- Email: [email protected]
gastro-oesophageal junctional cancers: should determination of gene amplification by SISH use HER2 copy number or HER2:CEP17 ratio?
Pathology (April 2014) 46(3), pp. 184–187
ANATOMICAL PATHOLOGY
HER2 status in gastric/gastro-oesophageal junctional cancers: should determination of gene amplification by SISH use HER2 copy number or HER2:CEP17 ratio? MARIAN PRIYANTHI KUMARASINGHE1,2, WILLEM BASTIAAN DE BOER1,2, TZE SHENG KHOR1, ESTHER M. OOI3, NIC JENE4, SURESHINI JAYASINGHE4 AND STEPHEN B. FOX4,5 1PathWest Laboratory Medicine, Perth, 2School of Pathology and Laboratory Medicine, University of Western Australia, Perth, 3School of Medicine and Pharmacology, Royal Perth Hospital Unit, University of Western Australia, Perth, WA, 4Department of Pathology, Peter MacCallum Cancer Centre, East Melbourne, and 5Department of Pathology and Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Vic, Australia
Summary The aim of this study was to compare HER2 amplification, as determined by the HER2 copy number (CN) and the HER2/ CEP17 ratio, with protein expression in gastric and gastrooesophageal junction (G/GOJ) adenocarcinoma. HER2 immunohistochemistry (IHC) and silver in situ hybridisation (SISH) were performed in 185 cases. Modified gastric criteria were used for IHC scoring. HER2 and CEP17 CNs were counted in at least 20 cancer cells and the ratio calculated as per previously defined protocols. These two SISH methods were statistically compared against the different IHC scores. Thirty-four cases showed amplification, by both methods in 29, and either method in five. IHC score was 3þ in 29 cases; 26 showed amplification by both methods, one by ratio only and two were not amplified. IHC score was 2þ in 24 cases; three showed amplification by both methods and two by either. One each of IHC 1þ and 0 showed an increased ratio but not CN. The HER2 CN and ratio for IHC score 3þ compared to scores 2þ, 1þ and 0 were significantly different (all p < 0.01). The CN for IHC 2þ vs IHC 1þ and IHC 0 was significantly different (both p < 0.01) but the ratio was not ( p ¼ 0.5711 and p ¼ 0.2857, respectively). The CN and the ratio for scores 1þ and 0 were not significantly different ( p ¼ 0.9823 and p ¼ 0.9910, respectively). The HER2 CN differentiates between the different IHC scores better than the HER2:CEP17 ratio. Cases that show IHC3þ and high CN may not require calculation of the ratio. Furthermore, consideration should be given to the CN when IHC negative cases appear amplified by the ratio only. Key words: Amplification, copy number, double probe method, gastric/gastroesophageal junctional carcinoma, HER2 status, ratio, single probe method. Received 17 June, revised 29 October, accepted 7 November 2013
INTRODUCTION HER2 positive status determines the eligibility for HER2 targeted therapy of advanced and metastatic gastric and gastro-oesophageal junction (G/GOJ) adenocarcinomas. The HER2 status can be determined by estimation of protein expression by immunohistochemistry (IHC) and/or assessment of HER2 gene copy number and centromeric probe 17 (CEP17)
Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000075
#
ratio by in situ hybridisation (ISH). ISH can be performed by bright field techniques such as chromogenic ISH (CISH) or silver ISH (SISH) or by dark field methods using fluorescence (FISH). Although there is evidence that bright field methods are superior to FISH in determining the gene amplification for G/GOJ cancers,1–3 there is no universal acceptance of a method for testing at this point.4,5–7 The Trastuzumab for Gastric Cancer (ToGA) study defined ISH positive status for G/GOJ cancer as HER2:CEP17 ratio 2.0 irrespective of the copy number.4 This criterion has been used for clinical testing and in most studies related to the HER2 status of gastric/ GOJ cancer. The majority of studies show a perfect or near perfect correlation between IHC3þ expression and gene amplification.4,8–10 However, they also report that a proportion of equivocal and negative IHC cases are ISH positive when the ratio is used. IHC is a semi-quantitative test that is subject to a variety of pre analytical and analytical issues involving tissue sampling, scoring criteria and inherent subjectivity of IHC interpretation. In contrast, ISH technique is considered superior, being more quantitative and reproducible. There are specific issues confounding HER2 testing in G/GOJ cancers by either method, such as the type of specimen used (i.e., endoscopic biopsies, resections or metastatic material) and a greater degree of heterogeneity reported in these cancers.2,8 Considering the special issues, a modified IHC scoring system was established for HER2 testing of G/GOJ cancers.9 Superiority of the modified/gastric cancer scoring system has been validated by others subsequently.4,10 A positive IHC reaction in G/GOJ cancers includes basolateral/lateral membrane staining as opposed to the requirement of complete membrane staining in breast carcinomas. Additionally the cut-off for a positive test is only 10% positive tumour cells for resections and a cluster of five positive tumour cells for endoscopic biopsies.9,10 There are two methods to establish gene amplification. One is the single probe method by counting the actual HER2 copy number (CN) per nucleus, the cut-off for amplification being 6.0. The other is using dual probes to calculate the ratio of HER2 genes to centromere 17 (HER2/CEP17), the cut-off for amplification being 2.0. The ratio distinguishes increased HER2 gene copy number secondary to extra copies of CEP17 that may occur due to true polysomy or
2014 Royal College of Pathologists of Australasia
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
HER2 AMPLIFICATION BY COPY NUMBER OR RATIO?
co-amplification. For the ToGA study, HER2 positive status was defined by either protein expression that showed a IHC 3þ score, or FISH gene amplification defined as a HER2/CEP17 ratio 2.0, not a HER2 CN 6.0. Significantly, subgroup analysis showed that there was no survival benefit with trastuzumab therapy for the HER2 amplified but IHC negative group (score 0 or 1þ) which accounted for 23% of the positive cohort. The exact reason for this has not been established to date. A recent review on the issue of HER2 testing in gastric cancer by an expert panel2 recommends taking into consideration the gene count, i.e., CN when the ratio suggests borderline amplification (a ratio close to 2.0). Furthermore, the GaTHER study conducted in Australia concluded that the inter-laboratory agreement on CISH/SISH scoring was good/ very good when HER2 copy number was used (k ¼ 0.68 to 0.86), but was reduced when HER2:CEP17 ratio was used.1 In spite of the above, there are no studies that have compared the performance of the ratio versus the HER2 copy number across the range of IHC. In our experience with HER2 testing of gastric/GOJ cancers, we have encountered a few cases that just reach the cut-off ratio of 2 whilst the actual HER2 copy number does not exceed 6. Therefore, we embarked on a detailed analysis of the HER2 copy numbers and HER2/ CEP17 ratio across the range of IHC scores, also comparing the results against tumour type/grade, and taking into account the potential impact on clinical testing.
METHODS Cases included in the study were: endoscopic biopsies (n ¼ 146, 80.7%), selected blocks of resections (n ¼ 30, 16.6%) and metastatic deposits (n ¼ 9, 2.7%) of G/GOJ adenocarcinomas with an adequate amount of tumour. HER2 status was assessed by IHC and SISH. IHC scoring (0–3þ) was performed using ‘modified gastric’ criteria. HER2 and CEP17 CNs were counted in at least in 20 cancer cells and the HER2:CEP17 ratio calculated as per standard protocol.1,4,9 These two methods were statistically compared across the range of IHC scores, as well as the tumour grade and type. Immunohistochemistry HER2 IHC was performed on formalin fixed, paraffin embedded tissue using the Ventana XT automated stainer (Ventana, USA), with polyclonal CERB2 antibody (Dako, Denmark) applied at 1:4000 dilution. The resection and/or biopsy specimens were scored for HER2 overexpression according to criteria by Hofman et al.9 Silver in situ hybridisation (SISH) HER2 amplification was performed on formalin fixed, paraffin embedded tissue using the automated Ventana INFORM HER2 Genomic probe platform. HER2 and CEP17 copy numbers were counted and the CN was averaged per cell. In each case HER2/CEP17 ratio was assessed. The cut-off for amplification by the copy number was 6 and by the ratio was 2 as per established guidelines.1,4,9 Table 1
185
Statistical analysis Skewed variables were logarithmically transformed where appropriate. Data among groups were compared using general linear models (SAS Proc GLM; SAS Institute, USA). All pair-wise comparisons and their corresponding p values for the test of no difference are reported. Tukey–Kramer test was applied to account for multiple comparisons for a given variable across the four IHC scores. The statistical significance was set at the 5% level.
RESULTS There were 185 G/GOJ carcinomas in which IHC and HER2 and CEP17 SISH had been performed. IHC scores and SISH results of CN and ratio A total of 34 cases demonstrated an increase in either CN or ratio (n ¼ 5, CN only in 1 and ratio only in 4) or both (n ¼ 29). Table 1 shows the comparison of the HER2 CN, HER2/CEP17 ratio and IHC score. Table 2 shows the raw figures of HER2 and CEP17 signals, ratio and the respective IHC score in discordant cases. Of 29 IHC 3þ cases, 26 showed amplification by both ratio and CN, one case by ratio only, and two cases were not amplified. Of the 24 IHC 2þ cases, three were amplified by both CN and ratio, with two additional cases amplified by either CN or ratio respectively, and 19 cases were not amplified. One each of IHC 1þ and 0 showed increased ratio but not CN. Mean CN for IHC 3þ cases was markedly elevated above 6 (18.92), while all other scores showed a mean copy number that was <6. Fig. 1 shows a concordant case (IHC 3þ, copy number 6 and ratio 2) and a discordant case (IHC 1, copy number <6 but ratio 2). Statistical comparison of ratio and copy numbers across IHC scores Table 3 shows the comparisons of the ratio and the CN across IHC scores and their corresponding p values. The HER2 CN and ratio for IHC score 3þ versus scores 2þ, 1þ and 0 were significantly different from each other (all p < 0.01). The CN for IHC 2þ versus IHC 1þ and IHC 0 was significantly different from each other (both p < 0.01) but the ratio was not ( p ¼ 0.5711 and p ¼ 0.2857, respectively). The CN and the ratio for scores 1 and 0 were not significantly different ( p ¼ 0.9823 and p ¼ 0.9910, respectively). HER2 status by definitions used across the world Using current ToGA (IHC 3þ and/or any ISH positivity by a HER2/CEP17 ratio of 2.0), European Medicines Agency [(EMEA); all IHC 3þ and IHC 2 and ISH positive by ratio of 2.0] and Belgian (all SISH positive by ratio of 2.0 irrespective of IHC) criteria, 36 (19%), 34 (18%) and 33 (17%) out of 185 cases would be positive, respectively. If the copy number was used, positivity would reduce to 34 (18%), 31 (17%), and 30 (16%), respectively. By using recently
Comparison of IHC score, HER2 CN and HER2/CEP17 ratio
IHC
SISHþ (ratio)
SISHþ (CN)
SISH– (ratio)
SISH– (CN)
Total
Mean CN
Mean ratio
0 1þ 2þ 3þ Total
1 1 4 27 33
0 0 4 26 30
98 32 20 2 152
99 33 20 3 155
99 33 24 29 185
2.48 2.57 4.14 18.92
1.15 1.26 1.94 9.57
CN, copy number; IHC, immunohistochemistry; SISH, silver in situ hybridisation.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
186
KUMARASINGHE et al.
Pathology (2014), 46(3), April
Table 2 Distribution of HER2 and CEP17 counts, ratio and IHC score of the discordant cases (n ¼ 5)
Case no. 64 110 2 137 163
HER2 signal/cell
CEP17 copy signals/cell
HER2/CEP17 ratio
IHC score
3.0 5.1 6.0 5.5 5.2
1.5 1.9 4.2 2.0 2.1
2.0 2.6 1.4 2.8 2.5
0 1 2 2 3
Table 3 numbers
Statistical comparison between IHC scores versus ratio and copy
IHC score p values HER2 signals per cell Ratio of HER2 to CEP17
0 1 2 0 1 2
vs vs vs vs vs vs
1: 2: 3: 1: 2: 3:
p ¼ 0.9823 p < 0.01 p < 0.01 p ¼ 0.9910 p ¼ 0.5711 p < 0.01
0 vs 2: p < 0.01 1 vs 3: p < 0.01
0 vs 3: p < 0.01
0 vs 2: p ¼ 0.2857 1 vs 3: p < 0.01
0 vs 3: p < 0.01
IHC, immunohistochemistry.
IHC, immunohistochemistry.
updated Australian guidelines (IHC 2þ or IHC 3þ plus gene copy >6 plus ratio of >2) the positive rate would be only 15% (29/185). The four of five discrepant cases that only showed an increased ratio were amplified at very low levels of 2.0, 2.6, 2.8 and 2.5 (Table 2).
as lapatinib as well as trials to determine the best partner chemotherapeutic agents in the management of G/GOJ adenocarcinomas.12,13 Protein overexpression is triggered by gene amplification and is manifested by increased copies of the HER2 gene. Hence protein overexpression is expected if there is significant gene amplification, although there may be other mechanisms that may interfere with this.14 Post-hoc subgroup analysis of the ToGA series indicates that positive treatment effects are limited to targeting tumours with higher levels of HER2 protein expression (IHC 2þ/FISHþ and IHC 3þ/regardless of the FISH status) compared to the those with low level of protein expression even if the FISH result was positive, i.e., amplified. EMEA subsequently defined HER2 positivity as an IHC 3þ reaction or IHC 2þ with a positive ISH result to be eligible for trastuzumab therapy. A positive ISH (FISH/SISH/CISH) result was defined by the ratio 2.0 irrespective of the copy number count. A recent review2 on the issue of HER2 testing in gastric cancer by an expert panel recommends taking into consideration the gene count when the ratio suggests borderline amplification (ratio close to 2.0). The GaTHER study demonstrated that gene amplification assessment based on HER2 CN
Comparison of HER2 copy number and ratio across severity of tumour grading and type Moderately differentiated (G2) carcinomas showed the highest mean CN (7.54) and ratio (3.76) . ‘Intestinal’ or gland forming adenocarcinomas showed a higher proportion of HER2 positivity (13/54, 24%), compared to the ‘diffuse ‘type (1/11, 9%) ( p ¼ 0.26, for both CN and ratio).
DISCUSSION Trastuzumanb is a monoclonal antibody that acts against the HER2 protein when it is overexpressed on the cell membrane blocking the downstream effects.11 Trastuzumab is currently the only approved targeted agent for G/GOJ adenocarcinomas following the results of the ToGA trial.4 However, there are several ongoing studies on other HER2 targeting agents such
A
B
C
D
E
F
Fig. 1 (A) HER2 immunohistochemistry (IHC) in carcinoma cells with no membrane positivity (IHC 1). (B) Corresponding HER2 silver in situ hybridisation (SISH) featuring an average CN of count of 5.1. (C) Corresponding CEP17 featuring an average count of 1.9 resulting in a ratio of 2.6. (D) HER2 immunohistochemistry in carcinoma cells with strong positive membrane staining (IHC 3). (E) Corresponding HER2 SISH featuring large and small clots with an average count of 18. (F) Corresponding CEP17 featuring an average count of 1.8 resulting in a ratio of 10.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
HER2 AMPLIFICATION BY COPY NUMBER OR RATIO?
was more reproducible than assessment based on the HER2/ CEP17 ratio, concluding that single-probe CISH or SISH is the optimal method for HER2 ISH testing in G/GOJ carcinomas in Australia.1 In this study, we provide evidence that HER2 copy number differentiated between IHC scores better than the ratio. This finding also suggests that (1) consideration should be given to the copy number when IHC negative cases show low level amplification by ratio only, and (2) calculation of the ratio is not necessary for cases that are IHC 3þ with a high copy number. Considering the costs and side effects of treatment, all factors that could potentially affect a positive result need to be evaluated carefully. Treatment failure due a to false positive result will have a negative impact on the ongoing advances in the management of upper GI adenocarcinomas with targeted therapy.4,12,13 The experience of the ToGA study that showed a lack of survival benefit in 23% of the positive cohort where the HER2 gene was amplified (by ratio) but IHC was negative (score 0 or 1þ) is also in support of caution. In our series, a very high proportion of cases that showed gene amplification both by the single probe (HER2 copy number) and the dual probe (HER2/CEP17 ratio) were IHC 3þ. The mean copy number for these IHC 3þ cases was 18.92, an unequivocal increase that is unlikely to be misinterpreted. Often these highly amplified cases show small or large clumps of gene copy number signals that are easily visible. In parallel to this, the GaTHER study showed very good inter-laboratory agreement for HER2 copy number for IHC 3þ cases. There may be no added value in the dual probe method for cases that are IHC 3þ with such an unequivocal increase in the copy number. In contrast, the mean copy number of IHC 2þ (equivocal), IHC 1þ (negative) and IHC 0 (negative) were 4.54, 2.67 and 2.54, respectively, and the difference between IHC 3þ and others was statistically significant according to our results. The four additional cases positive by ratio only showed very low amplification of the ratios, between 2.0 and 2.8 (2 cases were IHC negative, 1 case was equivocal and 1 was positive). Interestingly, Ru¨schoff and others recently reported low level of amplification (defined as ratio of 2–3) in most of their IHC negative and equivocal cases (76%), whereas only 16% of their IHC 3þ cases showed low levels of amplification.15 Our findings are similar. Currently there are no data comparing the outcomes of low and high level amplification irrespective of IHC results. Considering the technical and practical issues of IHC testing, unresolved issues of SISH testing, limited evidence of clinical outcomes, and special issues of G/GOJ carcinoma such as heterogeneity and variability of samples that may be tested, it would be prudent to rely on more than one method of testing as well as interpretation. Modifications and updated guidelines should continue to evolve as more experience is gained in clinical testing. The reasons for the discrepancy between the actual copy numbers and the ratio may be a result of several factors. Suboptimal staining of the CEP17 probe was noted in some samples, which could result in artificial elevation of HER2: CEP17 ratio. As stated above, this issue will not be a significant problem for cases with high/unequivocal amplification. Another confounding issue may be overlapping of nuclei that may be significant in gastric carcinomas, especially when testing is done on endoscopic biopsies. Technically, overlapping nuclei should not be counted, however, moderate to poorly differentiated gastric adenocarcinomas typically show intricate
187
glandular structures with marked nuclear overlapping and aneuploidy. Theoretically, loss of CEP17 signals may also lead to an artificially elevated ratio. For all these reasons when an IHC negative or equivocal sample shows a ratio 2, the result needs to be reassessed and the copy number considered. A second count by the same assessor or a second assessor and/ or repeat staining may be necessary. In conclusion, our findings show that the HER2 CN differentiated between IHC scores better than the ratio. Consideration should be given to the copy number when IHC negative cases appear amplified by the ratio only. IHC 3þ cases with high copy numbers do not need assessment of the ratio. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Professor M. P. Kumarasinghe, Department of Anatomical Pathology, PathWest, J Block, Queen Elizabeth II Medical Centre, Hospital Avenue, Nedlands, WA 6009, Australia. E-mail: [email protected]
References 1. Fox SB, Kumarasinghe MP, Armes JE, et al. Gastric HER2 Testing Study (GaTHER): an evaluation of gastric/gastroesophageal junction cancer testing accuracy in Australia. Am J Surg Pathol 2012; 36: 577–82. 2. Ru¨schoff J, Hanna W, Bilous M, et al. HER2 testing in gastric cancer: a practical approach. Mod Pathol 2012; 25: 637–50. 3. Powell WC, Zielinski D, Ranger-Moore J, et al. Determining the HER2 status in gastric cancer: a method comparison study of two patient cohorts. Gastrointestinal Cancers Symposium Orlando, USA, 2010; Abstract 17. 4. Bang Y-J, Van Cutsem E, Feyeereislova A, et al. Tratuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet 2010; 376: 687–97. 5. Jouret-Mourin A, Hoorens A, De Hertogh G, et al. Analysis of HER2 expression and gene amplification in adenocarcinoma of the stomach and the gastro-oesophageal junction: rationale for the Belgian way of working. Acta Gastroenterol Belg 2012; 75: 9–13. 6. European Medicines Agency. Assessment Report for Herceptin. Procedure No. EMEA/H/C/278/II/0047. London: European Medicines Agency, 2010. http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_ Assessment_Report_-_Variation/human/000278/WC500074921.pdf 7. National Cancer Institute. FDA approval for trastuzumab: HER2overexpressing metastatic gastric or gastroesophageal (GE) junction adenocarcinoma. Cited Jan 2013. http://www.cancer.gov/cancertopics/ druginfo/fda-trastuzumab 8. Lee S, de Boer WB, Fermoyle S, et al. HER-2 testing in gastric carcinoma: issues related to heterogeneity in biopsies and resections. Histopathology 2011; 59: 832–40. 9. Hofmann M, Stoss O, Shi D, et al. Assessment of a HER-2 scoring system for gastric cancer: results from a validating study. Histopathology 2008; 52: 797–805. 10. Park YS, Hwang HS, Park HJ, et al. Comprehensive analysis of HER2 expression and gene amplification in gastric cancers using immunohistochemistry and in situ hybridization: which scoring system should we use? Hum Pathol 2012; 43: 413–22. 11. Gajria D, Chandarlapaty S. HER2-amplified breast cancer: mechanisms of trastuzumab resistance and novel targeted therapies. Expert Rev Anticancer Ther 2011; 11: 263–75. 12. Hoffmann-La Roche. A study of capecitabine (Xeloda) in combination with trastuzumab (Herceptin) and oxaliplatine in patients with resectable gastric cancer. Cited Jan 2013. http://clinicaltrials.gov/show/NCT01130337 13. Iqbal S, Goldman B, Fenoglio-Preiser CM, et al. Southwest oncology group study S0413: a phase II trial of lapatinib (GW572016) as first-line therapy in patients with advanced metastatic gastric cancer. Ann Oncol 2011; 22: 2610–5. 14. Kanta SY, Yamane T, Dobashi Y, et al. Topoisomerase IIa gene amplification in gastric carcinomas: correlation with the HER2 gene. An immunohistochemical, immunoblotting, and multicolor fluorescence in situ hybridization study. Hum Pathol 2006; 37: 1333–43. 15. Ru¨schoff J, Dietel M, Baretton G, et al. HER2 diagnostics in gastric cancer – guideline validation and development of standardized immunohistochemical testing. Virchows Arch 2010; 457: 299–307.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
ANATOMICAL PATHOLOGY
HER2 status in gastric/gastro-oesophageal junctional cancers: should determination of gene amplification by SISH use HER2 copy number or HER2:CEP17 ratio? MARIAN PRIYANTHI KUMARASINGHE1,2, WILLEM BASTIAAN DE BOER1,2, TZE SHENG KHOR1, ESTHER M. OOI3, NIC JENE4, SURESHINI JAYASINGHE4 AND STEPHEN B. FOX4,5 1PathWest Laboratory Medicine, Perth, 2School of Pathology and Laboratory Medicine, University of Western Australia, Perth, 3School of Medicine and Pharmacology, Royal Perth Hospital Unit, University of Western Australia, Perth, WA, 4Department of Pathology, Peter MacCallum Cancer Centre, East Melbourne, and 5Department of Pathology and Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Vic, Australia
Summary The aim of this study was to compare HER2 amplification, as determined by the HER2 copy number (CN) and the HER2/ CEP17 ratio, with protein expression in gastric and gastrooesophageal junction (G/GOJ) adenocarcinoma. HER2 immunohistochemistry (IHC) and silver in situ hybridisation (SISH) were performed in 185 cases. Modified gastric criteria were used for IHC scoring. HER2 and CEP17 CNs were counted in at least 20 cancer cells and the ratio calculated as per previously defined protocols. These two SISH methods were statistically compared against the different IHC scores. Thirty-four cases showed amplification, by both methods in 29, and either method in five. IHC score was 3þ in 29 cases; 26 showed amplification by both methods, one by ratio only and two were not amplified. IHC score was 2þ in 24 cases; three showed amplification by both methods and two by either. One each of IHC 1þ and 0 showed an increased ratio but not CN. The HER2 CN and ratio for IHC score 3þ compared to scores 2þ, 1þ and 0 were significantly different (all p < 0.01). The CN for IHC 2þ vs IHC 1þ and IHC 0 was significantly different (both p < 0.01) but the ratio was not ( p ¼ 0.5711 and p ¼ 0.2857, respectively). The CN and the ratio for scores 1þ and 0 were not significantly different ( p ¼ 0.9823 and p ¼ 0.9910, respectively). The HER2 CN differentiates between the different IHC scores better than the HER2:CEP17 ratio. Cases that show IHC3þ and high CN may not require calculation of the ratio. Furthermore, consideration should be given to the CN when IHC negative cases appear amplified by the ratio only. Key words: Amplification, copy number, double probe method, gastric/gastroesophageal junctional carcinoma, HER2 status, ratio, single probe method. Received 17 June, revised 29 October, accepted 7 November 2013
INTRODUCTION HER2 positive status determines the eligibility for HER2 targeted therapy of advanced and metastatic gastric and gastro-oesophageal junction (G/GOJ) adenocarcinomas. The HER2 status can be determined by estimation of protein expression by immunohistochemistry (IHC) and/or assessment of HER2 gene copy number and centromeric probe 17 (CEP17)
Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000075
#
ratio by in situ hybridisation (ISH). ISH can be performed by bright field techniques such as chromogenic ISH (CISH) or silver ISH (SISH) or by dark field methods using fluorescence (FISH). Although there is evidence that bright field methods are superior to FISH in determining the gene amplification for G/GOJ cancers,1–3 there is no universal acceptance of a method for testing at this point.4,5–7 The Trastuzumab for Gastric Cancer (ToGA) study defined ISH positive status for G/GOJ cancer as HER2:CEP17 ratio 2.0 irrespective of the copy number.4 This criterion has been used for clinical testing and in most studies related to the HER2 status of gastric/ GOJ cancer. The majority of studies show a perfect or near perfect correlation between IHC3þ expression and gene amplification.4,8–10 However, they also report that a proportion of equivocal and negative IHC cases are ISH positive when the ratio is used. IHC is a semi-quantitative test that is subject to a variety of pre analytical and analytical issues involving tissue sampling, scoring criteria and inherent subjectivity of IHC interpretation. In contrast, ISH technique is considered superior, being more quantitative and reproducible. There are specific issues confounding HER2 testing in G/GOJ cancers by either method, such as the type of specimen used (i.e., endoscopic biopsies, resections or metastatic material) and a greater degree of heterogeneity reported in these cancers.2,8 Considering the special issues, a modified IHC scoring system was established for HER2 testing of G/GOJ cancers.9 Superiority of the modified/gastric cancer scoring system has been validated by others subsequently.4,10 A positive IHC reaction in G/GOJ cancers includes basolateral/lateral membrane staining as opposed to the requirement of complete membrane staining in breast carcinomas. Additionally the cut-off for a positive test is only 10% positive tumour cells for resections and a cluster of five positive tumour cells for endoscopic biopsies.9,10 There are two methods to establish gene amplification. One is the single probe method by counting the actual HER2 copy number (CN) per nucleus, the cut-off for amplification being 6.0. The other is using dual probes to calculate the ratio of HER2 genes to centromere 17 (HER2/CEP17), the cut-off for amplification being 2.0. The ratio distinguishes increased HER2 gene copy number secondary to extra copies of CEP17 that may occur due to true polysomy or
2014 Royal College of Pathologists of Australasia
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
HER2 AMPLIFICATION BY COPY NUMBER OR RATIO?
co-amplification. For the ToGA study, HER2 positive status was defined by either protein expression that showed a IHC 3þ score, or FISH gene amplification defined as a HER2/CEP17 ratio 2.0, not a HER2 CN 6.0. Significantly, subgroup analysis showed that there was no survival benefit with trastuzumab therapy for the HER2 amplified but IHC negative group (score 0 or 1þ) which accounted for 23% of the positive cohort. The exact reason for this has not been established to date. A recent review on the issue of HER2 testing in gastric cancer by an expert panel2 recommends taking into consideration the gene count, i.e., CN when the ratio suggests borderline amplification (a ratio close to 2.0). Furthermore, the GaTHER study conducted in Australia concluded that the inter-laboratory agreement on CISH/SISH scoring was good/ very good when HER2 copy number was used (k ¼ 0.68 to 0.86), but was reduced when HER2:CEP17 ratio was used.1 In spite of the above, there are no studies that have compared the performance of the ratio versus the HER2 copy number across the range of IHC. In our experience with HER2 testing of gastric/GOJ cancers, we have encountered a few cases that just reach the cut-off ratio of 2 whilst the actual HER2 copy number does not exceed 6. Therefore, we embarked on a detailed analysis of the HER2 copy numbers and HER2/ CEP17 ratio across the range of IHC scores, also comparing the results against tumour type/grade, and taking into account the potential impact on clinical testing.
METHODS Cases included in the study were: endoscopic biopsies (n ¼ 146, 80.7%), selected blocks of resections (n ¼ 30, 16.6%) and metastatic deposits (n ¼ 9, 2.7%) of G/GOJ adenocarcinomas with an adequate amount of tumour. HER2 status was assessed by IHC and SISH. IHC scoring (0–3þ) was performed using ‘modified gastric’ criteria. HER2 and CEP17 CNs were counted in at least in 20 cancer cells and the HER2:CEP17 ratio calculated as per standard protocol.1,4,9 These two methods were statistically compared across the range of IHC scores, as well as the tumour grade and type. Immunohistochemistry HER2 IHC was performed on formalin fixed, paraffin embedded tissue using the Ventana XT automated stainer (Ventana, USA), with polyclonal CERB2 antibody (Dako, Denmark) applied at 1:4000 dilution. The resection and/or biopsy specimens were scored for HER2 overexpression according to criteria by Hofman et al.9 Silver in situ hybridisation (SISH) HER2 amplification was performed on formalin fixed, paraffin embedded tissue using the automated Ventana INFORM HER2 Genomic probe platform. HER2 and CEP17 copy numbers were counted and the CN was averaged per cell. In each case HER2/CEP17 ratio was assessed. The cut-off for amplification by the copy number was 6 and by the ratio was 2 as per established guidelines.1,4,9 Table 1
185
Statistical analysis Skewed variables were logarithmically transformed where appropriate. Data among groups were compared using general linear models (SAS Proc GLM; SAS Institute, USA). All pair-wise comparisons and their corresponding p values for the test of no difference are reported. Tukey–Kramer test was applied to account for multiple comparisons for a given variable across the four IHC scores. The statistical significance was set at the 5% level.
RESULTS There were 185 G/GOJ carcinomas in which IHC and HER2 and CEP17 SISH had been performed. IHC scores and SISH results of CN and ratio A total of 34 cases demonstrated an increase in either CN or ratio (n ¼ 5, CN only in 1 and ratio only in 4) or both (n ¼ 29). Table 1 shows the comparison of the HER2 CN, HER2/CEP17 ratio and IHC score. Table 2 shows the raw figures of HER2 and CEP17 signals, ratio and the respective IHC score in discordant cases. Of 29 IHC 3þ cases, 26 showed amplification by both ratio and CN, one case by ratio only, and two cases were not amplified. Of the 24 IHC 2þ cases, three were amplified by both CN and ratio, with two additional cases amplified by either CN or ratio respectively, and 19 cases were not amplified. One each of IHC 1þ and 0 showed increased ratio but not CN. Mean CN for IHC 3þ cases was markedly elevated above 6 (18.92), while all other scores showed a mean copy number that was <6. Fig. 1 shows a concordant case (IHC 3þ, copy number 6 and ratio 2) and a discordant case (IHC 1, copy number <6 but ratio 2). Statistical comparison of ratio and copy numbers across IHC scores Table 3 shows the comparisons of the ratio and the CN across IHC scores and their corresponding p values. The HER2 CN and ratio for IHC score 3þ versus scores 2þ, 1þ and 0 were significantly different from each other (all p < 0.01). The CN for IHC 2þ versus IHC 1þ and IHC 0 was significantly different from each other (both p < 0.01) but the ratio was not ( p ¼ 0.5711 and p ¼ 0.2857, respectively). The CN and the ratio for scores 1 and 0 were not significantly different ( p ¼ 0.9823 and p ¼ 0.9910, respectively). HER2 status by definitions used across the world Using current ToGA (IHC 3þ and/or any ISH positivity by a HER2/CEP17 ratio of 2.0), European Medicines Agency [(EMEA); all IHC 3þ and IHC 2 and ISH positive by ratio of 2.0] and Belgian (all SISH positive by ratio of 2.0 irrespective of IHC) criteria, 36 (19%), 34 (18%) and 33 (17%) out of 185 cases would be positive, respectively. If the copy number was used, positivity would reduce to 34 (18%), 31 (17%), and 30 (16%), respectively. By using recently
Comparison of IHC score, HER2 CN and HER2/CEP17 ratio
IHC
SISHþ (ratio)
SISHþ (CN)
SISH– (ratio)
SISH– (CN)
Total
Mean CN
Mean ratio
0 1þ 2þ 3þ Total
1 1 4 27 33
0 0 4 26 30
98 32 20 2 152
99 33 20 3 155
99 33 24 29 185
2.48 2.57 4.14 18.92
1.15 1.26 1.94 9.57
CN, copy number; IHC, immunohistochemistry; SISH, silver in situ hybridisation.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
186
KUMARASINGHE et al.
Pathology (2014), 46(3), April
Table 2 Distribution of HER2 and CEP17 counts, ratio and IHC score of the discordant cases (n ¼ 5)
Case no. 64 110 2 137 163
HER2 signal/cell
CEP17 copy signals/cell
HER2/CEP17 ratio
IHC score
3.0 5.1 6.0 5.5 5.2
1.5 1.9 4.2 2.0 2.1
2.0 2.6 1.4 2.8 2.5
0 1 2 2 3
Table 3 numbers
Statistical comparison between IHC scores versus ratio and copy
IHC score p values HER2 signals per cell Ratio of HER2 to CEP17
0 1 2 0 1 2
vs vs vs vs vs vs
1: 2: 3: 1: 2: 3:
p ¼ 0.9823 p < 0.01 p < 0.01 p ¼ 0.9910 p ¼ 0.5711 p < 0.01
0 vs 2: p < 0.01 1 vs 3: p < 0.01
0 vs 3: p < 0.01
0 vs 2: p ¼ 0.2857 1 vs 3: p < 0.01
0 vs 3: p < 0.01
IHC, immunohistochemistry.
IHC, immunohistochemistry.
updated Australian guidelines (IHC 2þ or IHC 3þ plus gene copy >6 plus ratio of >2) the positive rate would be only 15% (29/185). The four of five discrepant cases that only showed an increased ratio were amplified at very low levels of 2.0, 2.6, 2.8 and 2.5 (Table 2).
as lapatinib as well as trials to determine the best partner chemotherapeutic agents in the management of G/GOJ adenocarcinomas.12,13 Protein overexpression is triggered by gene amplification and is manifested by increased copies of the HER2 gene. Hence protein overexpression is expected if there is significant gene amplification, although there may be other mechanisms that may interfere with this.14 Post-hoc subgroup analysis of the ToGA series indicates that positive treatment effects are limited to targeting tumours with higher levels of HER2 protein expression (IHC 2þ/FISHþ and IHC 3þ/regardless of the FISH status) compared to the those with low level of protein expression even if the FISH result was positive, i.e., amplified. EMEA subsequently defined HER2 positivity as an IHC 3þ reaction or IHC 2þ with a positive ISH result to be eligible for trastuzumab therapy. A positive ISH (FISH/SISH/CISH) result was defined by the ratio 2.0 irrespective of the copy number count. A recent review2 on the issue of HER2 testing in gastric cancer by an expert panel recommends taking into consideration the gene count when the ratio suggests borderline amplification (ratio close to 2.0). The GaTHER study demonstrated that gene amplification assessment based on HER2 CN
Comparison of HER2 copy number and ratio across severity of tumour grading and type Moderately differentiated (G2) carcinomas showed the highest mean CN (7.54) and ratio (3.76) . ‘Intestinal’ or gland forming adenocarcinomas showed a higher proportion of HER2 positivity (13/54, 24%), compared to the ‘diffuse ‘type (1/11, 9%) ( p ¼ 0.26, for both CN and ratio).
DISCUSSION Trastuzumanb is a monoclonal antibody that acts against the HER2 protein when it is overexpressed on the cell membrane blocking the downstream effects.11 Trastuzumab is currently the only approved targeted agent for G/GOJ adenocarcinomas following the results of the ToGA trial.4 However, there are several ongoing studies on other HER2 targeting agents such
A
B
C
D
E
F
Fig. 1 (A) HER2 immunohistochemistry (IHC) in carcinoma cells with no membrane positivity (IHC 1). (B) Corresponding HER2 silver in situ hybridisation (SISH) featuring an average CN of count of 5.1. (C) Corresponding CEP17 featuring an average count of 1.9 resulting in a ratio of 2.6. (D) HER2 immunohistochemistry in carcinoma cells with strong positive membrane staining (IHC 3). (E) Corresponding HER2 SISH featuring large and small clots with an average count of 18. (F) Corresponding CEP17 featuring an average count of 1.8 resulting in a ratio of 10.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
HER2 AMPLIFICATION BY COPY NUMBER OR RATIO?
was more reproducible than assessment based on the HER2/ CEP17 ratio, concluding that single-probe CISH or SISH is the optimal method for HER2 ISH testing in G/GOJ carcinomas in Australia.1 In this study, we provide evidence that HER2 copy number differentiated between IHC scores better than the ratio. This finding also suggests that (1) consideration should be given to the copy number when IHC negative cases show low level amplification by ratio only, and (2) calculation of the ratio is not necessary for cases that are IHC 3þ with a high copy number. Considering the costs and side effects of treatment, all factors that could potentially affect a positive result need to be evaluated carefully. Treatment failure due a to false positive result will have a negative impact on the ongoing advances in the management of upper GI adenocarcinomas with targeted therapy.4,12,13 The experience of the ToGA study that showed a lack of survival benefit in 23% of the positive cohort where the HER2 gene was amplified (by ratio) but IHC was negative (score 0 or 1þ) is also in support of caution. In our series, a very high proportion of cases that showed gene amplification both by the single probe (HER2 copy number) and the dual probe (HER2/CEP17 ratio) were IHC 3þ. The mean copy number for these IHC 3þ cases was 18.92, an unequivocal increase that is unlikely to be misinterpreted. Often these highly amplified cases show small or large clumps of gene copy number signals that are easily visible. In parallel to this, the GaTHER study showed very good inter-laboratory agreement for HER2 copy number for IHC 3þ cases. There may be no added value in the dual probe method for cases that are IHC 3þ with such an unequivocal increase in the copy number. In contrast, the mean copy number of IHC 2þ (equivocal), IHC 1þ (negative) and IHC 0 (negative) were 4.54, 2.67 and 2.54, respectively, and the difference between IHC 3þ and others was statistically significant according to our results. The four additional cases positive by ratio only showed very low amplification of the ratios, between 2.0 and 2.8 (2 cases were IHC negative, 1 case was equivocal and 1 was positive). Interestingly, Ru¨schoff and others recently reported low level of amplification (defined as ratio of 2–3) in most of their IHC negative and equivocal cases (76%), whereas only 16% of their IHC 3þ cases showed low levels of amplification.15 Our findings are similar. Currently there are no data comparing the outcomes of low and high level amplification irrespective of IHC results. Considering the technical and practical issues of IHC testing, unresolved issues of SISH testing, limited evidence of clinical outcomes, and special issues of G/GOJ carcinoma such as heterogeneity and variability of samples that may be tested, it would be prudent to rely on more than one method of testing as well as interpretation. Modifications and updated guidelines should continue to evolve as more experience is gained in clinical testing. The reasons for the discrepancy between the actual copy numbers and the ratio may be a result of several factors. Suboptimal staining of the CEP17 probe was noted in some samples, which could result in artificial elevation of HER2: CEP17 ratio. As stated above, this issue will not be a significant problem for cases with high/unequivocal amplification. Another confounding issue may be overlapping of nuclei that may be significant in gastric carcinomas, especially when testing is done on endoscopic biopsies. Technically, overlapping nuclei should not be counted, however, moderate to poorly differentiated gastric adenocarcinomas typically show intricate
187
glandular structures with marked nuclear overlapping and aneuploidy. Theoretically, loss of CEP17 signals may also lead to an artificially elevated ratio. For all these reasons when an IHC negative or equivocal sample shows a ratio 2, the result needs to be reassessed and the copy number considered. A second count by the same assessor or a second assessor and/ or repeat staining may be necessary. In conclusion, our findings show that the HER2 CN differentiated between IHC scores better than the ratio. Consideration should be given to the copy number when IHC negative cases appear amplified by the ratio only. IHC 3þ cases with high copy numbers do not need assessment of the ratio. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Professor M. P. Kumarasinghe, Department of Anatomical Pathology, PathWest, J Block, Queen Elizabeth II Medical Centre, Hospital Avenue, Nedlands, WA 6009, Australia. E-mail: [email protected]
References 1. Fox SB, Kumarasinghe MP, Armes JE, et al. Gastric HER2 Testing Study (GaTHER): an evaluation of gastric/gastroesophageal junction cancer testing accuracy in Australia. Am J Surg Pathol 2012; 36: 577–82. 2. Ru¨schoff J, Hanna W, Bilous M, et al. HER2 testing in gastric cancer: a practical approach. Mod Pathol 2012; 25: 637–50. 3. Powell WC, Zielinski D, Ranger-Moore J, et al. Determining the HER2 status in gastric cancer: a method comparison study of two patient cohorts. Gastrointestinal Cancers Symposium Orlando, USA, 2010; Abstract 17. 4. Bang Y-J, Van Cutsem E, Feyeereislova A, et al. Tratuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet 2010; 376: 687–97. 5. Jouret-Mourin A, Hoorens A, De Hertogh G, et al. Analysis of HER2 expression and gene amplification in adenocarcinoma of the stomach and the gastro-oesophageal junction: rationale for the Belgian way of working. Acta Gastroenterol Belg 2012; 75: 9–13. 6. European Medicines Agency. Assessment Report for Herceptin. Procedure No. EMEA/H/C/278/II/0047. London: European Medicines Agency, 2010. http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_ Assessment_Report_-_Variation/human/000278/WC500074921.pdf 7. National Cancer Institute. FDA approval for trastuzumab: HER2overexpressing metastatic gastric or gastroesophageal (GE) junction adenocarcinoma. Cited Jan 2013. http://www.cancer.gov/cancertopics/ druginfo/fda-trastuzumab 8. Lee S, de Boer WB, Fermoyle S, et al. HER-2 testing in gastric carcinoma: issues related to heterogeneity in biopsies and resections. Histopathology 2011; 59: 832–40. 9. Hofmann M, Stoss O, Shi D, et al. Assessment of a HER-2 scoring system for gastric cancer: results from a validating study. Histopathology 2008; 52: 797–805. 10. Park YS, Hwang HS, Park HJ, et al. Comprehensive analysis of HER2 expression and gene amplification in gastric cancers using immunohistochemistry and in situ hybridization: which scoring system should we use? Hum Pathol 2012; 43: 413–22. 11. Gajria D, Chandarlapaty S. HER2-amplified breast cancer: mechanisms of trastuzumab resistance and novel targeted therapies. Expert Rev Anticancer Ther 2011; 11: 263–75. 12. Hoffmann-La Roche. A study of capecitabine (Xeloda) in combination with trastuzumab (Herceptin) and oxaliplatine in patients with resectable gastric cancer. Cited Jan 2013. http://clinicaltrials.gov/show/NCT01130337 13. Iqbal S, Goldman B, Fenoglio-Preiser CM, et al. Southwest oncology group study S0413: a phase II trial of lapatinib (GW572016) as first-line therapy in patients with advanced metastatic gastric cancer. Ann Oncol 2011; 22: 2610–5. 14. Kanta SY, Yamane T, Dobashi Y, et al. Topoisomerase IIa gene amplification in gastric carcinomas: correlation with the HER2 gene. An immunohistochemical, immunoblotting, and multicolor fluorescence in situ hybridization study. Hum Pathol 2006; 37: 1333–43. 15. Ru¨schoff J, Dietel M, Baretton G, et al. HER2 diagnostics in gastric cancer – guideline validation and development of standardized immunohistochemical testing. Virchows Arch 2010; 457: 299–307.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.