Spanish 
Chinese 
Russian 
German 
French 
Japanese 
Portuguese 
Hindi Research Article, J Clin Exp Oncol Vol: 3 Issue: 2
HER2 Testing in Gastroesophageal Cancer Paired Biopsy and Resection Specimens, A Multi- Institutional Immunohistochemistry and Fluorescence In-Situ Hybridization Correlation Study
| Laura J. Tafe1,2*, Jackie Cao3, Marlene Sabbath-Solitare4, Teresita Cuyegkeng Redondo4, Jonathan Lara4 and Gregory J. Tsongalis1,2 | |
| 1Department of Pathology, Dartmouth-Hitchcock Medical Center and Norris Cotton Cancer Center, Lebanon, U.S.A. | |
| 2Department of Pathology, Theodor and Audrey Geisel School of Medicine at Dartmouth, Hanover, U.S.A. | |
| 3Department of Pathology, Baystate Health System, Springfield, MA, U.S.A. | |
| 4Department of Pathology, Saint Barnabas Medical Center, Livingston, NJ, U.S.A. | |
| Corresponding author : Dr. Laura J. Tafe, M.D Department of Pathology, Dartmouth Hitchcock, Medical Center, One Medical Center Drive, Lebanon, NH 03756, U.S.A Tel: 603-650-7211 E-mail: laura.j.tafe@hitchcock.org |
|
| Received: February 26, 2014 Accepted: May 12, 2014 Published: May 15, 2014 | |
| Citation: Tafe LJ, Cao J, Solitare MS, Redondo TC, Lara J, et al. (2014) HER2 Testing in Gastroesophageal Cancer Paired Biopsy and Resection Specimens – A Multi-Institutional Immunohistochemistry and Fluorescence In-Situ Hybridization Correlation Study. J Clin Exp Oncol 3:2. doi:10.4172/2324-9110.1000125 |
Abstract
Eligibility for treatment with trastuzumab (Herceptin) for metastatic gastroesophageal cancer (GE) requires demonstration of HER2 protein overexpression or HER2 gene amplification. However, in the US there are no guidelines for best practices of HER2 testing in GE. The goals of this multicenter study were to first evaluate whether the published 2007 ASCO/CAP guidelines established for breast cancer could be extrapolated to GE cancers; and second, to implement an inter-laboratory comparison of HER2 IHC and FISH performance in paired biopsy and tumor resection specimens. Fifty paired formalin-fixed paraffin-embedded (FFPE) biopsy and resection specimens of GE carcinomas were selected from the pathology archives of three institutions. HER2 FISH analysis was performed using the PathVysion DNA probe kit. Immunohistochemistry was performed on the Ventana Benchmark XT with Ventana’s Pathway anti- HER-2/neu (4B5) antibody. Scoring and interpretation of results were according to the ASCO/CAP guidelines for breast carcinoma. Two of the 50 paired cases, were excluded due to insufficient quantity of residual tumor in the biopsy specimen. 48 and 38 pairs were successfully evaluated by IHC and FISH, respectively. Overall, the agreement of the IHC and FISH scoring between institutions was good; Æ? = 0.76 (IHC) and very good; Æ? = 0.89 (FISH). The concordance between biopsies and resection specimens was 96% and 97% with IHC and FISH, respectively. We present data from an inter-laboratory collaborative study of the performance of HER2 IHC and FISH within GE tumors in paired biopsy and resection specimens. We found that the 2007 ASCO/CAP guidelines can be applied to GE cancer with good inter-observer reproducibility.
Keywords: HER2 FISH; Gastric Cancer; molecular diagnostic testing; HER2 testing; HER2 immunohistochemistry guidelines
Keywords |
|
| HER2 FISH; Gastric Cancer; molecular diagnostic testing; HER2 testing; HER2 immunohistochemistry guidelines | |
Introduction |
|
| In 1985, Coussens et al. discovered an EGFR homologous tyrosine kinase located at the same chromosomal location as the neu oncogene [1]. ERBB2 [commonly referred to as HER2/neu or simplyhuman epidermal growth factor receptor 2 (HER2 )] encodes for a 185-kd transmembrane tyrosine kinase receptor, a member of the epidermal growth factor family. Since this early description it has been shown that approximately 15-20% of breast cancers (as well as some other tumor types) have oncogenic activation of HER2 via gene amplification which confers a worse prognosis [2]. However, more than ten years elapsed from the discovery of this biomarker and the introduction of the first targeted therapy for HER2 amplified breast cancers. | |
| Trastuzumab (Herceptin), the first recombinant humanized monoclonal antibody therapy for human cancer, was approved by the US Food and Drug Administration (FDA) in 1998 and has become standard therapy for patients with HER2 - amplified breast cancer [2]. It soon became evident that other tumor types containing this amplified gene may benefit from similar therapy. Gastroesophageal (GE) cancers are aggressive with a 5-year survival rate less than 20%. There has been limited treatment success with conventional chemotherapy in these tumors and efforts have focused on identifying therapeutic molecular targets [3]. Approximately 20% of GE carcinomas harbor HER2 over expression or amplification and in 2010, based on the results of ToGA (trastuzumab for Gastric Cancer), a phase III, international, randomized controlled trial, the FDA approved trastuzumab for use in combination with cisplatin and either capecitabine or 5-fluorouracil in patients with metastatic GE cancer [4]. | |
| Earlier studies have shown more heterogeneity in HER2 expression and amplification in GE (4.8%) as compared to breast cancer (1.4%) and because of this IHC and FISH may have less correlation than in breast cancer [5]. This potentially complicates the analysis and interpretation of these tests, especially in small biopsy specimens. This is particularly relevant as many GE patients present with unresectable disease and a biopsy is the only tissue available for diagnosis and evaluation of HER2 status. | |
| Eligibility for treatment with trastuzumab for either breast or GE cancer requires demonstration of HER2 protein overexpression or HER2 gene amplification. Guidelines for companion diagnostic testing were initially established for breast cancer using immunohistochemistry (IHC) and/or fluorescence in situ hybridization (FISH) [6]. However, in GE carcinoma, the definition of HER2 positive tumors varies and the best HER2 testing strategy in these tumor types has not been defined. In the USA and Japan HER2 positivity is defined as IHC3+ or FISH-positive while in Europe (European Medicines Agency) it is defined as IHC 3+ or IHC 2+ confirmed by a positive FISH result [7]. While guidelines for HER2 testing in GE have been adopted by several countries (e.g. Belgium and Japan) [8,9], the US, unlike for breast cancer, has not issued guidelines for best practices of HER2 testing in GE cancer. | |
| The goals of this multicenter study were to first evaluate whether the published 2007 ASCO/CAP guidelines established for breast cancer could be extrapolated to GE cancers; and second to implement an inter-laboratory comparison of HER2 IHC and FISH performance in paired biopsy and tumor resection specimens. The three participating institutions included Dartmouth Hitchcock Medical Center (DHMC), Baystate Health System (BHS), and Saint Barnabas Medical Center (SBMC). | |
Materials and Methods |
|
| Fifty paired formalin-fixed paraffin-embedded (FFPE) biopsy and resection specimens of gastroesophageal (GE) carcinomas were selected from the pathology archives of three institutions. Original H & E slides were reviewed to document tumor content. All cases were blinded from patient identifiers and three H & E’s and nine unstained, 4 um sections were cut for each case. One H & E and three unstained slides were distributed to each of the collaborating institutions (Figure 1). Data was collected from pathology reports and included tumor histologic classification, differentiation and site of origin. Institutional review board (IRB) approval was obtained at all three institutions. | |
 |
Figure 1: Inter-laboratory study design. |
| Immunohistochemistry | |
| Immunohistochemistry (IHC) was performed at one of the participating laboratories (BHS) using the Ventana Benchmark XT with Ventana’s Pathway anti-HER-2/neu (4B5) Rabbit Monoclonal Primary Antibody (~6 μg/ml). IHC was scored using the 2007 ASCO/CAP HER2 IHC criteria by two pathologists (LJT and JC) housed at two of the three institutions. Briefly, 0-1+ staining was classified as negative (no staining or weak incomplete membrane staining in < 10% of cells), 2+ as equivocal (weak or non-uniform complete membrane staining in at > 10% but ≤ 30%), and 3+ as positive staining (uniform intense membrane staining of > 30% of invasive tumor cells) [6]. The correlation of the staining results between the two observers was determined. In addition, all cases were rescored by a single pathologist (LJT) using the Hofmann criteria which was used in the ToGA trial: 0, no staining or membranous reactivity in ≤ 10% of tumor cells; 1+, weak, barely perceptible membranous reactivity in > 10% of tumor cells; 2+, complete or basolateral membranous reactivity that is either non-uniform or weak in intensity in at least 10% of cells; and 3+, complete or basolateral membranous reactivity of strong intensity in ≥ 10% of cells [5]. Intratumoeral heterogeneity was noted when present. | |
| Fluorescence in-situ hybridization | |
| FISH analysis was performed at two of the participating laboratories (DHMC and SBMC) using the PathVysion DNA probe kit (Abbott Laboratories, Abbott Park, Illinois) according to each laboratory’s established clinical HER2 FISH protocol. A minimum of 40 cells were counted by two observers at each of the institutions using the 2007 ASCO/CAP guidelines for breast carcinoma to interpret results. A HER2:CEP17 ratio of < 1.8 was classified as negative for amplification, 1.8–2.2 as equivocal and > 2.2 as amplified [6]. | |
| Correlation between the two institutions was determined and heterogeneity was also noted when present. All IHC and FISH results were correlated for each case and biopsy vs resection scoring was examined. | |
| Statistical analysis | |
| Inter-observer concordance for IHC and FISH was analyzed using kappa (Ã?Â?) statistics. Analyses were performed using GraphPad QuickCalcs | |
Results |
|
| Twenty-four cases with paired biopsy and resection specimens were included from BHS, 17 from DHMC and 9 from SBMC. Two of the 50 cases were excluded due to insufficient quantity of residual tumor in the biopsy specimen. Of the remaining 48 cases there were 34 primary gastric and 14 primary gastro-esophageal junction (GEJ) tumors. Thirty-five (73%) were intestinal type (3 low grade, 22 intermediate grade and 10 high grade), 10 diffuse type (21%) and three mixed type (6%). | |
| HER2 Immunohistochemistry | |
| All of the cases were successfully evaluated by the IHC technique at two institutions (DHMC and BHS) using the 2007 ASCO/CAP scoring system. There was 100% correlation between the two institutions in the cases showing 3+ staining. Three biopsies and two excisions were scored as positive (3+) for HER2 overexpression by IHC. The one case that was scored as a 2+ on the excision specimen showed heterogeneous 3+ staining in <30% of the tumor cells; 30% being the cutoff of a positive result according to the 2007 scoring system. | |
| Scoring differences were observed mainly in the equivocal cases. One institution scored four biopsies and five excisions as equivocal (2+) compared to one biopsy and three excisions by the other institution. Overall, the agreement of the IHC scoring between the two institutions was good; Ã?Â?=0.76 (95% CI 0.56-0.96) [Ã?Â?=0.70 (95% CI 0.39-1.0) for the biopsies and Ã?Â?=0.82 (95% CI 0.57-1.0) for the excision specimens]. (Table 1) | |
|
Table 1: Immunohistochemistry scoring between the two institutions that evaluated IHC. (1 case 3+ on biopsy was down scored to 2+ on the excision with heterogeneity <30%) |
| HER2 FISH | |
| Out of 96 individual FISH slides examined (48 biopsy and 48 excisions) by two sites (DHMC and SBMC), four failed processing at one institution and nine failed at the other despite repeated attempts when adequate slides were available. Ten pairs did not have results from both institutions; therefore, only 38 paired biopsy and excision cases were successfully evaluated by FISH at both institutions. Thirtythree of the 38 pairs did not show HER2 amplification; the remaining five were amplified with 97% concordance between the laboratories. One pair showing low level amplification was discrepant between the two laboratories; the HER2:CEP17 ratios for the biopsy and the excision were 2.5 and 2.0 and, 2.0 and 2.4. All five amplified cases were intestinal type tumor with moderate to poor differentiation; two of gastric origin and three that arose in the GEJ. Overall, there was very good inter-observer agreement for the FISH scoring between the two institutions: Ã?Â?=0.89 (95% CI 0.75-1.0) (Table 2). | |
|
Table 2: FISH scoring between the two institutions that performed FISH; 38 pairs were successfully evaluated by both institutions. Only one pair was discrepant between the two labs; this was a low level amplified case with HER2:CEP17 ratio for the biopsy and the excision as 2.5 and 2.0 and 2.0 and 2.4, respectively for the two labs. |
| Concordance between biopsy and resection specimens | |
| Using IHC and FISH results from the DHMC laboratory we evaluated the scoring between biopsy and excision specimens. For IHC, excellent concordance (96%) was observed. Two of 48 paired specimens had discordant results; both showed intra-tumoral heterogeneity (4.2% of cases). In one case, the biopsy was scored as negative (1+) but 3+ heterogeneous staining was seen in the excised tumor in less than 30% of tumor cells and was thus scored as 2+ (equivocal according to the 2007 ASCO/CAP guidelines). The second case showed 3+ staining on the biopsy specimen with heterogeneous staining in <30% of tumor cells in the excision and was scored as 2+. | |
| Of the 38 pairs that were successfully scored by FISH, all but one had concordant results between biopsy and resection specimens (97%); in this case the biopsy was scored as amplified (HER2:CEP17 ratio of 2.5) while the excision was scored as equivocal (HER2:CEP17 ratio of 2.0). Of the two cases with heterogeneous staining by IHC, one was negative by FISH with noted focal heterogeneous amplification in the resection (<50% of tumor cells) (Vance 2009) and the other was HER2 amplified in both the biopsy and excision specimens (Figure 2). In addition to the heterogeneous cases, two cases that were scored as negative by IHC (0-1+) showed low level of HER2 amplification (HER2:CEP17 ratios of 2.0 and 2.8 and, 2.5 and 2.4 in the biopsies and excisions, respectively) (Figure 3). One pair was equivocal by IHC and negative by FISH. | |
|
Figure 2: A biopsy showing (A) heterogeneity by IHC and (B) amplification by FISH. |
|
Figure 3: In one case (A) the biopsy and (B) resection showed 1+ and 0 IHC staining, respectively. (C) By FISH, this case was clearly amplified. |
Discussion |
|
| Approximately 20% of gastroesophageal (GE) cancers show amplification of HER2 , providing these patients eligibility for treatment with trastuzumab [4]. Currently in the United States, there are no guidelines for the best practices of establishing HER2 overexpression or amplification in GE cancers. Multiple modalities have been applied to these tumors including IHC, FISH, CISH and SISH. European countries utilize the IHC scoring schemes based on the Hofmann criteria which was used for the ToGA trial with IHC being the primary test followed by FISH when necessary for confirmation [4,5,9]. The differences of the Hofmann criteria for GE cancer from the scoring scheme in breast cancer is the percentage of tumor cells showing staining (3+ is >10% vs. >30% for breast) and it accommodates the basolateral membranous pattern seen commonly in GE cancers. In addition, due to the known heterogeneity of these GE tumors, the Hofmann criteria score biopsy specimens positive when strong membranous staining is present, irrespective of the percentage of positive cells. Ruschoff, et al. recently proposed a minimum of 5 cohesive cells demonstrating strong complete or basolateral membranous staining before classifying a biopsy sample as HER2 positive [10]. | |
| In this study, we show that the 2007 ASCO/CAP breast cancer guidelines for scoring HER2 expression and gene amplification may be applied to GE cancers with excellent correlation and reproducibility observed amongst the three institutions. One advantage of applying the ASCO/CAP breast cancer guidelines is that it is familiar to pathologists and does not require utilization of multiple scoring criteria. In our study five of the pairs were amplified by FISH, a rate of 10%, which is lower than that reported in the literature for GE cancer. Consistent with other reports, all five cases HER2 positive by FISH were intestinal type with moderate to poor differentiation. | |
| There were several limitations recognized in this study. First, our study set had a relatively low rate of HER2 overexpression and gene amplification, possibly due to the selection of cases that had both biopsy and resection tissue available for the study. This specimen selection bias only included cases that had both a biopsy and resection specimen available which was not completely representative of sequential GE cases. Because the majority of cases were negative our inter-laboratory correlations were very good and, therefore, may not completely capture the variability between observers. Second, this study was performed using the 2007 ASCO/CAP guidelines for breast cancer currently in practice in the US. Revised guidelines for HER2 scoring in breast cancer have recently been introduced [11]. | |
| Few studies have focused on HER2 concordance between biopsies and subsequent resection specimens [12-14]. Watson, et al. studied paired samples from 228 patients by IHC with confirmatory CISH results in cases with an IHC score of 2+. Thirty-seven (17%) paired samples showed discrepancies in HER2 status by IHC (either 0/1+ on biopsy that was upscored to 2+/3+ on excision or vice versa) but once CISH was performed on the 2+ cases and factored into the comparison, the overall concordance rate between biopsies and surgical specimens reached 94% [12]. Lee, et al. found a 74.1% concordance of IHC HER2 status in 54 paired biopsies and gastrectomies; and a false negative IHC result (either biopsy or gastrectomy specimen) was seen in 19.4% of cases showing HER2 amplification by silver in-situ hybridization (SISH) [13]. Other studies have taken the approach of creating tissue microarrays (TMAs) from resection specimens to mimic biopsies and draw conclusions on the IHC or ISH correlation between the TMA cores and the excisions [10,15-17]. All of these studies utilized the ToGA IHC and FISH scoring criteria. | |
| The recently released updated ASCO/CAP guidelines for HER2 testing in breast cancer have some significant differences from the 2007 guidelines and are actually much more in alignment with the ToGA criteria (Table 3). | |
|
Table 3: (A): HER2 scoring of IHC according to the 2007 ASCO/CAP guidelines, 2013 ASCO/CAP guidelines and Hofmann criteria. (B): HER2 scoring of FISH according to the 2007 ASCO/CAP guidelines, 2013 ASCO/CAP guidelines and Hofmann criteria. |
| With the new IHC guidelines an IHC score of 0 is no staining or incomplete, weak membranous staining within ≤10%; 1+ incomplete, weak membranous within > 10%; 2+ incomplete weak/moderate within > 10%, or complete and circumferential intense membranous staining within ≤ 10% of tumor cells; and finally 3+ is complete, circumferential intense membrane staining in > 10% of tumor cells. This cutoff of 10% for 3+ is significantly different than the > 30% of the 2007 guidelines; which most likely will result in inclusion of a number of cases that previously would have been scored as false negative, especially in biopsies. | |
| In addition, these criteria essentially eliminate the previously poorly defined heterogeneous cases that fell into the < 30% amplification range. The revised HER2 FISH guidelines are also different from the 2007 guidelines in some important ways. The new criteria define HER2 amplification as HER2:CEP17 ratio ≥ 2.0 or an average HER2 copy number ≥ 6.0 signals per cell and negative as HER2:CEP17 ratio < 2.0 with an average HER2 copy number < 4.0 signals per cell. The prior equivocal range using HER2:CEP17 ratio of 1.8-2.2 has been eliminated and now equivocal FISH is determined based on average HER2 copy number per cell. Previous intratumoral heterogeneity was poorly defined for FISH as 5-50% [18], now with the new criteria, > 10% of cells with HER2 amplification is considered positive, significantly decreasing the definition of heterogeneity to tumors with < 10% amplification [11]. | |
| When the ToGA criteria were applied to our cases, 27 cases were reclassified on IHC; 25 of these were 1+ (any staining) by the 2007 ASCO/CAP criteria but were now scored to 0 with the Hofmann criteria (< 10% staining) (Table 4). | |
|
Table 4: Comparison of the 2007 ASCO/CAP IHC scoring with the Hofmann scoring used in the ToGA trial. Scoring was identical in the biopsy specimens; in the resection specimens two cases were upscored to 3+ from 2+ using the ToGA criteria. These two cases showed heterogeneous staining of less than 10% of tumor cells 3+. |
| Two biopsies were originally classified as 2+ but were now scored to 3+ with the Hofmann criteria. These biopsies showed heterogeneity of the IHC staining. For FISH, all five cases HER2 amplified with the 2007 criteria were also amplified using ToGA criteria. In addition, in one excision heterogeneity was identified with a small focus of amplification, this was called non-amplified with the 2007 criteria but would be considered amplified with the ToGA criteria; this case would also be considered positive with the revised ASCO/CAP criteria. | |
| The Hofmann criterion recognizes heterogeneity in GE cancers and on biopsy specimens cohesive IHC 3+ and/or FISH + clones are considered positive irrespective of size. However, there is no formal definition of heterogeneity and the reported rates vary from 4.8% up to 50% in GE cancers. The clinical significance of heterogeneous staining in respect to response to trastuzumab therapy is not yet understood; heterogeneity has implications for testing biopsies, with potential for false negative IHC results [13]. Similar to our findings, others report that heterogeneity is a frequent cause of discrepant IHC and FISH results. Kunz et al report a 4.7% rate of heterogeneity and on review of the resection specimens, there were demonstrable differences in HER2 expression and gene amplification in heterogeneous areas which corresponded to morphologically distinct areas of tumor [15]. | |
| In this study, two cases were negative by IHC and showed low level amplification on FISH. In contrast, no cases negative for HER2 amplification by FISH demonstrated 3+ over-expression by IHC. Similar findings have been reported in the literature; Ruschoff et al. reported that up to 32% of IHC 2+ cases and 5% of IHC 1+ cases were amplified by FISH [10]. Kunz et al. report three carcinomas negative by IHC that showed HER2 amplification by FISH; in their series none of the FISH amplified cases had IHC overexpression [15]. Heterogeneity and sampling issues are inherent disadvantages in biopsies however; biopsies may have better antigen preservation than resection specimens due to their smaller size and ease of fixation. IHC can be challenging to interpret for other reasons too, especially in small biopsies. As described by others, we also noticed that some cases have strong artifactual cytoplasmic and nuclear staining in the tumors which can either mask true staining or be over interpreted as false positive membranous staining. This finding is likely dependent on the antibody that is used; in such cases FISH confirmation is probably appropriate [3,19]. Due to these potential pitfalls of IHC interpretation, DHMC and SBMC do not screen cases with IHC but rather test all cases by FISH initially. While the ToGA trial established an apparent correlation between HER2 expression by IHC and response to trastuzumab therapy, a recent study shows that the level of HER2 gene amplification is a significant predictor of sensitivity to trastuzumab-based chemotherapy and overall survival in advanced gastric cancer, therefore, FISH rather than IHC may become the more prognostic test [4,20]. | |
Conclusion |
|
| In conclusion, we found that the 2007 ASCO/CAP guidelines can be applied to GE cancer with good inter-observer reproducibility. The concordance between biopsies and resection specimens was 96% and 97% with IHC and FISH, respectively. The new ASCO/CAP guidelines are very similar to the ToGA scoring criteria and, with further evaluation, may also be applicable to GE cancer. | |
Acknowledgments |
|
| The authors wish to thank the technical staff at the participating institutions and acknowledge the support of Abbott-Vysis for providing the FISH reagents. | |
References |
|
|
|
