Assessment of the Impact of Alternative Fixatives on HER2 Detection in Breast Cancer and Gastric Cancer Tumor Specimens

Abstract

The type of fixative used for preserving tumor specimens can significantly impact the performance of the immunohistochemistry and in situ hybridization assays used for assessing human epidermal growth factor receptor 2 (HER2) status. This study reports the prevalence of the use of alternative fixatives other than the guideline-recommended 10% neutral buffered formalin (NBF) during HER2 testing in a real-world setting. The effects of alternative fixatives [20% NBF and 10% unbuffered formalin (UBF) fixatives] on HER2 testing of breast cancer (BC) and gastric cancer (GC) cell lines and tissues are also assessed. Overall, 117,636 tumor samples received at a central laboratory from >8000 clinical trial sites across 60 countries were reviewed to determine the prevalence of alternative fixative usage. To investigate the impact of alternative fixatives, 27 cell lines (21 BC and 6 GC) and 76 tumor tissue samples (50 BC and 26 GC) were fixed in 10% NBF, 20% NBF, or 10% UBF, and evaluated for HER2 status by immunohistochemistry and in situ hybridization. Real-world data showed that 9195 (7.8%) tumor samples were preserved using an alternative fixative. In cell lines, overall percentage agreement, negative percentage agreement, and positive percentage agreement among the 3 fixatives were 100%. In tumor tissues, the agreement among 10% NBF, 20% NBF, and 10% UBF ranged between 94.7% and 96.6% for negative percentage agreement and 90.9% for overall percentage agreement compared with a range of 58.3% to 66.7% for positive percentage agreement. These results suggest that alternative fixatives may have the potential to convert HER2 status in tissues from positive to negative.

An increasing number of biomarkers have been identified, for which targeted cancer drugs are being discovered and exploited therapeutically to improve patient outcomes.^1,2 Human epidermal growth factor receptor 2 (HER2; also referred to as ERBB2), for example, has served as a predictive biomarker in breast and gastric/gastroesophageal cancers.^3,4 Therapies directed against HER2, including monoclonal antibody agents (trastuzumab, pertuzumab, and margetuximab), antibody-drug conjugates (ado-trastuzumab emtansine and trastuzumab deruxtecan), and small-molecule tyrosine kinase inhibitors (lapatinib, neratinib, and tucatinib) in combination therapy, have improved clinical outcome of patients with HER2-positive breast^5,6 and gastric cancers (GCs).⁷

Testing of HER2 status to determine eligibility for anti-HER2 therapy is routinely assessed by using a combination of immunohistochemistry (IHC, to evaluate HER2 protein expression levels) and in situ hybridization (ISH, to assess HER2 gene amplification).^8,9 HER2 status through IHC and/or ISH testing is primarily performed on formalin-fixed paraffin-embedded (FFPE) tissue blocks from patient biopsy or resected tumor tissue from the primary or metastatic site.⁹

Given the therapeutic importance of accurately assessing HER2 status in patients with breast cancer (BC)and GC, the American Society of Clinical Oncology (ASCO) and the College of American Pathologists (CAP) developed and recommended guidelines for determining HER2 status; first for BC in 2007,¹⁰ with revisions in 2013¹¹ and 2018,⁸ and later for GC in 2017.⁹ The ASCO/CAP guidelines also discuss the optimal preanalytical and analytical requirements for the performance and interpretation of HER2 testing using IHC and ISH.

Since the publication of the 2007 ASCO/CAP HER2 testing guideline, indirect evidence suggests that the performance of laboratories that conduct HER2 testing in the United States and Europe is improving,¹¹ but potential technical issues remain that may hinder correct assessment of HER2-expression levels and could lead to inter or intralaboratory variation. Furthermore, studies have found significant interlaboratory variation in validation practices and have revealed that many clinical testing laboratories do not follow consistent guideline procedures when validating IHC assays.^12,13 Following the ASCO/CAP 2007 guidelines, some laboratories did not minimize the frequency and effect of prolonged fixation (by either reducing the number of cases fixed longer than 48 h or retesting such specimens by fluorescence in situ hybridization). Other laboratories continued to assess HER2 by IHC using the scoring guidelines included in the test kit rather than the ASCO/CAP scoring criteria.^10,12

Although several studies report on the effects of some preanalytical variables such as cold ischemic time (delay to formalin fixation)^14–19 and prolonged formalin fixation for biomarker assay development,^14,18–21 few have addressed the technical aspects of fixative choice to define best practices for clinical studies of HER2 status.^22,23

The buffer’s qualities (ie, concentration and pH levels) in the formalin solution used for fixing tissue specimens have been shown to affect the quality of immunostaining.²⁴ Although published recommendations and guidelines for IHC advise fixation in 10% neutral buffered formalin (NBF),^8,25 phosphate-buffered saline is cited as the preferred buffer.²⁵ The ASCO/CAP guidelines and the Ventana PATHWAY anti-HER2/neu (4B5) antibody and INFORM HER2 dual ISH DNA Probe Cocktail package inserts specify a 10% NBF for optimal tissue fixation for accurate IHC and ISH HER2 testing.^8,9,26,27 Despite the guidelines’ recommendations, alternative NBF fixatives [eg, 20% NBF and 10% unbuffered formalin (UBF)] are still used.^28,29

This study is intended to assess the prevalence of alternative fixatives used in the real-world setting and to compare the effects of the guideline-recommended fixative (10% NBF) versus alternative fixatives (20% NBF or 10% UBF) on HER2 status of breast and GC tissues and cell lines.

MATERIALS AND METHODS

Real-world Central Laboratory Data Review

Laboratory testing data of 117,636 tumor samples received at LabCorp Central Laboratory Services (Indianapolis, USA; Geneva, Switzerland; Shanghai, China; and Singapore) between September 2, 2015, and August 12, 2021, and evaluated in 162 assays from 8401 clinical trials sites and across 60 countries were reviewed to determine the prevalence of alternative fixative usage.

Cell Line Samples

Twenty-seven cell lines (21 BC and 6 GC) were procured from Yale Medical School (Yale Pathology Tissue Services, New Haven, CT) and subcultured according to the American Type Culture Collection recommended conditions until cell numbers were sufficient (1×10⁸ cells/mL) to obtain 3 equal aliquots of ~3×10⁷ cells/mL each. Aliquots were centrifuged to create cell pellets that were fixed for ~24 hours with a minimum of 6 hours but no more than 72 hours in 10% NBF, 20% NBF, or 10% UBF. Fixed pellets were routinely processed and embedded into paraffin.

Tissue Samples

Thirty-two resected tissue samples (20 BC and 12 GC) were acquired from BioIVT, and 44 tissue samples (30 BC, 14 GC) were acquired from National Cancer Center Hospital East and Tokai University Hospital, both in Japan. The tumor tissue samples were ~10 mm³ in size according to ASCO/CAP criteria for BC and GC biopsy or surgical specimens.^8,9,11 Out of 76 total samples, 5 samples were not evaluable due to the lack of presence of tumor cells, lack of positive internal controls, and/or insufficient specimen size.

The fresh tumor tissue was divided into 3 equal parts, with the central portion of the tissue placed into 10% NBF and the bookend portions of the tissue were placed in 20% NBF or 10% UBF. The divided tissue was fixed for ~24 hours (≥6 h and ≤ 72 h). Fixed tissues were routinely processed and embedded in paraffin.

HER2 Detection

An IHC assay was performed for the detection of HER2 protein expression in FFPE tumor tissue specimens and cell lines. The FFPE tissue and cell line blocks were sectioned to a thickness of 4 to 5 μm and mounted on positively charged glass slides. Sections were dehydrated and heat-fixed in a laboratory oven at 60 °C for 1 hour to increase adherence of the tissue section to the glass slide. Sections were incubated with the PATHWAY anti-HER2/neu (4B5) rabbit monoclonal antibody (Ventana Medical Systems, Inc.), followed by incubation with a secondary antibody-horseradish peroxidase conjugate (ultraView Universal DAB Detection Kit, Ventana Medical Systems, Inc.) using the BenchMark ULTRA IHC/ISH System version 12.2 (Ventana Medical Systems, Inc.) per user instructions.

A dual ISH assay was performed to detect HER2 DNA amplification in FFPE tumor specimens that were IHC 2+. Sections were processed using the INFORM HER2 Dual ISH DNA Probe Cocktail, the ultraView SISH DNP Detection Kit, the ultraView Red ISH DIG Detection Kit (Ventana Medical Systems, Inc.), and accessory reagents using the automated BenchMark ULTRA IHC/ISH System version 12.2 (Ventana Medical Systems, Inc.) per user instructions.

HER2 Status Evaluation and Validation

Evaluation of HER2 protein expression and HER2 gene amplification was performed by board-certified pathologists using conventional light microscopy, with all 3 tissue specimens (10% NBF, 20% NBF, and 10% UBF) associated with a single case assessed by 1 pathologist. Scoring and interpretation were done according to the ASCO/CAP HER2 2013 BC and 2017 GC scoring guidelines (HER2-negative: IHC 0, 1+, 2+/ISH nonamplified; HER2-positive: IHC 2+/ISH amplified, and IHC 3+).^9,11 Concordance was assessed based on the category of HER2-positive or HER2-negative. All 95% CIs are based on the Person-Clopper exact method.

The validation set used to provide documentation of assay performance characteristics and to ensure the validity of the data produced comprised 103 (71 BC and 32 GC) cell line and tissue samples that were prospectively collected and processed using 10% NBF, 20% NBF, or 10% UBF. Sample distribution was based on HER2 status data from cell lines and tissue samples fixed in 10% NBF (Table 1).

TABLE 1 - Sample Distribution Based on HER2 Data From Cell Lines and Tissue Samples Fixed in 10% NBF

	Indication	IHC 0	IHC 1+	IHC 2+ ISH nonamplified	IHC 2+ ISH amplified	IHC 3+	Total no. samples
Cell lines	BC	7	2	2	0	10	21
	GC	5	—	—	—	1	6
Tissue samples	BC	14	9	14	5	3	45
	GC	12	3	7	1	3	26
HER2 status	—	Negative			Positive		98*
No. samples	—	75			23		—

^*Total number of samples (cell lines and tissue) is 103; HER2 status in 5 samples fixed in 10% NBF was not evaluable due to lack of presence of tumor cells, lack of positive internal controls, and/or insufficient size.

BC indicates breast cancer; GC, gastric cancer; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization; NBF, neutral buffered formalin.

RESULTS

Of the cancer tissue samples, 117,636 from 299 clinical studies conducted at 8401 clinical trial sites across 60 countries between 2015 and 2021 were analyzed. A review of the assays used for samples collected showed that 7.8% of these were fixed using a fixative other than 10% NBF (Table 2). Data on the individual rates of 20% NBF and 10% UBF use are not available.

TABLE 2 - Real-World Review of Tissue Fixation Protocols Conducted Between February 9, 2015, and December 8, 2021

Review parameters	N
Sponsors	88
Clinical studies	299
Patients	52,797
Clinical sites	8401
Countries involved	60
Assays	162
Tissue samples submitted	117,636
Tissue samples fixed in non-10% NBF	9195

NBF indicates neutral buffered formalin.

In a HER2 status concordance analysis using HER2-positive and HER2-negative BC and GC cell lines, the overall percentage agreement (OPA), negative percentage agreement (NPA), and positive percentage agreement (PPA) between the 10% NBF guideline-recommended fixative and the 20% NBF and 10% UBF alternative fixatives were 100% (Table 3). In a HER2 status concordance analysis using HER2-positive and HER2-negative BC and GC tissue samples, the agreement between the 10% NBF and 20% NBF was 94.7% (95% CI: 85.4%-98.9%) for NPA, 66.7% (95% CI: 29.9%-92.5%) for PPA, and 90.9% (95% CI: 81.3%-96.6%) for OPA. HER2 status concordance agreement between cancer tissues fixed with 10% NBF and 10% UBF was 96.6% (95% CI: 88.1%-99.6%) for NPA, 58.3% (95% CI: 27.7%-84.8%) for PPA, and 90.09% (95% CI: 80.5%-95.9%) for OPA (Table 4). Overall, PPA between 10% NBF and the alternative fixatives was not achieved in 3 out of 9 cancer tissues fixed with 20% NBF and in 5 out of 12 cancer tissues fixed with 10% UBF (Table 4). A total of 13 cases were discordant, which occurred in 10 tissue samples [9 cases occurred in 6 BC samples (with some overlap in the same samples) and 4 cases occurred in 4 distinct GC samples]. A shift from HER2-positive to HER2-negative status occurred in 8/13 cases [5 cases occurred in BC (n = 3 samples) and 3 cases occurred in GC (n = 3 samples)]. These discordant cases were due to a change in the ISH category [ISH+ to ISH−; 37.5% (3/8 cases, n = 3 BC)], a change from IHC 3+ to IHC 2+/ISH− [25.0% (2/8 cases, n = 2 GC)], and a change from IHC 2+ to IHC 1+ [37.5% (3/8 cases, n = 2 BC; n = 1 GC)]. NPA between 10% NBF and the alternative fixatives was not achieved in 3 out of 57 cancer tissues fixed with 20% NBF and in 2 out of 58 cancer tissues fixed with 10% UBF (Table 4). A shift from HER2-negative to HER2-positive status occurred in 5/13 discordant cases in both BC and GC tissue samples [4 cases occurred in BC (n = 3 BC samples) and 1 case occurred in GC (n = 1 GC sample)] and was most frequently due to a change from IHC 1+ to IHC 2+/ISH+ [60.0% (3/5 cases, n = 2 BC; n = 1 GC)]. One switch was due to a change from IHC 2+/ISH− to IHC 3+, and another switch was due to a change from IHC 2+/ISH− to IHC 2+/ISH+ [20.0% (1/5) each, both cases occurred in BC samples].

TABLE 3 - Concordance Rates of 20% NBF or 10% UBF to 10% NBF in Breast and GC Cell Line Samples

	10% NBF
HER2 status/concordance	HER2-negative (IHC 0, 1+, 2+/ISH−)*; n	HER2-positive (IHC 2+/ISH+, IHC 3+)*; n	OPA
20% NBF
HER2-negative	16	0	—
HER2-positive	0	11	—
Concordance with 10% NBF	NPA 100% (16/16) 95% CI (79.4%-100%)	PPA 100% (11/11) 95% CI (71.5%-100%)	OPA 100% (27/27) 95% CI (87.2%-100%)
10% UBF
HER2-negative	16	0	—
HER2-positive	0	11	—
Concordance with 10% NBF	NPA 100% (16/16) 95% CI (79.4%-100%)	PPA 100% (11/11) 95% CI (71.5%-100%)	OPA 100% (27/27) 95% CI (87.2%-100%)

^*HER2 status per American Society of Clinical Oncology/College of American Pathologists definition.

GC indicates gastric cancer; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization; NBF, neutral buffered formalin; NPA, negative percentage agreement; OPA, overall percentage agreement; PPA, positive percentage agreement; UBF, unbuffered formalin.

TABLE 4 - Concordance Rates of 20% NBF or 10% UBF to 10% NBF in Breast and Gastric Tumor Tissue Samples

	10% NBF
HER2 status/concordance	HER2-negative (IHC 0, 1+, 2+/ISH−)*; n	HER2-positive (IHC 2+/ISH+, IHC 3+)*; n	OPA
20% NBF
HER2-negative	54	3	—
HER2-positive	3	6	—
Not evaluable†	2	3	—
Concordance with 10% NBF	NPA: 94.7% (54/57) 95% CI (85.4%-98.9%)	PPA: 66.7% (6/9) 95% CI (29.9%-92.5%)	OPA: 90.9% (60/66) 95% CI (81.3%-96.6%)
10% UBF
HER2-negative	56	5	—
HER2-positive	2	7	—
Not evaluable†	1	0	—
Concordance with 10% NBF	NPA: 96.6% (56/58) 95% CI (88.1%-99.6%)	PPA: 58.3% (7/12) 95% CI (27.7%-84.8%)	OPA: 90.0% (63/70) 95% CI (80.5%-95.9%)

^*HER2 status per American Society of Clinical Oncology/College of American Pathologists definition.

^†Not evaluable samples were not included in the analysis.

HER2 indicates human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization; NBF, neutral buffered formalin; NPA, negative percentage agreement; OPA, overall percentage agreement; PPA, positive percentage agreement; UBF, unbuffered formalin.

Representative HER2 IHC staining of HER2 IHC 3+ BC specimens fixed with 10% NBF, 20% NBF, and 10% UBF are shown (Fig. 1). HER2 IHC staining was not changed between tissues fixed with 10% UBF and 20% NBF compared with the tissues fixed with 10% NBF in these representative HER2 IHC staining images.

FIGURE 1: Representative images of HER2 IHC staining of HER2 IHC 3+ breast cancer tissue fixed with 10% NBF, 20% NBF, and 10% UBF (×40 objective). HER2 indicates human epidermal growth factor receptor 2; IHC, immunohistochemistry; NBF, neutral buffered formalin; UBF, unbuffered formalin.

DISCUSSION

Molecular pathology has become fundamental to inform tumor diagnosis, prognosis, and therapeutic decisions. HER2 testing in clinical practice has been driven by the development of HER2-targeted therapy.⁴ Accurate, robust, and reliable assays for patient selection add considerable value to targeted cancer therapies such as anti-HER2 therapies. However, a review of real-world fixative methods reported in this study shows that ~8% of tissue samples from clinical trials are processed using alternative fixatives. This study further demonstrates that even as alternative fixatives (20% NBF and 10% UBF) may not impact HER2 status assessment in cell lines, they do have the potential to convert the HER2 status in tissues from positive to negative. Our results suggest that the use of alternative fixatives may negatively impact patient diagnosis and subsequent treatment options. Importantly, this study was not designed to address the accuracy of HER2-low status (IHC 1+ or 2+/ISH nonamplified) between alternative fixatives versus 10% NBF.

The chemical and physical properties of alternative fixatives make these reagents less appropriate for FFPE tissues and downstream DNA-based detection methods such as ISH compared with the guideline-recommended 10% NBF. The oxidation of UBF to formic acid reduces the pH to within a range of 4, which not only decreases fixation time, but also leads to lower amounts of amplifiable DNA.^30–33 DNA quality is adequately preserved in FFPE tissues processed with 10% NBF, but not with UBF or fixatives with a formaldehyde concentration higher than 10%.^34,35 FFPE tissues fixed with increased concentrations of formaldehyde, such as with 20% NBF, may also sustain increased DNA damage due to increased formaldehyde-induced crosslinks and DNA fragmentation³⁰ and may, therefore, impact results from an ISH assay.

In BC and GC cell lines, a minimal impact of alternative fixatives on HER2 status concordance was observed. In tumor tissue samples, however, there was only a 58% to 67% PPA between fixatives for HER2-positive and HER2-negative BC and GC tissue samples. The differences seen in staining concordance between cell lines and tumor tissue specimens may be due to the molecular composition of cultured cells compared with the complex cellular architecture of in vivo tumor tissue samples.³⁶ Findings from a reevaluation of HER2 status in 33 patients with HER2-positive advanced or recurrent GC refractory to trastuzumab showed that the use of fixatives other than 10% NBF significantly reduced the prevalence of HER2-positive status.²⁹ In contrast, a study using different types of fixatives on xenografted tumor tissues of human GC cell lines showed no differences in staining between fixatives of HER2 IHC 3+ specimens.²³ However, staining of HER2 IHC score 2+ specimens was reduced with fixatives other than 10% NBF.²³ Together, these results suggest that although alternative fixatives can negatively impact HER2 status, the level of impact may depend on the sample type and HER2 expression levels.

HER2-expression heterogeneity is also likely to contribute to a lower PPA in tissues, in addition to the type of fixative used. Tissue specimens may contain areas that are genetically heterogeneous for HER2 copy number due to biological heterogeneity.^8,9,37 Although this study was not designed to address the issue of tissue heterogeneity, heterogeneity was considered in the study design. The central portion of the tissue sample was fixed in 10% NBF, whereas the bookend portions of the same sample were fixed in 10% UBF and 20% NBF. All differently preserved sections from the same sample were reviewed by the same pathologist.

It has been further demonstrated that the presence of HER2 genetic heterogeneity is more common in cases with HER2-equivocal status by ISH and/or IHC.³⁷ In the present study, the most consistent trend for change from HER2-positive to HER2-negative status in tissue samples was due to a change in the ISH category (ISH+ to ISH−), followed by a change in the IHC category. However, among the samples with a change from ISH amplified to ISH nonamplified, most of the samples were ISH borderline or close, which may have led to discrepancies in HER2 status interpretation.

Discordant results of HER2 status within indications such as BC and GC can be further explained, to a certain degree, by preanalytical and analytical factors such as delay to, and prolonged, formalin fixation,¹⁹ tumor heterogeneity,^9,37 the diversity of kits and antibodies,³⁸ and data interpretation.³⁹ Although IHC testing methods have steadily improved with the introduction of automated staining platforms and improved antigen retrieval and detection systems, the need for assay validation and ongoing monitoring has not diminished given that results are still affected by various preanalytic and analytic factors. Importantly, biomarkers such as HER2 have the most clinical relevance when they are reproducibly and accurately measurable, clinically feasible, and prospectively validated in randomized clinical trials.

The limitations of this study include the small number of HER2-positive tissue samples, which led to large variations in 95% CI. In addition, BC and GC tissue concordance are analyzed together, and because HER2-expression in BC is more homogenous compared with HER2-expression in GC,^9,40 this can lead to large differences in CI. In addition, it is important to note that the current study was conducted before the latest (2018) ASCO/CAP guideline for BC was adopted by laboratories. Updates to the 2018 guidelines include a requirement for IHC-guided ISH rescoring by a second observer for ISH-equivocal samples (groups 2, 3, and 4). The 2018 guidelines also refined the definition of HER2 status (negative or positive) for ISH-equivocal (borderline) cases.⁸ Another limitation is the use of only 1 IHC/ISH assay to assess HER2 status. Based on these data, it is not possible to determine whether other assays are also impacted by fixative type and whether the effect is greater, or less, than that seen with the Ventana 4B5 assay.

CONCLUSION

The results of this study show that even though alternative fixatives may not impact HER2 status assessment in cell lines, the use of fixatives other than the guideline-recommended 10% NBF has the potential to convert the HER2 status in BC and GC tissues from positive to negative in both ISH and IHC. This suggests that the use of alternative fixatives may negatively impact patient diagnosis and subsequent treatment options.