Standardisation of HER2 analysis in carcinoma: Breast Cancer Lessons learned, lessons forgotten:

Transcription

1 6/7/ Standardisation of HER analysis in carcinoma: Breast Cancer Lessons learned, lessons forgotten: Dr John. M.S. Bartlett, Director of Transformative Pathology OICR HER testing in breast cancer. The rationale for HER testing. Which test is best? FISH or IHC?. Chromosome 7. Can I use CISH please?. Rogues! What is HER? HER receptor dimer transmembrane signal transduction pathway Epidermal Growth factor receptor homologue Type receptor tyrosine kinase (RTK) family member Extracellular ligand binding domain Transmembrane dimerisation domain Intracellular tyrosine kinase doman. 8 Kd protein. Plasma membrane Cytoplasm Growth factor Signal transduction to nucleus Binding site Tyrosine kinase activity Nucleus Gene activation CELL DIVISION HER-/neu gene amplification & expression Addition of Herceptin significantly improves overall survival Slamon et al. Science 89 Southern Northern Western IHC /9/98 (vol 7, tab 8, p977-8) al probability Estimated surviva Herceptin Chemotherapy alone p= Time (months) 6

2 6/7/ Herceptin Clinical Trial Program in Breast Cancer HER cloned mumab D Phase I Phase II Phase III End of Phase II Meeting Filed Herceptin in advanced Breast cancer 998 Herceptin in early breast cancer Herceptin in gastric cancer Lapatinib, Cetuximab, TDM etc etc. Accurate determination of HER status is now pivotal to the management of early breast and gastric cancers 7 L o c a l Error prone HER testing: Central Herceptest IHC FISH+ 7 6.% discrepancy between local (Mixed IHC methods) and central IHC, 6% with central FISH. Roche et al, JNCI 9:8-7 n = 9 IHC Errors: 9% small vs central.6% large vs central.8% Central vs FISH Central testing is not % accurate FISH errors.% small vs central.6% large vs central % Central vs NSABP Paik et al, JNCI 9: 8-. ASCO Guidelines Jan 7 Approximately % of current HER testing may be inaccurate. Such a disorganised practice and high rate of inaccuracy is not acceptable Wolff AC et al, Journal of Clinical Oncology :8-7 Suggests one in every five laboratory tests is incorrect! 9 Accuracy is paramount: A precise definition of accuracy is how close the measured values are to a supposed true value Which system most accurately determines HER status? Which is least error prone in routine clinical practice? Accurate determination of HER status must not be viewed exclusively in terms of benefit from antiher therapy upstream or downstream anomalies that render the interaction with trastuzumab ineffective. Wolff AC et al, Journal of Clinical Oncology :8-7 What is the most accurate assay for detection of HER overexpression? Accuracy, reproducibility, objectivity, response prediction. Concordance? FISH vs IHC vs FISH vs IHC vs? Based on faulty assumptions. External gold standard : FISH vs standard & IHC vs standard. Bartlett et al, J Pathology ; 9:- 8. Press MF, et al Journal of Clinical Oncology ;:9-. Concordance does not equal accuracy! Results test A If TEST A = 9% accurate & TEST B = 9% accurate Concordance will be 8%. If TEST A = % accurate & TEST B = 8% accurate Concordance will be 8%. Concordance does not equate to accuracy. Results Test B Concordant results Ca 8% of HER FISH/IHC results are concordant

3 6/7/ What is the most accurate assay for detection of HER overexpression? A precise definition of accuracy is how close the measured values are to a supposed true value Studies with External gold standard are rare: ae FISH vs standard & IHC vs standard. Bartlett et al, J Pathology ; 9:-8. Press MF, et al Journal of Clinical Oncology ;:9-. entage Perce HER in breast cancer: QIHC populations Unamplified Amplified. HER x normal Two separate populations of breast cancer wrto HER Robertson et al, Can Res 6:8-, 996. Overexpressing cancers are gene amplified: Tubbs RR et al, JCO 9:7-,, Press MF et al, CCR : , Precision and Accuracy HER expression FISH Herceptest /7/ HER Expression John Bartlett Relationship between HER protein expression measured by QIHC and results of diagnostic tests: Pathvysion precision =.9 (prediction over entire range) Accuracy = 9.7% Herceptest precision =.8 Accuracy = 87.% Bartlett et al, J Path., 9:-8. Scorer B Scorer Inter-Observer variation: FISH Scoring Scorer A IHC Scoring Scorer FISH Scoring Inter observer variation in IHC: κ =.67 Bartlett et al, J Path., 9:-8. Helin HJ et al Cancer 989; 6: van Diest PJ et al J Clin Pathol 997; : 8. Bartlett J, J Pharmacogenomics, :-, It s easier to count dots (FISH): κ =.97 Press M.F. et al, CCR : , Bartlett et al, J Path., 9:-8. BJC 998; 77:9 8.BJU Int 999; 8: Accuracy of HER tests Herceptest CB Pathvysion Precision Kappa Consensus 8% 6% 99.% Accuracy 87.% 8.8% 9.% When compared with analysis of true values for HER estimated on frozen tissues, FISH is a more accurate and reproducible diagnostic assay. Bartlett et al, J Path., 9:-8. Press MF, et al JCO ;:9-. If true this should be borne out in real world testing. Real world HER testing: Can accuracy be maintained in real world diagnostic laboratories? Data from central versus local laboratories? Data from External Quality Assurance schemes (UK NEQAS/US CAP) Independent d assessment of performance Standardised samples Assess performance over time and track changes Improvements/declines in performance Offer corrective action. 8

4 6/7/ Central laboratories: Methods matter? Local Test N Sens Spec Accuracy Local Test N Central (Fisher s exact test P <., All IHC Herceptest vs Herceptest 6 8.6% (.77-.8) FISH, local vs ( ) central). Herceptest (.76-.8) IHC 66 7.% (7.-78.) Ventana FISH 8 88.% (.7-.8) (8.6-9.) FISH ( ) Perez et al, JCO :-8 6 Press M.F. et al, CCR : , Choice of method explains 6.-% of errors Size of lab may play a role Choice of method has a critical impact on accuracy high levels of inaccuracy may persist. 9 External Quality Assurance Schemes Real world testing involves multiple hospitals/laboratories. Data from EQA schemes covers large numbers of centres, individually assessed: Can identify optimal and sub-optimal performance and correct it. UK NEQAS Bartlett et al, J Clin Path 6 webpub 8/9/6 (FISH data) US CAP - proficiency testing survey. Tubbs et al, Archives of Pathology & Laboratory Medicine 6:8-8, More on NEQAS tomorrow. Chromosome 7: is dual colour FISH/ISH important? Watters, Going Cooke & Bartlett, BCRT 77:9- What is CEP7 abnormality? How to define on tissue sections? True frequency and impact on HER diagnosis? Does HER copy reflect amplification? Copy false pos/negs Chr 7 HER Ratio Copy false pos/negs Copy false pos/negs Chr 7 HER Ratio Chr 7 HER Ratio

5 /7/ Chr 7 induced errors in HER copy only ISH Chromosome 7 in breast cancer (n = ) 6 Chr 7 < >< > Isola et al Clinical Cancer Research : : HER COPY % % % % % % % % % % % < < < < <6 <7 < < < > % 9% 8% 7% 6% % % % % % % % % Percent HER amplified by copy % 9% 68% 8% 98% 97% % % < < < < <6 <7 < < < > FISH Positivity ratio vs copy 9% 8% 7% 6% % % % 9.7% 9.% 7.% %.%.% % % FISH Copy Copy. Copy Copy 6 Neg Pos Under Over in 6. cases Misdiagnosed >6 p.a..%.9% 6.% Importance of reflux 7.%.9% 6.%.%.% 6.7% 7.%.% FISH Copy Copy. Copy Copy 6 7.6%

6 6/7/ HER copy and amplification % of cases with HER copies are gene amplified monosomy chromosome 7 9% of cases with 7 HER copies are NOT amplified Aneusomy or abnormal chromosome % of cases are misdiagnosed by excluding chromosome 7. In most cases this causes overtreatment with Herceptin. Acceptability of chromogenic ISH for HER IHC equivocal cases: John. M.S. Bartlett, Director of Transformative Pathology Ontario Institute for Cancer Research Chromosome 7 must be included for at least % of breast cancers. Chromogenic ISH Chromogenic ISH multiple chromogens: Silver SISH Gold GOLDFISH CISH Red/Green etc All are chromogenic assays. Key papers:.a UK NEQAS ISH multicentre ring study using the Ventana HER dual-color ISH assay: AJCP :7-6.A multicenter study comparing silver in situ hybridisation with FISH AJCP 9 :-.. Accuracy of HER status CISH has the advantage of not determination on breast coreneedle biopsies (IHC, FISH, CISH requiring fluorescence and SISH vs FISH) Modern microscopy, Pathology 6 th Jan p-8. CISH would have to.determination of HER demonstrate a similar high level amplification in primary breast of accuracy and interlaboratory cancer using dual-colour chromogenic in situ hybridisation is reproducibility in order to comparable to fluorescence in situ compete effectively with FISH as hybridisation: a European a diagnostic assay. multicentre study involving 68 specimens. Histopathology 6:7-8. Summary Ventana Silver CISH Ring study. Overall concordance = 96.% cases per lab 6 laboratories = 7 cases One lab failed 9% concordance threshold. Intraobserver variation: HER CISH=.9% (all labs) FISH=.% ( lab) RATIO CISH 8.% (all labs) FISH=.% ( lab) Ventana DUO-SISH One slide assay: DAKO duocish method: (Histopath ) DuoCISH N =67, concordance %. Overall concordance = 98.% by core 96.6% cases per lab laboratories = cases, cores Intraobserver variation: HER CISH=7.6% (all labs) RATIO CISH 7.7% (all labs) FISH=.9% ( lab) CISH applied to cores and resection specimens: Concordance 98%+ Modern Pathology :

7 6/7/ NEQAS ISH Pass rates FISH vs CISH PASS FISH n = CISH n = 8* Data over 6 runs. FISH CISH 7 CISH versus FISH Extremely high agreement between current dual colour CISH and FISH methods. Enzymatic development can lead to over/under development. Play extremely close attention to internal QC signals in normal cells. EQA data suggests use of CISH appropriate Some challenges exist see NEQAS talk tomorrow. Recommendation following careful central validation use with caution. 8 ROGUE Cases and HER HETEROGENEITY ca 9% of FISH/ISH cases are simple to interpret. There are a subset of challenging cases Collectively around % of all cases. These are rogues Clinical evidence for the treatment of these rogues is lacking due to their rarity. In order to make treatment recommendations there is a need to map these cases to what is known Further research is needed for some groups Starczynski et al American Journal of Clinical Pathology ; 7: 9-6 HER Rogues Audit of UK Birmingham Heartlands Hospital HER ISH cases 8-9: Referred due to IHC+ or other unusual staining patterns. 787 cases audited 7 (9.6%) exhibited unusual ISH staining patterns. Examples isolated reviewed by 9 UK NEQAS ISH reference centres (9) Suggested reporting and research objectives.,,,, Distribution of common rogues Percent Rogues Het AMP CEP7 loss HER loss CoAMP 7

8 6/7/ Simple Heterogeneity discrete amplified areas Either in close proximity or in different fields. CAP Guidelines on Heterogeneity HER genetic heterogeneity (GH) exists if there are more than % but less than % of infiltrating tumor cells with a ratio higher than.. For example, if cells are counted and at least one cell is identified with a HER/CEP7 signal ratio higher than., this specimen contains GH. Genetic Heterogeneity in HER Testing in Breast Cancer: Panel Summary and Guidelines Vance et al: Arch Pathol Lab Med Vol, April 9 What proportion of cases would then be defined as exhibiting heterogeneous amplification? Is this a valid definition? Do you use it? What impact is there on either outcome or Herceptin response? THERE IS NO DATA TO SUPPORT THIS GUIDELINE PERFORM AUDIT linked to OUTCOME. Bartlett AI, Starcyznski J, Robson T et al. Heterogeneous HER Gene Amplification. American Journal of Clinical Pathology ; 6: HER HETEROGENEITY AUDIT (CASES). Cases referred to UK Birmingham Heartlands Hospital over year + and equivocal IHC - eligible cases. Cases referred to UK Glasgow reference centre 67 eligible cases IHC -+. International TEAM pathology Study cases (with outcome) 7 eligible cases Total cases included in audit: 66 RESULTS ALL CASES Of 66 eligible cases 9 (.%) were amplified as defined by UK guidelines (HER/CEP7 ratio >.). Of these would be defined as borderline by the 7 ASCO/CAP guidelines (ratio between.-.). ) 66 cases (.%) exhibited between - % of cells with a HER/CEP7 ratio of greater than. and would under new CAP guidelines be regarded as exhibiting Heterogeneous amplification. The CAP guidelines identified over % of cases as either amplified or with heterogeneous amplification. 6 vival Disease Free Surv TEAM PATHOLOGY OUTCOME.6 Years Solid black line <% cells >. (n = 8) Green line -9.9% (n = 78) Blue line -9.9% (n= 97) Magenta line (N=9) Red line % (N=9) Dotted black line % (N=). 7 Conclusion: Using the new CAP panel guidelines for HER heterogeneous amplification in an audit of 9 cases identified 9 (.7%) cases of heterogeneous amplification, and raised the frequency of HER amplification/heterogeneous amplification to % of cases evaluated. Heterogeneous amplification of the HER oncogene is a real and challenging diagnostic finding. Evidence relating to the prognostic impact and in particular response to HER therapies is currently lacking for these cases. Guidelines should reflect this and seek to gather such evidence before implementing changes to diagnostic practice. Bartlett AI, Starcyznski J, Robson T et al. Heterogeneous HER Gene Amplification. American Journal of Clinical Pathology ; 6:

9 6/7/ The entire slide should be scanned before counting areas of apparent heterogeneity should be identified during this scan and/or by reference to an IHC stained slide. The number of chromosome 7 (CEP7) and HER signals should be counted in to 6 non-overlapping invasive cancer cell nuclei using at least three distinct tumour fields. If there is evidence of heterogeneity between fields (or less frequently within fields) additional cells (at least per field) and/or fields (up to 6) should be counted. The HER/CEP7 ratio should be calculated for each field individually. Where the mean HER/CEP7 ratio in any field is. or greater, the tumour should be regarded as amplified. Cases containing both amplified and nonamplified fields (simple heterogeneity) should be reported as exhibiting heterogeneous amplification. For all cases where the ratio is between.8-. results should be based on counting at least 6 tumour cells, and in cases where heterogeneity is suspected this should be 6 cells per field. In rare cases where amplified and non-amplified tumour cells are intermingled in a single field, interpretation is difficult and evidence is lacking. We suggest that for such cases only the presence of clearly l amplified cells, with multiple l HER signals, is considered evidence of heterogeneity. Current evidence does not support using the existence of small numbers of apparently amplified cells within an individual tumor field to identify heterogeneous amplification. Bartlett JMS, Starczynski J, Atkey N et al. HER testing in the UK: recommendations for breast and gastric in-situ hybridisation methods. J Clin Path ; 6: ca 9% of FISH/ISH cases are simple to interpret. How should we report cases with? Different patterns of heterogeneity It s notalwayssimple simple Loss of either CEP7 or HER signals co-amplification NB guidelines are opinion led and identify key research questions is there a right answer? Starczynski et al American Journal of Clinical Pathology ; 7: 9-6,,,, Distribution of common rogues Percent Rogues Het AMP CEP7 loss HER loss CoAMP Guiding principles:. For unusual cases there is no likelihood of a robust evidence base that will determine treatment efficacy with HER-directed therapies.. Report what you see: Provide a detailed, adequate description of the ISH/IHC staining pattern.. Reach a conclusion: Is there sufficient evidence of amplification/overexpression of HER? Existing treatments are all validated for HER overexpressing/amplified cases. Recognise HER status is not the final arbiter of clinical decisions Age, ER, PgR, Grade, Nodal status, size, Ki67, comorbidity etc etc all play a part.. Identify ways to improve certainty research recommendations Base funding for OICR is provided by the Government of Ontario through the Ministry of Economic Development and Innovation 9

10 6/7/ Simple Heterogeneity discrete areas Either: close proximity or different fields. Report heterogeneous amplification per UK guidelines. Provide ratio for each area scored. Complex Heterogeneity intermixed amplified and non-amplified cells. 6 Reporting guideline Case B amplified and nonamplified cells appear intermingled across the field. Some cells appear normal and some clearly amplified, but it is impossible to distinguish separate fields. Scoring Recommendation. Score 6 tumor cells across areas, recording results for amplified and nonamplified cells on the same reporting sheet. Interpretation. This tumor contains intermixed breast cancer clones. Reporting: Report as a case exhibiting potential intermixed heterogeneous amplification, Mean HER/CEP7 ratio for all cells counted Also report percentage of amplified cells (with ratios >.). Reporting guideline Case C: Isolated cells with high-level amplification Ensure that the signal is genuine and not a processing artifact Scoring Recommendation. Score without biasing the count toward amplified or non-amplified cells. Interpretation. Tumor may contain isolated cells with HER amplification significance unclear. Current evidence - limited prognostic impact. Reporting: Record isolated cells only if clearly amplified. Report mean HER/CEP7 ratio and base diagnosis on this ratio. This case would be regarded as not amplified for HER. 7 8 Complex Heterogeneity: research recommendation. Loss of CEP7 or HER signals Almost % of cases (% of rogues ) show loss of either CEP7 or HER signals. In most cases this does not impact on diagnosis (amplified or not). For some cases with loss of CEP7 and duplication of HER interpretation may impact treatment. Laser capture: Single-cell analysis might yield information on biology but would not clarify the prognostic impact. 9 6

11 6/7/ Deletion of Chromosome 7? Monosomy 7, borderline HER amplification 6 Amplification with loss of CEP7? HER copy number.7 (s.d..9) CEP7 copy number. Ratio HER/CEP7:.7 a) IF: Loss of non amplified chr7. = True amplification b) IF: Loss of CEP7 signal on duplicated chromosomes = false positive 6 Amplification with loss of CEP7? Interpretation. This is an amplified tumor by HER ratio (HER/CEP7) representing a reduction to monosomy of chromosome 7 with HER amplification on the retained chromosome. Scoring Recommendation. No evidence scoring additional cells will alter diagnosis. Reporting: HER gene amplified - report ratio. Append note of abnormality of CEP7 counts. 6 Complete loss of Chromosome 7: Complete loss of Chromosome 7: Total loss of all CEP7 signals in the tumor cells; note CEP7 in normal cells HER copy number is 6.8 copies/cell (SD,.78). Scoring: Score HER in tumor cells Interpretation. Research suggests copy number is an extremely poor guide of HER gene amplification. Reporting: Report the case as amplified on the basis of HER gene copy number alone (only if 6 copies of HER/cell). Include note to explain the loss of the CEP7 signal in tumor cells. 6 66

12 6/7/ Loss of CEP7 or HER signals Research recommendation HER amplified cases with monosomy 7 may have reduced response to Herceptin Risio et al Oncol Reports (metastatic disease). Need to discriminate between loss of whole chromosome/region and focal loss of CEP7/HER. CGH or MLPA assays would provide insight into specific mechanisms underlying ISH patterns. Outcome data unlikely. HER/CEP7 co-amplification Accounts for ca -% of all ISH cases. High copy number of both HER and CEP7 Reduced HER/CEP7 ratio to around Signals may co-localise HER CEP7 co-amplification T Only HER amplified II HER + one or more contiguous targets amplified, followed by telomeric non- amplified or deleted targets III Signals co-localise : HER copy., CEP7. RATIO HER/CEP.! 69 All targets amplified Complex patterns I IV John Bartlett 6/7/ Interpretation and reporting Interpretation: HER/CEP7 ratio = : high HER/CEP7. Opinion is that these rare cases should be interpreted as amplified. Preliminary evidence suggests patients with extended amplicons may exhibit reduced response (again metastatic disease) Scoring: Score HER and CEP7 signals under single-color filters, taking care to match cells within fields to ensure accurate counting of red and green signals. Counting additional cells is not likely to alter the diagnosis. Reporting: Amplified with centromere co-amplification Based on HER copy and associated balanced CEP7 copy number, plus note co-localization of HER/CEP7 signals in the report. Co-Amplified or not? high HER and CEP7 Non-overlapping discrete signals. HER/CEP7 ratio. mean HER 7. and CEP

13 6/7/ Co-Amplified or not? HER/CEP7 ratio = not amplified HER and CEP7 signals high. HER >6 often, not uniformly, associated with HER amplification. Balanced increase in HER and CEP7 copy numbers unlikely to reflect multiple intact copies of chromosome 7. significant room for error in over or under-interpreting FISH results. No clear evidence (colocalized signals) that HER/CEP7 are coamplified. Difficult to assume this represents true HER amplification. Reporting: Not amplified comment on the high HER and CEP7 copy numbers should be included in the report. Area of greatest discussion/disagreement between experts When is Co-AMP present? + signals, +, etc? What about + discrete signals. Co-amplificaton of CEP7/ HER Research recommendation CGH or MLPA assays would provide insight into specific mechanisms underlying ISH patterns. Outcome data unlikely. All research recommendations require collaborative efforts to identify sufficient cases % of breast cancers are HER amplified Rogues = -% of these cases =.-.% of all breast cancers Test + cases to identify rogues with similar ISH pattern! 7 7 Rogues: Conclusion Rogues present significant challenges to diagnosis of HER status. Further research is essential but very challenging challenging Small numbers of specific case types multiinstitutional collaboration/audit. Guidelines can produce debate and synthesise data More likely to improve consistency of diagnosis than accuracy? 7