Article Text

Download PDFPDF

HER2 expression in ovarian carcinoma: caution and complexity in biomarker analysis
  1. Hugh McCaughan,
  2. InHwa Um,
  3. Simon P Langdon,
  4. David J Harrison,
  5. Dana Faratian
  1. Edinburgh Breakthrough Research Unit, Division of Pathology, Institute of Genetics and Molecular Medicine, University of Edinburgh, Edinburgh, UK
  1. Correspondence to Dr Dana Faratian, Division of Pathology, Institute of Genetics and Molecular Medicine, University of Edinburgh, Crewe Road South, Edinburgh EH4 2XU, UK; d.faratian{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

We read with interest the recent article by Yan et al,1 in which a small cohort (n=85) of ovarian tumours were assessed for HER2 amplification by immunohistochemistry (IHC), fluorescence in situ hybridisation (FISH) and chromogenic in situ hybridisation. HER2 amplification and protein overexpression were identified in 35.3% (6/17) and 29.4% (5/17) of primary mucinous carcinomas, respectively, with none of the other histological subtypes demonstrating HER2 amplification or overexpression. In another recent study, Tan et al.2 used array-based comparative genomic hybridisation to establish the genomic profiles and regions of recurrent copy number change in 50 pure ovarian clear cell carcinomas (OCCCs). In that study, 7/50 (14%) OCCCs were found to harbour HER2 gene amplification and protein overexpression, representing <2% of the ovarian cancer population as a whole. The authors in both studies quite rightly conclude that HER2 may therefore be a potential therapeutic target in non-serous ovarian tumours. However, since there is documented evidence of HER2 overexpression in other histological subtypes of ovarian cancer, including serous carcinomas3–5 and some HER2-directed therapies (such as pertuzumab), may be effective against tumours which express non-amplified levels of the receptor,6 we sought to better define the frequency, quantity and distribution of HER2 expression in a large retrospective series of clinically annotated ovarian cancers representing all histological subtypes. In doing so, we confirm that HER2 gene amplification and protein overexpression is not just limited to mucinous or OCCCs, and present in a larger proportion of ovarian carcinomas of all histological types. Several of these tumours show marked intra-tumoural heterogeneity in both morphological pattern and gene or protein expression. Quantitative expression of HER2 defines a larger proportion of poor-prognosis tumours which may benefit from HER2-directed therapy. The exact impact of target quantity and heterogeneity on therapeutic efficacy has yet to be established, but should be considered early in the assay development process.

We constructed tissue microarrays in triplicate taken from 468 formalin-fixed, paraffin-embedded specimens of epithelial ovarian cancers collected in Scotland between 1988 and 2006 as previously described (see Faratian et al7 and table 1); mixed histology tumours were represented in replicate cores. The study was approved by the Lothian Research Ethics Committee (06/S1101/41). The median overall survival from diagnosis was 32.7 months (range 1.2–211.0 months) and the 5-year overall survival was 21%. HER2 protein expression was assessed using IHC (Herceptest, DAKO) and quantitative immunofluorescence (AQUA, HistoRx7). Gene amplification was assessed by FISH using the Pathvysion HER2 DNA Probe Kit (Vysis). HER2:CEP17 ratios >2.0 were considered amplified. Overall, HER2 gene amplification was present in 20/456 (4.4%) ovarian cancers, 19 of which were 2+ or 3+ positive and one case 1+, by IHC. HER2 gene amplification was closely associated with HER2 protein expression (average HER2 histoscore 27.5 vs 244 in HER2 non-amplified vs HER2 amplified, p<0.0001, Mann–Whitney U test). The distribution of HER2 amplification between histological types was 7/259 (3%) serous papillary carcinomas, 2/92 (2.1%) endometrioid, 3/12 (25%) cases of mucinous type, 1/25 (4%) of clear cell type and 7/60 (11.9%) of mixed type. Therefore, HER2 amplification is not confined only to OCCCs and the other histological types may also benefit from HER2-targeted therapy.

Table 1

Clinical and pathological information

Since 7/20 (35%) HER2-amplified tumours were of mixed histological pattern, not uncommon in epithelial ovarian cancer, we further examined the histological patterns represented in these tumours to assess whether any contained mucinous or clear cell areas to account for the relatively low frequency of HER2 amplification in these subtypes in this cohort, compared to the 35.3% of mucinous observed by Yan et al, and 14% OCCCs observed by Tan et al. It was found that 2/7 tumours contained areas of clear cell differentiation. To assess whether similar heterogeneity was observable at the molecular level, we also compared IHC and FISH scores between tissue microarray cores in HER2-positive cases. Four (20%) cases with HER2-positive areas showed marked heterogeneity in expression by IHC, an example of which is shown in figure 1. This marked intratumoural heterogeneity has been noted previously,5 and we have recently demonstrated variable protein expression of HER2 in ovarian cancer using quantitative immunofluorescence.8

Figure 1

Photomicrograph (× 20) showing marked heterogeneity of HER2 expression in a single tumour.

There was no association between IHC or FISH positive cases and progression-free survival in univariate analysis (figure 2A), probably due to the overall low frequency of positivity (log-rank p=0.60). We therefore sought to use a more quantitative and sensitive HER2 protein detection methodology (AQUA quantitative immunofluorescence) in order to better assess the prognostic value of HER2 expression and define patient populations who may benefit from HER2-directed therapies (such as pertuzumab) which may also be effective in HER2 expressing, but unamplified, population. Quantitative immunofluorescence may have advantages over traditional IHC in terms of sensitivity, wide dynamic range of detection, and the reproducibility and reduction in bias afforded by computer-assisted analysis. Using optimal cutpoint determination method, AQUA scores >1869 (representing 58% of the total cohort) were significantly associated with reduced progression-free survival in univariate analysis (log-rank p=0.008, figure 2). When taken together with the other significant variables (stage and histological type, log-rank p<0.00017) in multivariable analysis, HER2 expression remained a significant prognostic factor (Cox Regression model, HR 1.35, 95% CI 1.08 to 1.68; p=0.008) along with stage (HR 1.50, 95% CI 1.25 to 1.73; p<0.0001).

Figure 2

Prognostic value of HER2 expression. (A) Kaplan–Meier analysis shows no significant difference in progression-free survival between patients with HER2 positive and negative tumours measured by immunohistochemistry/fluorescence in situ hybridisation (log-rank p=0.60). (B) When measured by quantitative immunofluorescence, high HER2 expression is associated with shorter progression-free survival. This difference was retained in multivariate analysis (Cox Regression: p=0.008).

In conclusion, the overall rates of HER2 overexpression and amplification in this large retrospective series are consistent with previously published data,4 ,5 and are not confined to mucinous or OCCCs, which might unduly limit the target population in ovarian cancer patients as a whole. Heterogeneity of HER2 expression has been previously observed5 ,8 and was also observed at relatively high frequency in tumours in this series. This may have a significant impact on the efficacy of HER2-targeted treatments. Finally, tumours with higher quantitative HER2 expression are associated with poorer prognosis, as described in some previous series,3 and given the availability of agents which are not reliant on HER2 amplification, may define a larger population of high-risk patients who may benefit from HER2-directed therapy. HER2 expression in ovarian carcinomas is complex and caution must be applied when defining the target population.


The authors would like to acknowledge the Breakthrough Breast Cancer and Scottish Funding Council for support of this work.


View Abstract


  • Competing interests None.

  • Ethics approval The study was approved by Lothian Research Ethics Committee (08/S1101/41).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles