Aims To determine the frequency of HER2 genetic heterogeneity according to the recent American Society of Clinical Oncology (ASCO) and College of American Pathologists (CAP) definition (2009) in invasive breast carcinoma, and to identify clinicopathological features that characterise breast carcinomas with HER2 genetic heterogeneity.
Methods 530 invasive breast carcinomas were retrospectively analysed for HER2 genetic heterogeneity, and investigated for a potential association of HER2 genetic heterogeneity with other HER2 FISH findings, clinicopathological parameters, oestrogen/progesterone receptor expression and DNA cytometric parameters in breast carcinomas with an equivocal (2+) HER2 immunohistochemical score.
Results The overall frequency of HER2 genetic heterogeneity was 14.7% in a cohort of 218 consecutive breast carcinomas. HER2 genetic heterogeneity was most frequent in invasive breast carcinomas with an equivocal (2+) HER2 immunohistochemical score. Among the 151 carcinomas lacking HER2 amplification, 16.1% showed HER2 genetic heterogeneity. In an extended cohort of 345 carcinomas with a (2+) HER2 score, the frequency of HER2 genetic heterogeneity was 41%, and was associated with the absence of HER2 gene clusters, chromosome 17 polysomy, histological tumour grade, DNA ploidy category and 5c exceeding rate.
Conclusion HER2 genetic heterogeneity according to the ASCO/CAP definition is frequent in breast carcinoma, and is most often present in carcinomas with an equivocal (2+) HER2 score. Many carcinomas with HER2 genetic heterogeneity have a negative HER2 amplification status, although they contain a significant number of tumour cells with HER2 gene amplification. Single cell scoring of the HER2/17 centromeric probe (CEP17) ratio is necessary to identify carcinomas with HER2 genetic heterogeneity, because they lack specific clinicopathological characteristics.
- Breast cancer
- genetic heterogeneity
- liver disease
- molecular pathology
Statistics from Altmetric.com
Trastuzumab treatment in patients with invasive breast cancer is based on the HER2 status of carcinoma cells.1 Intratumoral heterogeneity may hamper reliable evaluation of the HER2 status.2–9 Intratumoral heterogeneity of both HER2 protein expression and HER2 gene amplification has been observed. Furthermore, differences in HER2 protein expression and HER2 gene amplification between primary carcinoma and metastatic manifestations can occur within the same patient.10–13 As a result of the usage of different definitions, the reported frequencies of HER2 intratumoral heterogeneity have varied significantly. However, most studies have concluded that the frequency of HER2 intratumoral heterogeneity is low, at least in breast carcinomas with high-grade HER2 gene amplification, and that HER2 intratumoral heterogeneity does not affect the selection of patients for trastuzumab therapy.6–8 14
Recently, guidelines of the American Society of Clinical Oncology (ASCO) and the College of American Pathologists (CAP) have been published that define HER2 genetic heterogeneity as the presence of more than 5%, but less than 50%, of infiltrating tumour cells with a HER2/17 centromeric probe (CEP17) ratio higher than 2.2.15 The frequency and clinical significance of HER2 genetic heterogeneity, as defined by these ASCO/CAP guidelines, in breast carcinoma is unclear.
The aims of this study were to determine the frequency of HER2 genetic heterogeneity in breast carcinoma, and to identify clinicopathological features that characterise breast carcinomas with HER2 genetic heterogeneity.
Materials and methods
Patients and tissue samples
Breast carcinomas were classified according to the WHO classification of tumours of the breast and the female genital tract (2003). The first cohort comprised resection specimens of 218 consecutive carcinomas, the second cohort 176 core biopsies (51%) and 169 resection specimens (49%) of HER2 (2+) carcinomas, respectively.
Histology and immunohistochemistry
Haematoxylin–eosin staining was performed using standard histological techniques. For immunohistochemistry, sections were incubated with primary antibodies for oestrogen receptor (clone 6F11, dilution 1:100; Novacastra Laboratories Ltd, Newcastle, UK), progesterone receptor (clone 16, dilution 1:300; Novacastra), or Ki-67 (clone MIB-1, dilution 1:200; Dako, Baar, Switzerland) following antigen retrieval. Staining procedures were performed on Dako autostainer using the Dako REAL detection system with peroxidase/Diaminobenzidine (DAB)+. The cut-off for oestrogen receptor positivity and progesterone receptor positivity was 10% tumour cells with nuclear staining. The percentage of carcinoma cells showing nuclear Ki-67 immunoreactivity was determined in 500–1000 invasive carcinoma cells in randomly selected high-power fields (magnification ×400) at the periphery of the tumour. A cut-off level of 20% was used to dichotomise between carcinomas with high (>20%) or low (≤20%) Ki-67 labelling index.16
HER2 analysis was performed on invasive carcinoma cells by selecting areas with invasive carcinoma on haematoxylin–eosin stained sections. Carcinoma in situ was excluded from the analysis. HER2 protein status was assessed using the HercepTest (Dako). HER2 protein expression was scored as 0 (no staining), 1+ (weak and incomplete membrane staining), 2+ (strong, complete membrane staining in ≤30% of tumour cells or weak/moderate heterogeneous complete membrane staining in ≥10% of tumour cells), or 3+ (strong, complete, homogeneous membrane staining in >30% of tumour cells).1
HER2 fluorescence in-situ hybridisation (FISH) was performed using the PathVysion HER2 DNA probe (Abbott Molecular Inc., Downers Grove, Illinois, USA). At least 60 invasive carcinoma cells in at least three randomly selected areas of invasive carcinoma were scored for nuclear HER2 and chromosome CEP17 signals using a Leica DM6000B fluorescence microscope system (Leica Microsystems, Heerbrugg, Switzerland). The FISH result for each carcinoma cell was recorded in a table, which was used to determine the frequency of tumour cells with a HER2/CEP17 ratio greater than 2.2. If the average HER2/CEP17 ratio of 60 carcinoma cells was 1.8–2.2 (equivocal), another 40 invasive carcinoma cells were scored, and the final ratio of the tumour was calculated from the total of 100 cells. Using the ASCO/CAP recommendations,1 a tumour was classified as HER2 non-amplified (HER2/CEP17 ratio <1.8), equivocal (HER2/CEP17 ratio 1.8–2.2), or amplified (HER2/CEP17 ratio >2.2). A HER2/CEP17 ratio of greater than 2.2 to less than 4 was defined as low-grade amplification; a HER2/CEP17 ratio of 4 or greater as high-grade amplification. HER2 genetic heterogeneity was defined as more than 5% but less than 50% of invasive tumour cells with a ratio higher than 2.2.15 The percentage of carcinoma cells with HER2 gene cluster (defined as more than 16 signals per nucleus) was also determined. HER2 cluster-positive carcinomas were defined by the presence of HER2 gene clusters in 1% or more of tumour cells. For calculation of the HER2/CEP17 ratio, cells with HER2 clusters were considered to have 16 HER2 signals. Chromosome 17 copy number changes were classified based on the mean number of CEP17 signals per cell according to Ma et al.17
DNA image cytometry
Image cytometry was performed on Feulgen-stained imprints taken from resection specimens. The nuclear DNA content of tumour cells was measured using the ACAS cytometry analysis system (Ahrens, Bargteheide, Germany). For each imprint, 300–400 morphologically selected tumour cell nuclei were analysed. Diploid values were obtained from normal leucocytes. A mean ±2 SD nuclear DNA content was defined as 2c region. Based on the DNA histograms, carcinomas were classified into three ploidy categories: diploid (stem line at 1.8c–2.2c, and no cells exceeding 5c), tetraploid (stem line at 3.8c–4.2c, and <5% of the cells exceeding 5c), or aneuploid (stem line with one of more peaks outside the diploid or tetraploid region).18 Furthermore, the 5c exceeding rate and the stem line scatter index (SSI) were determined. The SSI is a measure of the percentage of tumour cells with non-modal DNA content values, or of the degree of scattering of DNA histograms.19 The SSI is defined as the sum of the percentages of cells with DNA content values in the S-phase region, the percentage of cells with DNA content values exceeding twice the modal value plus 1c (G2 exceeding rate) and the coefficient of variation of the respective tumour stem line (SSI = S phase + G2 exceeding rate + coefficient of variation). Based on the SSI, carcinomas were classified as genomically stable (SSI <8.8%) or unstable (SSI >8.8%).19 Genomically unstable tumours display a high cell-to-cell variability in DNA content, whereas genomically stable tumours have low variability.
Correlation analysis of nominal variables was performed using contingency table analysis, the χ2 test and two-sided Fisher's exact test (SPSS V.16 software). p Values less than 0.05 were regarded as significant. Tumours were grouped as ductal, lobular and others (tubular, medullary). Patients with neoadjuvant chemotherapy were not included in the association analysis between HER2 genetic heterogeneity and some pathological parameters (tumour size, pT stage, nodal status and histological grade).
HER2 genetic heterogeneity in breast carcinoma
We first analysed the frequency of HER2 genetic heterogeneity in a cohort of 218 consecutive invasive breast carcinomas. The overall frequency of HER2 genetic heterogeneity was 14.7%. The frequencies of HER2 genetic heterogeneity in carcinomas with HER2 immunohistochemical scores 0, 1+, 2+ and 3+ were 12.9% (11/85), 12.7% (8/63), 27.3% (9/33), and 10.8% (4/37), respectively. Among the 151 carcinomas lacking HER2 amplification, 29 (16.1%) showed HER2 genetic heterogeneity (figure 1A).
HER2 genetic heterogeneity was associated with both scattered carcinoma cells and small clusters of carcinoma cells with HER2 amplification (figure 1B). Therefore, the intratumoral variability of HER2 genetic heterogeneity was analysed in more detail in 32 carcinomas with HER2 genetic heterogeneity using resection specimens. The percentage of carcinoma cells with a HER2/CEP17 ratio greater than 2.2 was determined in three randomly selected areas of invasive breast carcinoma. We found comparable percentages of carcinoma cells with HER2 amplification in the majority of carcinomas (table 1).
We also performed HER2/CEP17 FISH on corresponding core biopsy and resection specimens of five carcinomas with HER2 genetic heterogeneity. HER2 genetic heterogeneity identified on core biopsy was confirmed on the corresponding resection specimen in all five cases (data not shown).
To correlate HER2 protein and HER2 gene findings at the cellular level, HER2 immunostaining and HER2/CEP17 FISH were performed on serial sections of five carcinoma core biopsies with HER2 2+ score. HER2 staining and HER2/CEP17 FISH results of carcinoma cells were compared in corresponding areas of the serial core biopsy sections. We focused on areas with heterogeneous HER2 staining of carcinoma cells. In these areas, carcinoma cells with and without HER2 gene amplification were present suggesting that HER2 staining heterogeneity of carcinoma cells was associated with HER2 genetic heterogeneity (data not shown).
Association between HER2 genetic heterogeneity and clinicopathological characteristics
To analyse the clinicopathological features of breast carcinomas with HER2 genetic heterogeneity, we extended the subgroup of HER2 (2+) carcinomas by adding 312 carcinomas with equivocal (2+) HER2 scores. The clinicopathological characteristics of the extended cohort are summarised in table 2. In the extended cohort, HER2 FISH analysis revealed HER2 genetic heterogeneity in 39.7% (137/345) and HER2 gene amplification in 8.7% (30/345) of table 3 HER2 (2+) carcinomas, respectively (table 2). The frequency of HER2 genetic heterogeneity in HER2 (2+) carcinomas was 35.2% for core biopsies and 44.4% for resection specimens, respectively. Among the 315 HER2 (2+) carcinomas lacking HER2 amplification, 130 (41.3%) showed HER2 genetic heterogeneity (figure 2). HER2 genetic heterogeneity was significantly associated with the chromosome 17 polysomy and the absence of HER2 gene clusters. For the various clinicopathological parameters analysed, we found an association between HER2 genetic heterogeneity and high histological tumour grade, but no other characteristics (table 4). No association between HER2 genetic heterogeneity and oestrogen/progesterone receptor expression was observed.
Association between HER2 genetic heterogeneity and DNA content
To study a potential association between HER2 genetic heterogeneity and genomic instability, image DNA cytometry was performed on a subset of 101 carcinomas from our series (table 5). DNA ploidy, SSI and 5c exceeding rate were used to assess genomic instability. HER2 genetic heterogeneity was associated with the presence of DNA aneuploidy and high 5c exceeding rate indicating an association between HER2 genetic heterogeneity and genomic instability (table 4).
We found that 14.7% of breast carcinomas display HER2 genetic heterogeneity, based on the recent ASCO/CAP definition.15 HER2 genetic heterogeneity was most frequent in breast carcinomas with an equivocal (2+) HER2 score, and was associated with a negative HER2 amplification status in 16% of all carcinomas and 42% of HER2 (2+) carcinomas, respectively. These results demonstrate that HER2 amplification negative carcinomas include a large subgroup of HER2 genetic heterogeneity positive carcinomas that harbour a significant subpopulation (>5%) of tumour cells with HER2 amplification, but do not qualify for trastuzumab treatment based on current recommendations due to a HER2/CEP17 ratio below 2.2.
In contrast to our findings, Tubbs et al9 found HER2 genetic heterogeneity in 5% of breast carcinomas using the same definition for HER2 genetic heterogeneity. Other studies on HER2 genetic heterogeneity in breast carcinoma defined HER2 genetic heterogeneity as regional variation in HER2 amplification status.2 6–8 20 Lewis et al2 performed HER2/CEP17 FISH on up to four different blocks of the same tumour in a small series of 21 (2+) HER2 breast carcinomas and observed substantial HER2/CEP17 ratio variability among blocks of the same tumour. In a subset of tumours, HER2/CEP17 ratio variability led to different classification of the same tumour as HER2 amplified or non-amplified depending on the tumour block analysed. In a similar approach, Brunelli et al7 analysed HER2 genetic heterogeneity in a series of 30 ductal carcinomas with HER2 (3+) immunostaining using whole tissue sections and tissue microarray cores derived from different areas of the same carcinoma. Of 20 tumours with high-grade HER2 amplification on whole tissue sections, 30% showed at least one tissue core with low-grade HER2 amplification. The usage of different definitions for HER2 genetic heterogeneity and of different approaches for its quantitative assessment has hampered the comparison of HER2 genetic heterogeneity frequencies between studies. The ASCO/CAP definition for HER2 genetic heterogeneity provides a basis for future studies to analyse the clinical significance of this HER2 FISH finding.
We also analysed a potential association between HER2 genetic heterogeneity and other tumour characteristics in the subgroup of carcinomas with an equivocal (2+) HER2 score. Among the various parameters analysed, we observed an association between HER2 genetic heterogeneity and the presence of HER2 gene clusters, chromosome 17 polysomy, histological tumour grade, DNA ploidy category and 5c exceeding rate. Despite these results, HER2 genetic heterogeneity was unpredictable based on the histological, immunohistochemical and DNA cytometric characteristics of a breast carcinoma. The identification of HER2 genetic heterogeneity therefore relies on single cell scoring and documentation of the HER2/CEP17 ratio in a sufficient number of carcinoma cells to determine the percentage of carcinoma cells with a HER2/CEP17 ratio greater than 2.2.
The prognostic and predictive significance of HER2 genetic heterogeneity is unknown. Due to the lack of follow-up data, we were not able to investigate the clinical significance of HER2 genetic heterogeneity with respect to treatment response and prognosis. Based on current criteria, most carcinomas with HER2 genetic heterogeneity do not qualify for trastuzumab treatment, because they show HER2/CEP17 ratios of 2.2 or less. However, patients with HER2 genetic heterogeneity may benefit from trastuzumab treatment, because these carcinomas harbour a significant proportion of carcinoma cells with HER2 gene amplification. Future studies are necessary to address this clinically important question.
HER2 genetic heterogeneity according to ASCO/CAP guidelines (2009) is frequent in breast carcinoma.
Carcinomas with HER2 genetic heterogeneity lack specific clinicopathological characteristics.
Detection of HER2 genetic heterogeneity depends on single cell scoring of the HER2/CEP17 ratio in a sufficient number of tumour cells.
Competing interests None.
Ethics approval Ethics approval was provided by Kantonale Ethikkommission St Gallen.
Provenance and peer review Not commissioned; externally peer reviewed.
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.