Original Title: Performance diagnostique des techniques de détermination du statut HER-2 dans le cancer du sein

Authors' objectives: This report is a systematic review of the literature on the diagnostic performance of IHC, CISH, SISH and real-time PCR and RT-PCR in breast cancer. The gold standard used for the analysis is the FISH which utilizes both the HER-2 and CEP17 probes. The distribution of IHC scores in breast cancer patients is estimated, and the concordance between FISH and CISH results for each IHC score (0, 1+, 2+ and 3+) is presented. The interobserver and interlaboratory variability for each technique is evaluated as well. This report includes an analysis of studies published from January 2000 up to and including November 2007.

Authors' results and conclusions: Nearly three fourths (approximately 73%) of breast cancer patients test negative on IHC (a score of 0 or 1+), and the vast majority (97.4%) of them can be confi rmed by FISH. Asfor patients who test positive on IHC (a score of 3+), only 89.9% test positive on FISH. Thus, it can be said that 10.1% of these cases are discordant. However, the concordancebetween IHC and FISH results is better when the tests are performed at central or reference laboratories as opposed to local laboratories. The diagnostic performance of CISH is very good, but only when the HER-2 and CEP17 probes are used in a complementary manner. Otherwise, both the sensitivity andspecificity of this test fall below 95%. For cases with an equivocal IHC result (a score of 2+), the CISH result varies and is thus less concordant with the FISH result. Only two studies comparing SISH and FISH were identifi ed, one of which took into account the correction for polysomy 17. The use of two probes (HER-2 and CEP17) seems to improve the performance of SISH, but its sensitivity nonetheless remains low. Real-time PCR and RT-PCR have good specifi city but have difficulty detecting FISHpositivecases. Selecting the tumour cells to be analyzed by laser microdissection seems to increase the accuracy of PCR, but the evidence in this assessment is based on only onestudy. The level of agreement among pathologists in interpreting IHC test results varies considerably from study to study and is not always satisfactory, given the issues involved. Although less studied, the interobserver variability in FISH, CISH and SISH seems to be very good. As for the tests’ interlaboratory reproducibility, it is low to average in the case of IHC and seems to be better between laboratories that use the same antibodies. Several studieshave reported, for local laboratories, high proportions (18.4 to 36.4%) of positive IHC results (a score of 3+) that subsequently turned out to be negative when the cases wereretested with IHC at a central or reference laboratory. The concordance of FISH results is good between local and central laboratories and excellent between central laboratories. Results for the interlaboratory reproducibility ofCISH are scarcer but seem to be good, whereas those for SISH and real-time PCR and RT-PCR are not yet available.

Authors' recommendations: The use of IHC as the initial test seems to be a good strategy for rapidly and less expensively identifying patients who are not likely to benefi t from trastuzumab therapy.On the one hand, they account for the vast majority of breast cancer cases, and on the other, negative IHC results tally well with FISH results. However, many IHC-positive cases (3+) turn out to be negative on FISH. Given the variability of IHC results, some of these cases could be false positives due to technicaldiffi culties, problems calibrating the method, or interpretation problems. However, their proportion out of the total number of discordant cases is not known because certain biological mechanisms can also lead to the overexpression of a given protein without there being any gene amplifi cation. At this time, the response to trastuzumab indiscordant cases still appears to be uncertain. However, since the treatment targets the protein, the Canadian consensus guidelines automatically consider eligible for trastuzumab therapy cases that test positive on IHCperformed on a good-quality sample at laboratories with recognized expertise that have internal and external quality assurance measures. To improve diagnostic accuracy, several laboratories in the province perform, for each patient, not one, but two IHC tests with antibodies of different sensitivities. They thenperform ISH on all discordant or equivocal cases. Even if it seems reasonable to think that this algorithm can improve test accuracy, there is still little evidence to support this. Whatever the case, this strategy of performing two rather than one IHC test should not make up for the absence, in Québec, of a formal external quality assurance programspecific to these tests or for the insufficient number of IHC tests performed by some Québec laboratories. Indeed, it appears that several laboratories do not perform tests onthe minimum annual case load stipulated in the Canadian standards to ensure the desired level of competence. The Canadian consensus guidelines recognize CISH as an alternative to FISH, but this technique, when performed with an HER-2 probe alone, does not appear to be reliableenough to replace FISH. In Québec, CISH is performed without a chromosome 17 centromere probe, with the result that a certain number of these tests have to be repeatedwith FISH to confirm the diagnostic accuracy. Furthermore, the Canadian recommendation of using SISH as a possible alternative to FISH seems to be supported more by expert opinions than evidence. The diagnosticperformance reported in the literature is clearly inadequate. Just the same, this technique is in the process of being introduced in Québec. None of the evidence presented in this report supports real-time PCR or RT-PCR as an alternative to FISH in a clinical setting.

Authors' methods: Review

Project Status: Completed

Year Published: 2008

URL for published report: https://www.inesss.qc.ca/publications/repertoire-des-publications/publication/performance-diagnostique-des-techniques-de-determination-du-statut-her-2-dans-le-cancer-du-sein.html

English language abstract: An English language summary is available

Publication Type: Not Assigned

Country: Canada

Organisation Name: Agence d'évaluation des technologies et des modes d'intervention en santé

Contact Address: 2021, avenue Union, Bureau 10.083,Montreal, Quebec H3A S29, Canada.Tel: +1 514 873 2563; Fax: +1 514 873 1369

Contact Name: demande@inesss.qc.ca

Contact Email: demande@inesss.qc.ca

Copyright: Agence d'Evaluation des Technologies et des Modes d'Intervention en Sante (AETMIS)