Assessment of ERBB2/HER2 Status in HER2-Equivocal Breast Cancers by FISH and 2013/2014 ASCO-CAP Guidelines.

Importance: The 2013/2014 American Society of Clinical Oncology and College of American Pathologists (ASCO-CAP) guidelines for HER2 testing by fluorescence in situ hybridization (FISH) designated an "equivocal" category (average HER2 copies per tumor cell ≥4-6 with HER2/CEP17 ratio <2.0) to be resolved as negative or positive by assessments with alternative control probes. Approximately 4% to 12% of all invasive breast cancers are characterized as HER2-equivocal based on FISH. Objective: To evaluate the following hypotheses: (1) genetic loci used as alternative controls are heterozygously deleted in a substantial proportion of breast cancers; (2) use of these loci for assessment of HER2 by FISH leads to false-positive assessments; and (3) these HER2 false-positive breast cancer patients have outcomes that do not differ from clinical outcomes for patients with HER2-negative breast cancer. Design, Setting, and Participants: We retrospectively assessed the use of chromosome 17 p-arm and q-arm alternative control genomic sites (TP53, D17S122, SMS, RARA, TOP2A), as recommended by the 2013/2014 ASCO-CAP guidelines for HER2 testing, in patients whose data were available through Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) and whose tissues were available through the Breast Cancer International Research Group clinical trials. We used data from an international cohort database of invasive breast cancers (1980 participants) and international clinical trial of adjuvant chemotherapy in invasive, node-positive breast cancer patients. Main Outcomes and Measures: The primary objectives were to (1) assess frequency of heterozygous deletions in chromosome 17 genomic sites used as FISH internal controls for evaluation of HER2 status among HER2-equivocal cancers; (2) characterize impact of using deleted sites for determination of HER2-to-internal-control-gene ratios; (3) assess HER2 protein expression in each subgroup; and (4) compare clinical outcomes for each subgroup.
Results: Of the 1980 patients in METABRIC,1915 patients were fully evaluated. In addition, 100 HER2-equivocal breast cancers by FISH and 100 comparator FISH-negative breast cancers from the BCIRG-005 trial were analyzed. Heterozygous deletions, particularly in specific p-arm sites, were common in both HER2-amplified and HER2-not-amplified breast cancers. Use of alternative control probes from these regions to assess HER2 by FISH in HER2-equivocal as well as HER2-not-amplified breast cancers resulted in high rates of false-positive ratios (HER2-to-alternative control ratio ≥2.0) owing to heterozygous deletions of control p-arm genomic sites used in ratio denominators. Misclassification of HER2 status was observed not only in breast cancers with ASCO-CAP equivocal status but also in breast cancers with an average of fewer than 4.0 HER2 copies per tumor cell when using alternative control probes. Conclusions and Relevance: The indiscriminate use of alternative control probes to calculate HER2 FISH ratios in HER2-equivocal breast cancers may lead to false-positive interpretations of HER2 status resulting from unrecognized heterozygous deletions in 1 or more of these alternative control genomic sites and incorrect HER2 ratio determinations.

Introduction

Amplification/overexpression of human epidermal growth factor receptor 2 gene (ERBB2, formerly HER2), is associated with shortened disease-free (DFS) and overall survival (OS) in patients whose breast cancers contain this alteration.[1,2,3] Because targeted therapies using anti-HER2 humanized monoclonal antibodies,[4,5,6,7] small molecule inhibitors of HER2 kinase,[8,9,10] and antibody-drug conjugates[11,12] effectively treat patients with HER2-positive breast cancer, accurate assessment of HER2 status is critically important for treatment selection. Since HER2 protein overexpression is a direct consequence of HER2 amplification, a variety of companion diagnostics are used to identify patients for targeted therapy. During the past decade, the American Society of Clinical Oncology (ASCO) and College of American Pathologists (CAP) have specified criteria for clinical assessment of HER2 amplification status.[13,14,15,16]

The most recent full ASCO-CAP guidelines for HER2 testing by in situ hybridization (ISH) changed the evaluation for HER2 amplification requiring formalized assessment of both average HER2 gene number per tumor cell and ratio of average HER2-to-internal control chromosome 17 centromere (CEP17) for assessment of HER2 status by fluorescence in situ hybridization (FISH).[13,14] This scoring algorithm identifies 5 different breast cancer FISH groupings using HER2 FISH ratio and average HER2 copy number per nucleus.[13,14,17,18] One of these 5 ASCO-CAP FISH groups, designated as ASCO-CAP FISH group 4,[17,18] is considered “HER2-equivocal” (neither amplified nor not-amplified).[13,14] According to the guidelines, this ambiguous status may be resolved with alternative controls to replace CEP17 for assessment of HER2 FISH ratios using genes other than CEP17. This approach was widely adopted by both commercial testing laboratories and academic centers. As described,[19,20,21,22] when any of these alternative control probes leads to a HER2-to-control ratio of 2.0 or more, the breast cancer is designated “ISH-positive.”

This approach to HER2-equivocal breast cancers contrasts with our experience using alternative controls.[23,24] Based on early studies of control genes on chromosome 17[2,3,25] and preliminary findings from the Breast Cancer International Research Group/Translational Research in Oncology (BCIRG/TRIO) central laboratories, we hypothesized that use of chromosome 17 p-arm controls to establish HER2 gene status of ISH-equivocal breast cancers could lead to false-positive classifications in a significant proportion of cases. In this study, we evaluate these hypotheses: (1) genetic loci used for alternative control probes are heterozygously deleted in a substantial proportion of human breast cancers, especially ISH-equivocal cancers; (2) use of these loci for FISH assessment of HER2 status leads to HER2-to-control ratios greater than or equal to 2.0 and, therefore, false-positive assessments of HER2 status; and (3) these HER2 false-positive breast cancers have outcomes that do not differ from behavior established for HER2-negative breast cancers.

Methods

Our study was conducted in 3 parts. The first involved analyses of Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) breast cancer data sets to determine relative frequency of deletions in chromosome regions corresponding to probes used as alternative controls on the chromosome 17 p-arm (LIS1, TP53, D17S122, RAI1, and SMS [Smith-Magenis syndrome] region) and q-arm (TOP2A and RARA). The second part involved reassessment of HER2 status in breast cancers from women accrued to BCIRG-005 (ClinicalTrials.gov Identifier NCT00312208) whose cancers had HER2-to-CEP17 FISH ratios of less than 2.0 and an average of 4.0 to 5.99 HER2 genes per tumor cell (ISH-equivocal) or cancers with FISH ratios of less than 2.0 and average HER2 gene copies less than 4.0 per tumor cell (ISH-negative) using alternative control probes to reassess HER2 status, according to the 2013/2014 ASCO-CAP guidelines.[19,20,22,26] Finally, we assessed clinical outcomes in these subgroups.

Analyses of METABRIC Data Set

Publicly available breast cancer genomic copy number data sets, including 1980 patients from the METABRIC[27] cohort (European Genome-phenome Archive, https://ega-archive.org/dacs/EGAC00001000484) have been profiled on the Affymetrix single-nucleotide polymorphism (SNP) 6.0 array, which includes more than 906 600 SNPs and more than 946 000 copy number probes. Corresponding clinical data were obtained, including tumor estrogen receptor α/progesterone receptor (ER/PR) status, HER2 evaluation by IHC (most samples) or FISH. Linear copy number values for HER2 and alternative probes were compared, as were their copy number states as inferred based on Genomic Identification of Significant Targets in Cancer (GISTIC, version 2.0.23) (https://www.ncbi.nlm.nih.gov/pubmed/21527027/) using default settings. In addition, allele-specific copy number analysis of tumors (ASCAT)[28] total allele counts (corrected for ploidy and purity) were obtained for the METABRIC cohort from Pereira et al.[29] To define the alternative probe regions of interest (LIS1, TP53, D17S122, RAI1, SMS, TOP2A, and RARA), each region was delineated by the position of the gene/genes contained (D17S122 contains CDRT7, PMP22, TEKT3, CDRT4, and TVP23C; SMS contains LLGL1, FLII, TOP3A, and SHMT1; RARA-TOP2A contains RARA and TOP2A). Individual cells were compared by selecting a representative gene from each region (TEKT3 for D17S122, TOP3A for SMS, and TOP2A for RARA-TOP2A).

Patients From the BCIRG-005 Trial

Because the second and third parts of our current study were reassessments of breast cancers designated as ISH-equivocal and ISH-negative by 2013/2014 ASCO-CAP FISH guidelines and because such cases were accrued to the BCIRG-005 trial, our focus was on this trial[30,31] (Figure 1) (Supplement). This randomized trial of concurrent TAC or sequential AC-T adjuvant anthracycline-containing chemotherapy demonstrated the regimens were equally efficacious but differed in levels of toxic effects.[30,31] This trial was previously approved by human investigations committees of each institution that accrued patients. Written informed consent was obtained from each patient. The central laboratory obtained institutional review board approval for the characterization of HER2 status of tumor samples from each patient.

Figure 1.

Participant Flow Diagram and Specimen Accountability

Breast cancers from patients were evaluated in 1 of 2 central laboratories (laboratory) as either human epidermal growth factor receptor 2 gene HER2-not-amplified or HER2-amplified for eligibility to 1 of 3 concurrently conducted clinical trials (BCIRG-005, BCIRG-006, and BCIRG-007). One of the trials, BCIRG-005, required patients whose breast cancers were HER2-not-amplified and the other 2 trials, BCIRG-006 and BCIRG-007, required patients whose breast cancers were HER2-amplified, as determined with fluorescent in situ hybridization (FISH). Although 10 948 patients were screened in the Breast Cancer International Research Group central laboratories for trial accrual, complete HER2 FISH assay results were available from 10 468 patients for a variety of reasons, including lack of invasive carcinoma in samples submitted, tissue sections that detached from slides during processing, and FISH assay failure owing to lack of probe hybridization. BCIRG-005 randomized patients with HER2-not-amplified breast cancers to sequential (arm 1) or concurrent (arm 2) anthracycline, cyclophosphamide, and docetaxel chemotherapy. BCRIG-006 randomized patients with HER2-amplified breast cancers to standard anthracycline-containing chemotherapy (arm 1, AC-T) alone, AC-T with trastuzumab (arm 2, ACTH) or a nonanthracycline chemotherapy regimen with trastuzumab (arm 3, TCH). The breast cancers from these trials were subsequently pooled according to the ASCO-CAP guidelines for HER2 testing by FISH as recommended into 5 in situ hybridization (ISH) groups, identified in the lower portion of the figure (ASCO-CAP ISH group 5, ASCO-CAP ISH group 4, ASCO-CAP ISH group 3, ASCO-CAP ISH group 2, and ASCO-CAP ISH group 1) and reanalyzed for correlations with HER2 protein expression and clinical outcomes.[17] Since ASCO-CAP ISH group 4 is composed exclusively of HER2-equivocal breast cancers by FISH, the focus of the current investigation, group 4 served as the source of all HER2-equivocal breast cancers (N = 100) characterized in this study by FISH with alternative control probes. As a comparator group, ASCO-CAP ISH group 5, breast cancers were selected for similar analyses by FISH using the same alternative control probes (N = 100). AC-T indicates anthracycline, cyclophosphamide, and docetaxel; ACTH, anthracycline, cyclophosphamide, docetaxel, and trastuzumab; TAC, taxotere, docetaxel, and cyclophosphamide; TCH, docetaxel, carboplatin, and trastuzumab.

We included all patients whose breast cancers were both ISH-equivocal by standard FISH (HER2/CEP17) (183 participants) (Figure 1) and whose breast cancers were successfully hybridized with all alternative controls used in this study to replace CEP17 (TP53, SMS, D17S122, RARA, and TOP2A) (100 participants) (Figure 1) to recalculate the HER2 FISH ratio, as recommended.[13,14,19,21,22] This portion of our study is based on 100 ISH-equivocal and 100 ISH-negative cases from BCIRG-005 successfully reanalyzed with 5 alternative controls by FISH. For comparison with ISH-equivocal, we selected 100 HER2-not-amplified breast cancers that, according to the guidelines, are ISH-negative (our ASCO-CAP ISH group 5 breast cancers[17,18]) beginning with those samples that had an average HER2 gene copy number just less than 4.0 and selected every case in sequential order beginning with 3.99 copies per tumor cell until we had a comparison group of 100 cases that also had successful hybridization with all alternative control probes used in the study (100 participants) (Figure 1).

Laboratory Methods

Tissue sections[6,17,32] or tissue microarrays[33] previously analyzed for HER2 gene amplification status were rehybridized and reanalyzed using alternative control chromosome 17 probes (TP53, D17S122, RAI1, SMS, TOP2A, and RARA) by FISH. The primary invasive breast carcinomas of these patients were previously analyzed for HER2 protein expression using immunohistochemical analysis.[17,18,32] Among these, 80 ISH-equivocal and 100 ISH-negative cases had HER2 immunohistochemical results available for comparison (Supplement).

HER2 FISH Assays

HER2 FISH assays were performed using the PathVysion assay (Abbott-Molecular, Inc) (Supplement).[17,18,32,34,35] We characterized HER2-equivocal and HER2-negative breast cancers by FISH with alternative controls (TP53, D17S122, SMS, TOP2A, RARA) according to current full (2013/2014) ASCO-CAP guidelines[13,14] (Figure 1) (eFigure 1 in the Supplement).

HER2 Protein Expression by Immunohistochemistry

The HercepTest (Dako) as well as a laboratory-developed HER2 10H8-IHC assay were used to evaluate HER2 protein expression in the BCIRG-005 trial (Supplement).[17,18,32,35]

Statistical Methods

A 2-tailed Fisher exact test was used to compare the frequencies of deletions or gains among q and p chromosome arms using the fisher.test function in the R Statistical Programming language (version, 3.4.4; R Foundation, Inc); 95% confidence intervals are reported. A Mann-Whitney test was used for comparing copy number values among different regions at chromosome 17p using the wilcox.test function in R. The Mann-Whitney was selected because copy number data do not follow a gaussian distribution. The Spearman correlation among copy number values at chromosome 17p was calculated with cor function in R statistical software. Log-rank tests were used to compare BCIRG-005 DFS and OS between different subgroups.

Results

To evaluate heterozygous deletions of chromosome 17 p-arm and q-arm genomic sites, we used publicly available SNP array data from the METABRIC data set, and confirmed those findings using the same genetic loci by FISH in BCIRG-005 breast cancer specimens.

Evaluation of Heterozygous Deletions in Alternative Chromosome 17 Genomic Regions Based on METABRIC

We identified genomic regions of chromosome 17 reported as alternative control sites for assessment of HER2 FISH ratios.[19,20,21] These sites were assessed in METABRIC for copy number gains and losses relative to ERBB2/HER2 gains and losses (Table 1) (Figure 2) (eFigure 1 in the Supplement).

Table 1.

Chromosome 17 Regional Gene Copy Gains and Losses Based on GISTIC Among Alternative Control Genomic Sites Compared With ERBB2/HER2 Gene Copy Gains and Losses in the METABRIC Cohort Including 1915 Participants

Alternative Control (region)	ERBB2/HER2 Gene Copy No. Status, No. (%)				Total
	HER2 Loss	HER2 Normal	HER2 Gain	HER2 Amp
LIS1
Gain	9 (2.5)	25 (2.7)	61 (18.7)	13 (4.5)	108 (5.6)
Normal	46 (12.7)	622 (66.5)	64 (19.6)	87 (29.9)	819 (42.8)
Loss	308 (84.8)	288 (30.8)	201 (61.7)	191 (65.6)	988 (51.6)
Total	363 (100)	935 (100)	326 (100)	291 (100)	1915 (100)
TP53
Gain	3 (0.8)	15 (1.6)	60 (18.4)	4 (1.4)	82 (4.3)
Normal	41 (11.3)	624 (66.7)	59 (18.1)	81 (27.8)	805 (42.0)
Loss	319 (87.9)	296 (31.7)	207 (63.5)	206 (70.8)	1028 (53.7)
Total	363 (100)	935 (100)	326 (100)	291 (100)	1915 (100)
D17S122 (TEKT3)
Gain	10 (2.7)	11 (1.2)	55 (16.9)	8 (2.7)	84 (4.4)
Normal	50 (13.8)	643 (68.8)	67 (20.5)	84 (28.9)	844 (44.1)
Loss	303 (83.5)	281 (30.0)	204 (62.6)	199 (68.4)	987 (51.5)
Total	363 (100)	935 (100)	326 (100)	291 (100)	1915 (100)
RAI1
Gain	7 (1.9)	39 (4.2)	78 (23.9)	35 (12.0)	159 (8.3)
Normal	57 (15.7)	640 (68.4)	64 (19.9)	89 (30.6)	851 (44.4)
Loss	299 (82.4)	256 (27.4)	183 (56.1)	167 (57.4)	905 (47.3)
Total	363 (100)	935 (100)	326 (100)	291 (100)	1915 (100)
SMS (TOP3A)
Gain	12 (3.3)	42 (4.5)	79 (24.2)	41 (14.1)	174 (9.1)
Normal	54 (14.9)	647 (69.2)	71 (21.8)	94 (32.3)	866 (45.2)
Loss	297 (81.8)	246 (26.3)	176 (54.0)	156 (53.6)	875 (45.7)
Total:	363 (100)	935 (100)	326 (100)	291 (100)	1915 (100)
TOP2A/RARA (TOP2A)
Gain	2 (0.5)	19 (2.0)	284 (87.1)	117 (40.2)	422 (22.0)
Normal	22 (6.1)	899 (96.1)	36 (11.0)	77 (26.5)	1034 (54.0)
Loss	339 (93.4)	17 (1.8)	6 (1.8)	97 (33.3)	459 (24.0)
Total	363 (100)	935 (100)	326 (100)	291 (100)	1915 (100)

Abbreviations: D17S122, the 17p-arm genomic locus which is duplicated in Charcot-Marie-Tooth disease; GISTIC, genomic identification of significant targets in cancer; HER2, human epidermal growth factor receptor 2 gene; METABRIC, Molecular Taxonomy of Breast Cancer International Consortium; RARA, retinoic acid receptor-alpha gene; SMS, Smith-Magenis Syndrome locus; TOP2A, topoisomerase-II-alpha gene.

Genomic Identification of Significant Targets in Cancer, a tool to identify genes targeted by somatic copy-number alterations; note that a subset of cases was not evaluable by GISTIC, thus 1915 of 1980 cases are reported here.

Figure 2.

Relative Copy Number of ERBB2 and Genomic Sites Used as Alternate Controls for Assessment of HER2 Status by FISH (METABRIC Cohort, SNP Chip Data for 1980 Patients)

A, Schematic illustration of the positions of alternative control genomic sites (p-arm: LIS1, TP53, D17S122, RAI1, SMS; and q-arm: RARA, TOP2A) relative to ERBB2/HER2 on chromosome 17. The location of chromosome 17 centromere is highlighted in red. B, Relative copy number of ERBB2/HER2 and genomic sites used as alternative controls for assessment of HER2 Status by FISH (METABRIC SNP array data for 1980 patients). Samples were ordered by their HER2 CN value and plotted alongside the copy number profiles for alternative probes. A linear regression line was fit for each gene (probe). Top bar shows annotations for samples based on IHC, ASCAT, and GISTIC.

Homozygous deletion of any alternative control site was distinctly unusual; however, heterozygous deletions were relatively common, especially on the p-arm of chromosome 17. Alternative control sites on the p-arm showed a higher rate of heterozygous deletion than q-arm alternative control sites. A scatterplot of ERBB2/HER2 linear copy numbers compared with alternate control sites (LIS1, TP53, D17S122, RAI1, SMS, TOP2A, and RARA) demonstrated relatively high rates of gene copy loss using the alternative control genomic sites located on the p-arm (Table 1) and a significantly lower rate of loss among q-arm markers (P < .001; OR, 0.29; 95% CI, 0.25-0.34; Fisher exact test for the frequency of TOP2A [q-arm] was lower than the frequency of TETK3 [p-arm] in the METABRIC cohort) (Table 1) (Figure 2B) (eFigure 1 in the Supplement). Deletions in LIS1, TP53, D17S122, RAI1, and SMS were observed in 288 (31%), 296 (32%), 281 (30%), 256 (27%), and 246 (26%) of 935 cases, respectively, with concurrent HER2 normal copy numbers (Table 1). In contrast, only 17 (1.8%) of 935 METABRIC samples showed loss of RARA and TOP2A (q-arm) with a normal HER2 copy number state (Table 1).

In METABRIC, 32% of samples demonstrated either a copy number gain or amplification of ERBB2/HER2 with 15% of samples showing high-level gains, considered amplifications. We considered those breast cancers with HER2 copy number “gain” but not “high gain” (amplification) to be most likely representative of ISH-equivocal cancers with increased HER2 (4-6 copies per tumor cell), but not HER2 amplification. Among those with HER2 copy number gain but not amplification, in the p-arm LIS1, TP53, D17S122, RAI1, and SMS regional losses occurred in 201 (62%), 207 (64%), 204 (63%), 183 (56%), and 176 (54%) of 326 cases, respectively (Table 1) (Figure 2B) (eFigure 1 in the Supplement). Regional loss on the q-arm (TOP2A and RARA genes) was significantly less frequent (1.84%) (P < .001; OR, 0.01 [95% CI, 0.003-0.025]; Fisher exact test for TETK3 in p-arm and TOP2A in q-arm).

The rate of p-arm alternative control region deletions was even higher among breast cancers with high-level gain or amplification than in the HER2 copy gain cancers (Table 1) (Figure 2B) (eFigure 1 in the Supplement). Therefore, more than half of the samples with HER2 gain or amplification showed a loss of p-arm regions.

Comparison of HER2 IHC values revealed that IHC 3+ cases in METABRIC exhibited deletions of the alternative control probe loci (eFigure 1B in the Supplement). A similar deletion pattern in the control probes were observed when using the total count of alleles as estimated by ASCAT (correcting by ploidy and purity) (eFigure 1C in the Supplement). Differences were noted between alternative probes located in 17p-arm such that genes located at 17p13-12 (LIS1, TP53, and D17S122) were more frequently deleted than genes at 17p11 (DSS and RAI1) (P < .001; θ = −0.0261; 95% CI, −0.0351 to 0.0171; Mann-Whitney test for TP53 vs DSS in METABRIC). This difference was magnified in amplified samples (P < .001; θ = −0.0533; 95% CI, −0.077 to −0.0299; Mann-Whitney test in METABRIC amplified samples). The negative correlation between HER2 copy-number state (loss, neutral, gain) and alternative probe values was also higher in the 17p13-12 region (LIS1, TP53, and D17S122) compared with 17p11 (RAI1 and SMS) (Spearman correlation ρ = −0.142 [P < .001], −0.158 [P < .001], −0.172 [P < .001], vs −0.078 [P < .001], −0.05 [P = .03] in METABRIC).

Chromosome 17 Alternative Control Regions Demonstrate Heterozygous Deletion by FISH

Among 100 ASCO-CAP FISH group 4 (HER2-equivocal) breast cancers, we identified heterozygous deletions (Table 2) (eTable 1 and eFigure 2 in the Supplement) in 65 for the SMS locus, 46 for D17S122, 43 for TP53, 8 for TOP2A, none for RARA, and none for HER2 (eTable 1 in the Supplement). The p-arm alternative control loci had a significantly higher rate of deletions relative to q-arm loci (eTable 1 in the Supplement)(P < .001) with SMS the most frequently lost, followed by D17S122 and TP53. The frequency of q-arm heterozygous deletions among HER2-equivocal cancers was low, with only TOP2A reaching 8%. Using the observed average HER2 copy number (varied from 4 to 5.99) per tumor cell divided on a case-by-case basis by the corresponding observed average alternative control copy numbers demonstrated that ratios greater than or equal to 2.0 were observed for 61 cases using SMS, 65 using TP53, and 30 using D17S122 as the denominators in the HER2-to-control ratio calculations. Use of RARA or TOP2A resulted in fewer cases with HER2-to-control ratios greater than 2.0 (7% and 25%) (eTable 1 in the Supplement).

Table 2.

Criteria for Evaluation of Heterozygous Deletions at Alternative Control Genomic Sites on Chromosome 17 by FISH

Chromosome 17 Arm	Gene / Locus	Ratio	Interpretation	Ratio	Interpretation
p-arm	SMS	<0.75a	SMS with heterozygous deletion relative to RARA	>1.25b	RARA with heterozygous deletion relative to SMS
q-arm	RARA
p-arm	TP53	<0.75c	TP53 with heterozygous deletion relative to TOP2A	>1.25d	TOP2A with heterozygous deletion relative to TP53
q-arm	TOP2A
p-arm	D17S122	<0.75e	D17S122 with heterozygous deletion relative to HER2	>1.25f	HER2 with heterozygous deletion relative to D17S122
q-arm	HER2

Abbreviations: D17S122, the 17p-arm genomic locus which is duplicated in Charcot-Marie-Tooth disease; FISH, fluorescence in situ hybridization; HER2, human epidermal growth factor receptor 2 gene.; RARA, retinoic acid receptor-alpha gene; SMS, Smith-Magenis syndrome locus; TOP2A, topoisomerase-II-alpha gene; TP53, tumor protein 53 tumor suppressor gene.

Subjective assessment of signals also requires the observation that SMS signals are loosely paired with RARA signals plus additional individual, scattered RARA signals. These paired and individual signals are scattered randomly within tumor cell nuclei, not clustered or aggregated as observed with amplified genes, illustrated in eFigure 2 in the Supplement.

Subjective assessment of signals also requires the observation that most RARA signals are loosely paired with SMS signals plus additional individual, scattered SMS signals. These paired and individual signals are scattered randomly within tumor cell nuclei, not clustered or aggregated as observed with amplified genes.

Subjective assessment of signals also requires observation of TP53 signals loosely paired with TOP2A signals plus excess additional TOP2A signals. These paired and individual signals are scattered randomly within tumor cell nuclei, not clustered or aggregated as observed with amplified genes, illustrated in eFigure 2 in the Supplement.

Subjective assessment of signals also requires the observation that most TOP2A signals are loosely paired with TP53 signals. These paired and individual signals are scattered randomly within tumor cell nuclei, not clustered or aggregated as observed with amplified genes.

Subjective assessment of signals also requires the observation that D17S122 signals are loosely paired with HER2 signals plus excess unpaired HER2 signals. These paired and individual signals are scattered randomly within tumor cell nuclei, not clustered or aggregated as observed with amplified genes, illustrated in eFigure 2 in the Supplement.

Subjective assessment of signals also requires the observation that most HER2 signals are loosely paired with D17S122 signals. These paired and individual signals are scattered randomly within tumor cell nuclei, not clustered or aggregated as observed with amplified genes.

Among 100 ASCO-CAP FISH group 5 (HER2-not-amplified or ISH-negative) breast cancers, we identified heterozygous deletions in 35 cases for D17S122, 30 for SMS, 3 for TOP2A, 1 for RARA, and none for TP53 (eTable 1 in the Supplement). The p-arm SMS and D17S122 alternative control loci had significantly higher deletion rates than other loci (eTable 1 in the Supplement) (P < .001). The frequency of p-arm heterozygous deletions among HER2-negative breast cancers was lower than rates observed among HER2-equivocal breast cancers. Nevertheless, use of these alternative controls in ASCO-CAP FISH group 5 breast cancers would lead to significant up-grading of HER2 status to ISH-positive (eTable 1 in the Supplement). Using the observed average HER2 copy number, which varied from 3.99 to 3.25 per tumor cell divided on a case-by-case basis by the correspondingly observed average alternative control copy number, we demonstrated ratios greater than 2.0 for 37 cases using SMS, 12 with RARA, 11 with D17S122, 2 with TP53, and 1 with TOP2A as the denominators in the HER2-to-control ratio calculations. Similar to HER2-equivocal breast cancers, the use of heterozygously deleted alternative control loci was associated with HER2-to-control ratios equal to or greater than 2.0, consequently, false-positive ratios through status upgrading.

HER2 Protein Expression and Alternative Control Subgroups

The ASCO-CAP FISH group 4 (HER2-equivocal) and ASCO-CAP FISH group 5 (HER2-negative) breast cancers included in this study were associated with low levels of HER2 protein expression by immunohistochemical analysis (eTable 2 in the Supplement). Separation of group 4 and group 5 according to FISH ratios using the various alternative controls did not alter this association for any subgroup by ratio greater than or equal to 2.0, determined with any alternative control probe (eTable 2 in the Supplement), consistent with the interpretation that these alternative control ratios of 2.0 or more do not identify a subgroup of cancers which should be considered for upgrading of HER2 status.

Clinical Outcomes by Alternative Control Subgroups

Within the ISH-equivocal breast cancer patients whose cancers were upgraded to ISH-positive using alternative control probes, there was no subgroup of patients who had a significantly worse disease-free or overall survival compared with either the overall ISH-equivocal or ISH-negative patients or those with a HER2-to-alternative control probe ratio of less than 2.0 (Figure 3) (eTable 3 and eFigure 3 in the Supplement). Similar observations were made among the patients whose cancers were ISH-negative breast cancer and were upgraded to ISH-positive by the use of alternative control probes.

Figure 3.

Comparison of Clinical Outcomes for ASCO-CAP Group 4 (HER2-Equivocal) and ASCO-CAP Group 5 (HER2-Negative) Patients With Breast Cancer

A, Disease-free survival of ASCO-CAP FISH group 4 (HER2-equivocal) compared with ASCO-CAP FISH group 5 (HER2-not-amplified). There was no significant difference in disease-free survival for the 100 patients with ASCO-CAP FISH group 4 (HER2-equivocal) breast cancers compared with the 100 with ASCO-CAP FISH group 5 (HER2-not-amplified) breast cancer. B, Overall survival of ASCO-CAP FISH group 4 (HER2-equivocal) compared with ASCO-CAP FISH group 5 (HER2-not-amplified) patients with breast cancer. There was no significant difference in overall survival for the 100 patients with ASCO-CAP FISH group 4 (HER2-equivocal) breast cancers compared with the patients with 100 ASCO-CAP FISH group 5 (HER2-not-amplified) breast cancer. C, ASCO-CAP FISH group 4 (HER2-equivocal): OS for Alternative Control Probe D17S122 (HER2/D17S122) ratios ≥2.0 vs <2.0. Among women with HER2-equivocal breast cancers with an D17S122 alternative control probe ratio ≥2.0 does not identify a subgroup with a worse overall survival. D, ASCO-CAP FISH group 4 (HER2-equivocal): OS for alternative control probe SMS (HER2/SMS) ratios ≥2.0 vs <2.0. Among women with HER2-equivocal breast cancers, those who had a SMS alternative control probe ratio ≥2.0 appear to have a slightly better overall survival than those whose breast cancers had a SMS alternative control probe ratio <2.0; however, this difference was not significant.

Discussion

ISH-equivocal, as defined by the 2013/2014 ASCO-CAP guidelines[13,14] and the recent 2018 update,[36] represent approximately 4% to 12% of all breast cancers, or 7000 to 21 000 patients annually.[17,18,20,37,38,39,40,41] These cases present patients with breast cancer, oncologists, and pathologists some of their most frequent clinical challenges, because questions remain regarding how HER2 status should be resolved and confusion about HER2-targeted therapies for these patients. Several studies reported the use of alternative controls among patients whose cancers had ISH-equivocal status.[19,20,21,22,26] In these studies, if any HER2-to-alternative probe ratio was 2.0 or more, the breast cancer was upgraded to ISH-positive.[19,20,21,26]

The potential for internal comparator controls to impact assessment of HER2 status was not addressed by previous studies but has concerned us since we began characterizing this alteration.[2,3,25] In 2 early investigations of HER2 amplification using the same 345 breast cancers, we used myeloperoxidase gene (MPO) as the internal control (HER2-to-MPO ratio)[3] whereas another group used TP53 (HER-to-TP53 ratio)[42] for assessment of HER2 status. The studies came to different conclusions about both amplification frequency in the cohort (27% vs 33%) and associations with clinical outcomes (significantly associated with DFS and OS[3] vs neither[42]) using the same shared HER2 gene data by Southern hybridization.[3,42] The choice of control gene to calculate ratios was the only difference.

The 2013/2014 ASCO-CAP guidelines did not provide data to support use of chromosome 17 alternative probes and the recent 2018 Focused Update did not provide data justifying discontinuation of this approach to resolve the status of ISH-equivocal cancers. Here we used METABRIC data as well as cancers from the BCIRG-005 trial to demonstrate heterozygous deletions among various p-arm and q-arm genomic sites previously used as alternative controls. Our findings indicate heterozygous deletions, particularly at p-arm genomic sites, are relatively common and are not restricted to breast cancers with ISH-equivocal status. In addition, among cancers upgraded by HER2-to-alternative-control ratios of 2.0 or greater there was no significant association with either HER2 protein overexpression or worse clinical outcomes. Our conclusion, that ISH-equivocal breast cancers are HER2-not-amplified is also supported by our previous data,[17,18] which demonstrate that ASCO-CAP FISH group 4 (HER2-equivocal) breast cancers lack HER2 protein overexpression[17,18] and have patient outcomes[17] that are not significantly different from ASCO-CAP FISH group 5 (HER2-negative or HER2-not-amplified) breast cancer patients. These findings are consistent with the perspective that HER2-equivocal cancers upgraded to HER2-positive are the result of heterozygously deleted alternative control probes. These findings do not mean use of alternative controls for assessment of amplification has no value, only that selection of internal controls should be done carefully with the possibility of heterozygous deletions clearly considered. In our laboratory we have established criteria for the use of such alternative control probes (Table 2) (eFigure 2 in the Supplement).

Because p-arm genomic sites frequently have heterozygous loss, using a HER2-to-p-arm alternative control ratio in ISH-equivocal breast cancers can lead to overestimation of the HER2-amplification status. There is a lack of evidence to support the view that HER2 upgraded, ISH-equivocal breast cancers have a clinical disease similar to HER2-amplified breast cancers. We and others[20] show that women with ISH-equivocal cancers upgraded to HER2-positive by p-arm alternative controls do not have significantly worse outcomes than those who are not upgraded. ISH-equivocal cancer patients show outcomes similar to those with ISH-negative disease with no difference in DFS or OS that warrants upgrading to ISH-positive status. We predict lack of responsiveness to HER2-targeted therapies among these patients whose cancers lack HER2-amplification, consistent with previous trials of lapatinib[9] and, more recently, the NSABP-B47 trial of adjuvant trastuzumab in 3270 patients with breast cancers having weakly positive HER2 expression by IHC (1+) or moderately positive HER2 expression by IHC (2+) but HER2-not-amplified (HER2-negative) by FISH.[43] Therefore, use of HER2-directed therapies in a population falsely classified as HER2-positive is expected to produce inferior clinical and pharmacoeconomic outcomes.

Limitations

The study has some noteworthy limitations. Although a large number of breast cancer cases were available from the METABRIC database for analyses, HER2 gene copy number estimates were based on SNP array data, not HER2 gene copy numbers determined by FISH. In contrast, HER2 gene copy number data were available from the BCIRG-005 trial; however, the number of patients in this trial with long-term clinical follow-up who had breast cancers with HER2 FISH-equivocal (ASCO-CAP FISH group 4) status was limited.

Conclusions

Assessment of HER2 gene status in HER2-equivocal breast cancers is important and clinically relevant. Although approximately half of such HER2-equivocal cases have been upgraded to HER2-positive by FISH using alternative control probes, deletions in these alternative control genomic sites on chromosome 17 can lead to false-positive HER2-to-control ratios (greater than or equal to 2.0). We show the genomic sites used to assess these HER2-to-internal control ratios, especially those on the p-arm of chromosome 17, have frequent heterozygous deletions. These deletions, when used for assessment of HER2-equivocal (ASCO-CAP FISH group 4) breast cancers lead to frequent false-positive HER2-to-internal control ratios greater than 2.0. HER2-equivocal breast cancers with these false-positive ratios do not have HER2 protein overexpression and these patients do not have clinical outcomes that differ from either other patients with HER2-equivocal breast cancers not upgraded to positive or from HER2-negative disease. Our findings indicate that HER2 FISH-equivocal breast cancers are HER2-not-amplified.