The human epidermal growth factor receptor type 2 gene (HER2), also known as ERBB2, is amplified and overexpressed in approximately 20% of unselected breast cancers.1,2 The gene encodes a membrane receptor protein expressed at relatively low levels on lateral and basal surfaces of virtually all normal epithelial cells3 including normal breast epithelium (Figures 1A, 1B, 1C). Amplification of this gene leads to high levels of protein expression, referred to as overexpression (Figures 1D, 1E), which is associated with shorter disease-free survival (DFS) and overall survival (OS) in patients with breast cancer.2,4 HER2 overexpression has now been effectively targeted for therapeutic intervention using humanized monoclonal antibodies to the extracellular domain (trastuzumab,5-7 pertuzumab,8 and T-DM19 ) and small molecular inhibitors to the intracellular kinase domain (lapatinib,10 neratinib11). These HER2-targeted agents have substantially improved both DFS and OS in both the adjuvant and metastatic settings for HER2-positive breast cancer patients.
However, to achieve these benefits, correct recognition of which breast cancers have the alteration and which do not is critically important for appropriate patient selection. The expense and potential adverse effects of these therapeutics should preclude their use in patients who are not likely to benefit. Equally problematic is the possibility of withholding an effective therapy from someone who may potentially benefit. Given this, the accuracy of the testing methodology as well as the scoring criteria used for reporting a cancer as HER2-positive or HER2-negative is of paramount importance.
The 2 most common modalities used for testing breast cancer specimens for the presence or absence of the HER2 alteration are immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH). Based on the variabilities in accuracy for HER2 testing that have already been previously reported,12-14 particularly using IHC,12 the American Society of Clinical Oncology (ASCO) and the College of American Pathologists (CAP) convened a panel to standardize approaches to HER2 testing.13 Subsequently, the ASCO-CAP guidelines panel for HER2 testing was reconvened to modify the initial recommendations.15,16 While we have already reported on many of the contentious aspects of HER2 testing by IHC,12 here we summarize some of the issues related to the current ASCO-CAP guidelines for HER2 testing by FISH.15,16
The primary purposes of the initial ASCO-CAP guidelines for HER2 testing were multifold and aimed at improving accuracy of HER2 testing through standardization of preanalytic tissue-processing procedures (eg, anoxia tissue time, fixative type [formalin] and duration [6-48 hours], and methods of tissue processing), analytical procedures, and postanalytical procedures when testing was performed in a CAP-accredited laboratory environment. This included guidelines for interpretation with algorithms for scoring based on a ratio of the average HER2 gene copy number-toaverage CEP17 copy number per tumor cell. The 3 categories were defined as negative with a ratio of <1.8, equivocal when the ratio is 1.8 to 2.2, and positive when the ratio is >2.2. Prior to that, the US FDA had approved clear criteria for defining HER2-positive disease as cancers with a FISH ratio ≥2.0 and HER2-negative cancers as those with a ratio <2.0; the criteria included a method for resolution of cases when ratios are within 10% of the 2.0 cut-off—ie, 1.8 to 2.2—without further testing.4,17 Despite this, the ASCO-CAP guidelines identified a new “equivocal” category and recommended additional assessment for resolution. Of note, these “equivocal” cases represented only 2% of all breast cancers.12 In the subsequent 2013/2014 ASCO-CAP guidelines, the designation of “equivocal” was retained; however, the definition of what constituted a “HER2-equivocal” breast cancer was modified and the number of cases increased to between 4% and 12%.18-25
According to the new and current ASCO-CAP guidelines for HER2 testing, in situ hybridization (ISH) assay results, including FISH, should now be divided into 5 groups based on a formalized assessment of both average HER2 gene copy number and HER2 FISH ratio
At the time these guidelines were published, no clinical or demographic data were available using this classification schema, and basic information such as the prevalence of each FISH group in the general breast cancer population was not known. Moreover, data regarding whether these new ASCO-CAP groups correlated with HER2 protein expression or, more importantly, clinical outcomes, were also not available. To better address these questions, we conducted 2 retrospective studies of breast cancer specimens previously characterized for HER2 status in our laboratories: one set was from a cohort of an academic consultation practice,21 and the other set was from breast cancers screened for entry to Breast Cancer International Research Group (BCIRG)/Translational Research In Oncology (TRIO) clinical trials.20
Eligibility requirements for inclusion in the cohort study and BCIRG/TRIO trials are described in detail elsewhere.20,21 In brief, all consecutive, primary, invasive breast carcinomas submitted to the clinical consultation practice of one of us (MFP) from April 1999 until September 2015 that had both HER2 gene amplification status determined by FISH and HER2 protein level determined by IHC were eligible for inclusion in the cohort study of HER2 status by FISH (n = 7526).21 The study of BCIRG/TRIO clinical trials breast carcinoma samples included primary invasive breast carcinomas from 10,468 patients who were screened for enrollment in either of 2 central laboratories (MFP and GS) for HER2 gene amplification status determined by FISH as an enrollment criterion for eligibility to 3 different trials: BCIRG-005,26 BCIRG-006,7 and BCIRG-007.27
Those patients whose breast cancers were HER2-amplified were eligible for BCIRG-006 or -007, whereas those whose breast cancers were not HER2-amplified were eligible for BCIRG-005. The BCIRG-006 trial (n = 3222) is a randomized, 3-arm study of adjuvant chemotherapy with or without trastuzumab in patients with HER2-amplified stage I to III breast cancer who were accrued between April 2001 and March 2004.7 Therapy in the control arm was adjuvant anthracycline, cyclophosphamide, and docetaxel (AC-T) with or without hormonal therapy depending on tumor estrogen receptor and progesterone receptor status at site investigator discretion. Therapy in the 2 experimental arms involved trastuzumab with patients randomly assigned to either standard AC-T adjuvant chemotherapy or nonanthracycline chemotherapy with docetaxel and a platinum salt—again, with or without hormonal therapy depending on tumor estrogen receptor and progesterone receptor status. This trial demonstrated significant improvement in DFS for both trastuzumab-containing treatment arms compared with control AC-T adjuvant chemotherapy alone. BCIRG-005 clinical trial (n = 3298) is a randomized study of concurrent (taxotere, adriamycin, and cyclophosphamide) or sequential (AC-T) adjuvant anthracycline-containing chemotherapy in patients with HER2-not-amplified, stage II and III breast cancer who were accrued from August 2000 to February 2003. This trial demonstrated that sequential and combination regimens that incorporated 3 drugs were equally efficacious but differed significantly in toxicity profile.26,28 The BCIRG-007 trial (n = 263), a randomized phase III trial of docetaxel and trastuzumab compared with docetaxel, carboplatin, and trastuzumab in women with HER2-amplified metastatic breast cancer, was screened for HER2 status by FISH concurrently with BCIRG-005 and BCIRG-006. Data for HER2 gene amplification and expression are included in the study20; however, outcome information is not included as this trial had no control, nontrastuzumab treatment arm.27
We performed analyses of prevalence by FISH group, association with HER2 expression by IHC, and clinical outcomes. We compared the original FDA-approved criteria for HER2 gene amplification with current ASCO-CAP guidelines, assessed the number of cases in each guidelines group, and determined whether or not the new ASCO-CAP FISH testing criteria used to define each of the 5 HER2 FISH groups are correlated with those characteristics already known to be associated with HER2 gene amplification, such as HER2 protein overexpression, poorer clinical outcomes (DFS/OS) in the absence of HER2-targeted therapy, and significant improvement in DFS and OS when such patients are treated with HER2-targeted therapy.
As expected, in both study cohorts the majority of breast cancers had a HER2 FISH ratio <2.0 with an average HER2 gene copy number <4.0 (group 5) and the second largest group were those with a HER2 ratio ≥2.0, with an average HER2 gene copy number ≥4.0 (group 1) (Table 1). These are the breast cancers traditionally considered “HER2-negative” and “HER2-positive,” respectively, by FISH assay. Groups 2 and 3 each represented less than 1% of the study population and ASCO-CAP Group 4, the “equivocal” breast cancers, represented 4% to 5% of each study population.20,21
Because only ASCO-CAP groups 1 and 5, designated respectively as ISH-positive and ISH-negative, corresponded to the interpretations we assigned in our consultation practice and in our central laboratory for entry to BCIRG/TRIO trials, we also wanted to evaluate association with HER2 protein expression levels by IHC to determine agreement between ASCO-CAP FISH guidance and protein expression category by IHC. Contrary to the ASCO-CAP designations, we found that ASCO-CAP groups 2 and 4 were significantly associated with low HER2 protein expression, not overexpression. ASCO-CAP group 3 appeared to be composed of 2 different subgroups: a larger subgroup (our group 3N) HER2- negative with low expression, and a smaller subgroup (our group 3A) HER2-amplified that had protein overexpression (Tables 2 and Table 3).
Given the fact that the findings with IHC in 3 new ASCO-CAP FISH groups (groups 2, 3, and 4) appeared to contradict the assigned designation of the ASCO-CAP guidelines for HER2 testing, we evaluated known clinical outcomes in the BCIRG/ TRIO clinical trials. These trials have long-term clinical follow-up data available7,26,28 that allow for a determination of whether or not the new HER2 guidelines for FISH/ISH testing are predictive of known clinical outcomes and, therefore, clinically useful.
As described above, we found that breast cancers in ASCO-CAP FISH group 1 had HER2 protein IHC levels that correlated with HER2 overexpression (Tables 2A and 2B). Those patients whose cancers were in this group were accrued to the BCIRG-006 trial of adjuvant trastuzumab, and those randomly assigned to a trastuzumab-plus-chemotherapy treatment arm experienced significant improvements in DFS (hazard ratio [HR], 0.71; 95% CI, 0.60-0.83; P <.0001) and OS (HR, 0.69; 95% CI, 0.55-0.85; P <.0006) compared with similar patients assigned to standard chemotherapy alone.20
It is worth noting that approximately 20% of ASCO-CAP FISH group 1 (HER2-amplified) breast cancers were IHC-negative (IHC 0/1+) (Table 2A). We have previously used a variety of approaches to confirm that IHC-negative, HER2-amplified breast cancers are predominantly the result of tissue processing artifacts that impact IHC, but not FISH.29 For example, we assessed HER2 gene amplification status, HER2 mRNA expression, and HER2 protein expression by western immunoblot; assessed HER2 protein expression by frozen section IHC in frozen breast cancer samples; and compared HER2 status with HER2 protein expression by IHC in the corresponding formalin-fixed, paraffin-embedded (FFPE) breast cancers. We found that a significant number of HER2-amplified, overexpressed breast cancers were IHC 0 in the corresponding FFPE tissue sections.2 Although the percentage of IHC false-negative breast cancers varies depending on tissue processing and IHC assay methods used, this is a recurring observation, not only in our own studies,2,12,30-32 but in the published literature comparing IHC with FISH.33-36 As expected, these patients with IHC false-negative, FISH-amplified HER2 status currently do not receive trastuzumab or other anti-HER2 therapy and have been shown to experience a statistically significantly worse distant DFS compared with similarly treated patients with IHC-negative, FISH-not-amplified breast cancers.32
Although few breast cancers are in ASCO-CAP FISH group 2, these cases have HER2 IHC scores indicating low HER2 protein expression in our clinical consultation practice and BCIRG trials cohort (Tables 2A and 2B). Nevertheless, in BCIRG trials, the majority of these patients were accrued to BCIRG-006 due to our use of the FDA-approved FISH criteria for HER2 gene amplification (ratio >2.0 without regard for the average HER2 gene copy number per tumor cell). When applying the new ASCO-CAP FISH guidelines to patients (n = 46) randomized to receive adjuvant trastuzumab in BCIRG-006, no significant improvement in either DFS (HR, 1.1; 95% CI, 0.31-3.89; P = .89) or OS (HR, 3.15; 95% CI, 0.35-28.63; P = .28) was observed when compared with patients randomized to receive standard anthracycline-cyclophosphamide followed by taxane chemotherapy alone.
Interestingly, and as expected, the small number of patients in ASCO-CAP group 3 (Table 1) appears to be not a single group of “ISH-positive” breast cancers as specified by the ASCO-CAP guidelines, but a group with at least 2 subgroups, which we have referred to as subgroup 3N (not amplified) and subgroup 3A (amplified). In our consultation practice as well as in the BCIRG clinical trials cohort, the larger 3N subgroup of breast cancers (Table 3) have relatively modest increases in average HER2 gene copy number per tumor cell of 6.8 and 7.4, respectively. As described above, these breast cancers are associated with low HER2 protein expression, while the members of the other, even-less-numerous 3A subgroup have substantially higher average HER2 gene copy numbers per tumor cell of 12.3 and 16.3, respectively. In our pathology consultation practice as well as in the BCIRG trials cohorts, there is a clear association with HER2 protein overexpression20,21 only in the group 3A breast cancers (Table 3). Based on this latter association, we expect the ASCO-CAP FISH 3A subgroup to be associated with worse OS in the absence of HER2-targeted therapy, and, conversely, improved DFS and OS with such treatment.
The ASCO-CAP FISH group 4 (HER2 FISH ratio <2.0; average HER2 gene copy number per tumor cell ≥4.0 to <6.0) breast cancers, currently labeled “ISH-equivocal,” are associated with low HER2 protein expression, and, in the absence of trastuzumab treatment, have clinical outcomes that are not significantly worse than those of other patients whose breast cancers lack HER2 gene amplification and have low HER2 protein expression (IHC 0/1+).20 When outcomes of these “ISH-equivocal” patients, our ASCO-CAP FISH group 4, are compared with outcomes of ASCO-CAP FISH group 5 patients, who are those considered HER2-not-amplified or “ISH-negative,” there is no significant difference in either DFS (HR, 0.92; 95% CI, 0.679- 1.224; P = .58) or OS (HR, 0.88; 95% CI, 0.609-1.267; P = .49). Similar observations were made by Sneige et al in a study of 3630 patients analyzed by FISH for HER2 status.24
Confirmation of this “HER2-not-amplified” status in ASCO-CAP FISH group 4 or “ISH-equivocal” breast cancers can be supported by using alternative control probes in addition to the chromosome 17 centromere control routinely used to calculate the HER2 FISH ratio.24,37 However, this approach also has some important shortcomings. The most important pitfall is the lack of recognition that these alternative control genomic regions, especially those on the p-arm of chromosome 17, may show heterozygous deletion, which leads to an increased HER2- to-control probe ratio >2.0 based exclusively on heterozygous deletion of the control genomic site rather than true gene amplification38 (Joshi H, Press MF; unpublished data). An independent study from other investigators has shown that those breast cancer cases converted from “ISH-equivocal” to “ISH-positive” based on the use of p-arm alternative controls for conversion of a HER2 FISH ratio from <2.0 to >2.0 demonstrate DFS and OS rates similar to those of patients whose cancers continued to have a HER2 FISH ratio <2.0 after evaluation with these same alternative controls24 (Figure 3).
HER2 gene amplification status is critically important to select the most appropriate patients with breast cancer for HER2-targeted therapies, such as trastuzumab, pertuzumab, T-DM1, neratinib or lapatinib treatments. The current ASCO-CAP guidelines for HER2 testing are widely accepted by pathologists and clinicians for assessment of this status, yet there are inconsistencies with the available data for at least 5% of patients, based on either correlative expression data or actual clinical outcome data or both. These data suggest that these patients should be assigned differently than currently recommended by the most recent guidelines. We have reviewed these inconsistencies and suggested appropriate remedies based on currently available data.
This work was supported in part by grants from the Breast Cancer Research Foundation, Tower Cancer Research Foundation (Jessica M. Berman Senior Investigator Award), and a gift from Dr. Richard Blach.
Author affiliations: Michael F. Press, MD, PhD, and Yanling Ma, MD, are with the Departments of Pathology and Laboratory Medicine, and Susan Groshen, PhD, is with the Department of Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California. Dr Press is also with the Norris Comprehensive Cancer Center, Los Angeles, California. Guido Sauter, MD, PhD, is with the Department of Pathology, University of Hamburg, Hamburg Germany; and Dennis J. Slamon, MD, PhD, is with the Department of Medicine, Geffen School of Medicine at UCLA, Los Angeles, California.
Corresponding author: Michael F. Press, MD, PhD, Norris Comprehensive Cancer Center, NOR 5409, 1441 Eastlake Ave, Los Angeles, CA 90033; Tel: (323) 865-0563; E-mail: firstname.lastname@example.org.
Financial disclosures: One of the co-authors (MFP) has disclosed research funding to his institution for work conducted in his laboratory from Cepheid, Inc, Eli Lilly & Company, Novartis Pharmaceuticals, and Hoffmann La-Roche, Inc. He has served as a consultant for these companies as well as for Karyopharm Therapeutics, Puma Biotechnology, Halozyme Therapeutics, ADC Therapeutics, and Biocartis. The other co-authors have no financial disclosures or potential conflicts of interest.
Support: This work was supported by grants from the Breast Cancer Research Foundation, Tower Cancer Research Foundation (Jessica M. Berman Senior Investigator Award), and a gift from Dr Richard Blach.
To the Editor,
Recently, Press et al raised serious concerns regarding the accuracy and usefulness of the various definitions of HER2-positive disease that were added in the American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) guidelines and published in 2013.1,2
In the article, the authors state that “correct recognition of which breast cancers have the alteration and which do not is critically important” and also state that the accuracy of the testing methodology is of “paramount importance.” The many reports since the 2013 ASCO/CAP guidelines, including those cited by Press et al, which prove or disprove concordance between HER2 gene amplification or mRNA expression and HER2 protein expression, correlation among different methods of assessing HER2 gene amplification, and the trials that report risk of recurrence as a surrogate for HER2-positivity, are important, but not of “paramount” or “critical” importance. What is critically important and of paramount importance is whether there are clinical data showing a potential benefit from anti-HER2 therapy for patients defined as having HER2-positive disease based solely on a definition added in 2013.
There is a key distinction between proving potential efficacy of a targeted therapy once a biomarker has been identified versus proving that a different metric can be used to predict whether a tumor likely harbors that biomarker. Concordance and correlative studies may be important after first proving efficacy of anti-HER2 therapies for patients whose tumors are defined as positive based solely on 1 of the defining criteria of HER2-positivity added by ASCO/CAP in 2013. This is clearly underscored by the focus of recent landmark studies showing that a myriad of very different biomarkers (eg, PD-L1 expression, deficient mismatch repair [dMMR] gene expression/microsatellite instability–high , mutational burden) predict checkpoint inhibitor therapy efficacy in particular settings as the key first studies, rather than the focus being toward investigations of concordance between, for example, dMMR somatic mutations and normal MMR protein expression.3 Both are important, but first identifying whether the particular biomarker(s) predicts efficacy is of paramount or critical necessity in order to endorse the therapy.
When the authors of the ASCO/CAP 2013 guideline expanded the definition of HER2-positivity to include tumors that would not previously have been defined as HER2-positive, they stated that they did so to ensure that “the right patient receives the right treatment” and they recommended clinical trials to study anti-HER2 therapy efficacy for patients meeting only 1 of the new definitions.2
Precision medicine relies fundamentally on clinical trials that verify the efficacy of a targeted therapy for tumors harboring an identified biomarker. The now reported National Cancer Institute–sponsored trial, “Chemotherapy with or without trastuzumab after surgery in treating women with invasive breast cancer,” does aim to determine whether patients with “HER2-low” tumors benefit from adjuvant trastuzumab (NCT01275677). Different clinical trials are also needed to investigate whether patients with tumors that would be deemed HER2-positive based solely on 1 of the added ASCO/CAP 2013 definitions of HER2-positivity might benefit from anti-HER2 therapy.
Steven Sorscher, MD
Professor of Medicine
Wake Forest School of Medicine
The authors respond
We appreciate Dr Sorscher’s interest in our article, “Controversies in HER2 Oncogene Testing: What Constitutes a True Positive Result in Patients with Breast Cancer?”,1 which summarizes our findings on HER2 status according to the fluorescent in situ hybridization (FISH) groups recommended by the 2013-2014 ASCO/CAP guidelines for HER2 testing.2,3 We compared our conclusions, determined by FISH testing in 2 large breast cancer populations (N = 10,468 and N = 7526),4,5 with the assigned designations of the 2013-2014 ASCO/CAP guidelines, created in the absence of any previous data supporting their FISH algorithm.
We addressed 2 issues: 1) Does a breast cancer FISH group have the genetic alteration of HER2 amplification? and 2) Do patients whose cancers have particular alterations (either HER2 amplification or a lack thereof) have worse clinical outcomes in the absence of HER2-targeted therapy or do they respond to HER2-targeted therapies?
HER2 gene amplification is directly associated with HER2 protein overexpression. In the 3 decades since our initial demonstration of this association,6 there has been no other mechanism for HER2 overexpression that has been demonstrated or validated. Breast cancers originally thought to be HER2–not-amplified by solid matrix blotting, but to have HER2 overexpression,6 have since been demonstrated to be HER2-amplified by FISH.1,7,8 We have used these known biological associations of HER2 amplification with overexpression as well as correlations with more aggressive disease behavior to re-establish the expected associations (or lack thereof) for each ASCO/CAP FISH guideline category. These analyses of our data clearly contradict the designations assigned by the ASCO/CAP guidelines committee for 3 of the 5 HER2 FISH groups (Groups 2, 3, and 4). In the absence of a significant association with HER2 protein overexpression, one should doubt whether the ASCO/CAP FISH group 2, group 3N or group 4 breast cancers are anything other than HER2–not-amplified, a conclusion that contradicts the ASCO/CAP designations of “ISH-positive,” “ISH-positive,” and “ISH-equivocal.”
In the pre-trastuzumab era, numerous studies demonstrated that patients whose breast cancers have HER2 amplification/overexpression experience significantly worse disease-free survival (DFS) and overall survival (OS) than similar patients whose cancers lack this alteration.4,6,9 We demonstrated that ASCO/CAP FISH group 4 (“ISH-equivocal”) breast cancer patients, in aggregate, lack HER2 protein overexpression4,5 and have DFS and OS that do not differ from ASCO/CAP group 5 (ISH-negative) breast cancer patients.4 Others report that use of alternative FISH probes as internal controls in HER2 “ISH-equivocal” breast cancers identify 2 subgroups having HER2-to-alternative control ratios <2.0 and >2.0.10-13 However, the DFS and OS outcomes of these 2 subgroups of women are not significantly different from one another,1,2 consistent with our unpublished data showing the subgroup with HER2/alternative control ratios >2.0 are related to heterozygous deletions of alternative control genomic sites used in the ratio denominator.
We have provided clinical outcome data for ASCO/CAP group 2 (HER2-to-CEP17 ratio >2.0 with an average HER2 gene copy number per tumor cell <4.0) breast cancer patients randomized in the BCIRG-006 trial of adjuvant trastuzumab. As expected with a prevalence rate of <1% of breast cancers with this FISH profile,1,4,5 few patients had this profile, but follow-up was relatively long (>10 years) and demonstrated no significant benefit from chemotherapy plus trastuzumab compared with chemotherapy alone in terms of DFS or OS.4
Dr Sorscher offers no data to contradict our findings or our classifications. He does, however, want additional outcome data from randomized clinical trials of HER2-targeted therapies for patients in each of these ASCO/CAP FISH groups, much like the NSABP-B47 trial (NCT01275677), which demonstrated no benefit from trastuzumab among 3270 randomized women whose breast cancers lacked HER2 amplification and had low expression of HER2 with IHC 1+ or IHC 2+ immunohistochemical staining.14 Such clinical trials require patients to make commitments with their lives to these studies and are very expensive. We think clinicians are well-advised to treat their patients according to the preponderance of current data until such specialized clinical trials demonstrate our findings to be in error.
Michael F. Press, MD, PhD
University of Southern California
Norris Comprehensive Cancer Center
Los Angeles, CA
Dennis J. Slamon, MD, PhD
University of California Los Angeles
Jonsson Comprehensive Cancer Center
Los Angeles, CA