positive cases were scored as equivocal or positive by our LDT. The concordance rate between FISH and the LDT was 100%. The overall concordance rate between the FDA kit and FISH testing was also high (90.4%) in our study. Similar concordance rates were reported by others.[29,39]

The biggest limitation of our study is the small sample size. The second limitation could be the selection bias since we randomly selected the cases with a target number of HER2+ cases, which accounted for a higher proportion of HER2+ cases than that in an unselected population. These limitations may give a falsely high discordance rate between the FDA-approved HercepTest kit and FISH analysis in HER2+ cases. However, other studies also reported such a discrepancy between the HercepTest kit and FISH results. Kovács and Stenman[29] studied 103 cases with discordant IHC results between the HercepTest and the A0485 HER2 antibody. FISH analysis was performed in these 103 cases, and 11 were HER2 gene amplified by FISH. All these 11 FISH-positive cases were scored negative by HercepTest. Jørgensen et al[39] showed that eight (3.2%) of 247 FISH-positive cases were scored as HER2 negative by the HercepTest. In our study, the five discrepant cases were retested by a reference laboratory using the same HercepTest kit. Therefore, it is important to be aware that some HER2+ cases could be missed by the FDA-approved HercepTest kit. Some LDTs could outperform the HercepTest kit in this regard.

In conclusion, although the ASCO/CAP guideline preferentially recommends the use of an assay that has received FDA approval, our and other data showed that an FDA-approved method may miss some HER2+ cases and thus lead to suboptimal treatments. Our results further showed that a significant number of cases that were scored as equivocal by our LDT were scored as negative by the HercepTest kit. Our results indicate that an LDT can be superior to the FDA-approved HercepTest kit in terms of sensitivity and accuracy.