Determination of Hormone Receptors, Human Epidermal Growth Factor Receptor 2 and Ki67 Status in Invasive Breast Carcinoma: A Concordance Study between Immunohistochemistry, Fluorescence in Situ Hybridization, and GeneXpert® Breast Cancer STRAT4 Assay

The accurate assessment of hormone receptors (ER and PR), HER2 and Ki-67 proliferative index provides meaningful information about breast cancer prognosis and prediction of therapy response. Immunohistochemistry, the most common method for evaluating these prognostic biomarkers, can be impacted by numerous variabilities due to pre-analytical/analytical factors and subjective interpretation by pathologists. The Xpert® Breast Cancer STRAT4, a RT-qPCR based system, can be used to classify breast invasive carcinomas based on the assessment of these 4 biomarkers. In this study, we investigated the accuracy of RT-qPCR based mRNA expression levels in a closed, single-use cartridge, automated system compared with the current gold standard, immunohistochemistry (IHC), and uorescent in situ hybridization (FISH) for HER2 equivocal cases. We evaluated ESR1, PGR, ERBB2 and MKi67 mRNA expression by Xpert Breast Cancer STRAT4 and ER, PR, HER2 and Ki67 by IHC (FISH for HER2 IHC 2+) in 200 formalin-xed paran-embedded (FFPE) tissue blocks with invasive breast cancer, collected from the Pathology Department of Casablanca Ibn Rochd University Hospital. provide an additional, objective, and quantitative assessment of tumor receptor status in breast cancer. : IHC

In spite of long-term use, immunohistochemical assays have not been adequately standardized across labs and IHC and FISH results can be impacted by pre-analytical or analytical limitations, including tissue xation, choice of antibodies, use of manual vs computer assisted scoring methods, and interpretation of results in assay performance, all of which can signi cantly affect the accuracy and reproducibility of results for these four biomarkers [10,12,16,17].
The Xpert® Breast Cancer STRAT4 test is a CE-IVD* test (*In vitro diagnostic medical device. May not be available in all countries. Not available in the U.S.) that offers a semi-quantitative assay with qualitative cut-off values for Estrogen Receptor (ESR1), Progesterone Receptor (PGR), HER2/ERBB2, and Marker of Proliferation Ki-67 (MKi67) mRNAs isolated from FFPE invasive breast cancer samples [16,17].
Xpert® Breast Cancer STRAT4 measures target (ESR1, PGR, ERBB2, and MKi67) and reference gene (CYFIP1) mRNAs isolated from FFPE breast cancer tissue in a self-contained cartridge using the Cepheid® GeneXpert® (GX) System which automates and integrates the incartridge sample processing, including RNA isolation, ampli cation, and detection of the target sequences in FFPE samples using real-time reverse transcriptase, polymerase chain reaction assays (RT-PCR) [16,18].

Methods
Two hundred blocks of FFPE tissue specimens aged ≤5 years archived in the Pathology Department of Casablanca Ibn Rochd University Hospital were included in our study. The histopathology of all samples remaining in the blocks was reviewed, and only specimens still containing invasive breast carcinoma cells were included in the study. This analysis enrolled retrospectively de-identi ed FFPE tissue sections obtained from core biopsies or surgical specimens, from a selection of patients with invasive breast cancer whose tumor samples were collected and routinely evaluated for breast cancer biomarkers (ER, PR, HER2 and Ki67) according to the standard of care (SOC) IHC and/or FISH assays at the Pathology laboratory. The immunohistochemical status (IHC) was sought on 4 µm tissue sections treated and incubated with the antibodies ER : FLEX Monoclonal Rabbit Anti-Human Estrogen Receptor α Clone EP1 Ready-to-Use; PR : FLEX Monoclonal Mouse Anti-Human Progesterone Receptor Clone PgR 636 Ready-to-Use and Ki67 : FLEX Monoclonal Mouse Anti-Human Ki-67 Antigen Clone MIB-1 Ready-to-Use, according to the Dako protocol on the Autostainer Link 48 IHC platform. HER2 status is rst assessed by IHC, using the antibody Ventana Pathway Anti-HER-2/neu (4B5) Rabbit Monoclonal Primary Antibody on a Ventana GX automated platform. Tumors were classi ed as ER positive or PR positive when ≥1 % invasive tumor cells showed de nite nuclear staining, irrespective of staining intensity. A tumor was considered to be HER2 positive if an IHC score equal to 3+ was found and HER2 negative if a score of 0 or 1+ was observed (ASCO/CAP guidelines) [25]. HER2 Equivocal (IHC 2+) results were subsequently tested by FISH with manual technique using the probes (HER2 IQFISH pharmDx) to con rm nal HER2 status.
The patient tumors selected for our study represent the various breast cancer subtypes as determined through surrogate IHC subtyping by the routine assays performed at our laboratory, as follows: The mRNA levels of ESR1, PGR, ERBB2 (HER2), and MKi67 were assessed by quantitative gene expression readouts using the Xpert® Breast Cancer STRAT4 assay. Breast Cancer FFPE tissue samples were prepared for the assay as tissue scrolls (10 µm thickness) and placed into a tube. FFPE samples were rst treated with the recommended volumes of FFPE lysis reagent (1.2 ml) and proteinase K (20 µL) provided by the Xpert® FFPE Lysis Kit (CE-IVD*) prior to use in Xpert® Breast Cancer STRAT4. The solution was then incubated in a heat block at 80°C for 30 minutes. Then 1.2 ml of ≥95% Ethanol was mixed with the sample.
Once the tissue lysate is prepared, a 520 µL aliquot was placed into the appropriate sample chamber in the Xpert® Breast Cancer STRAT4 cartridge. The testing cartridge was inserted into a module of a GeneXpert® System for processing where nucleic acid puri cation, ampli cation, and real-time detection are all fully automated and completely integrated by the system. The nal results of STRAT4 testing are available in approximately 70 minutes after starting the test.

Statistical analysis
Statistical analysis was done in GraphPad Prism Software. For each of the four biomarkers studied, agreement measurements between Xpert® Breast Cancer STRAT4 and IHC and/or FISH, which were considered as the reference methods, were based on contingency table analysis and included overall concordance (overall percent agreement), positive percent agreement (sensitivity) de ned as the number of samples classi ed positive by both IHC and Xpert® Breast Cancer STRAT4 divided by the number of positive samples using immunohistochemistry, negative percent agreement (speci city), and Cohen's κ coe cient scores. The Kappa (κ) statistic numeric values are categorized into slight agreement (≤0.2), fair agreement (between 0.21 and 0.40), moderate agreement (between 0.21 and 0.40), substantial agreement (between 0.61 and 0.80) and almost perfect agreement (between 0.81 and 1.00). All measurements were associated with 95% con dence intervals (95% CI), compared using Fisher's exact test and considered signi cant for P < 0.05 [27].

Results
To assess the concordance between the Xpert® Breast Cancer STRAT4 and IHC+HER2 FISH methods, we used a cohort of 200 specimens of formalin-xed para n-embedded invasive breast carcinomas diagnosed at the Pathology Department of Casablanca Ibn Rochd University Hospital between 2016 and 2020.
For each sample, we evaluated mRNA results by Xpert® Breast Cancer STRAT4 and compared them to the results obtained by the already routinely performed IHC+HER2 FISH .
The Cohen's κ coe cient score was equal to 0.488 (95% con dence interval: From 0.372 to 0.603). Six cases with "indeterminate" Xpert® Breast Cancer STRAT4 PGR results were excluded from this analysis.
The concordance rate obtained when the population was strati ed rst by ER status was 89.04% for HER2 ER+ and 100% for HER2 ER-, including all cases.
The Cohen's κ coe cient was equal to 0.838 (95% con dence interval: From 0.735 to 0.940) when comparing Xpert® Breast Cancer STRAT4 ERBB2 dCt results to HER2 IHC results, excluding equivocal cases (HER2 IHC2+). The Cohen's κ coe cient was equal to 0.387 (95% con dence interval: From 0.073 to 0.700) when we analyzed IHC 2+ cases with Xpert® Breast Cancer STRAT4 ERBB2 dCt and FISH HER2 results. For determination of HER2 status taking into account both reference methods (IHC and FISH) as recommended by ASCO/CAP Guidelines [28] the Kappa coe cient is equal to 0.771 (95% con dence interval: From 0.666 to 0.877). Considering the ER status strati ed subset only for the comparison of Xpert® Breast Cancer STRAT4 ERBB2 dCt and HER2 results by IHC+HER2 FISH, the statistical Kappa was 0.521 (95% con dence interval: From 0.320 to 0.722) for ER-positive subset and 1.00 (95% con dence interval: From 1.000 to 1.000) for ERnegative subset.
Last, we examined the Xpert® Breast Cancer STRAT4 MKi67 dCt values and Ki67 results by using Ki67 IHC cutoff of 10% and 20% to discriminate "high proliferation rate" from "low proliferation rate", while we used an intermediate zone (equivocal results) between 10 and 20% for the MKi67 dCt distribution. We excluded sixteen cases with « Indeterminate » Xpert® Breast Cancer STRAT4 MKi67 status from this study.

Discussion
Immunohistochemical assays are the current gold standard for evaluating ER, PR, HER2 and Ki67 status in breast invasive carcinoma. The main advantages of IHC for the assessment of these markers are that it is rapid and simple, it can be performed in most pathology laboratories, and (when compared with other assays) it is relatively inexpensive. However, IHC assay reliability has been questioned because alterations during tissue processing, manipulation and xation, as well as the antibody clone, internal controls and scoring system used may affect the precision of the results. In addition to that, inter-observer variability in interpretation may play also a role as IHC remains a semiquantitative and non-standardized method [18][19][20].
The ambiguity encountered in the interpretation of HER2 IHC results especially in cases with HER2 equivocal scores (HER2 IHC =2+) may also represent an issue that most laboratories in resource-constrained settings may not be able to overcome as the gold standard would be, in these cases, FISH for quantifying HER-2 gene ampli cation. Indeed, FISH has the advantage of being a quantitative method and is considered as the gold standard method for con rming the HER-2 status, not only to resolve IHC 2+ cases, but also for all other cases where it has an excellent correlation with the HER2 IHC results [31,32] Major disadvantages are that FISH is technically complicated to execute, arduous to establish, has a long run time, and is costly, making it not routinely available in all pathology laboratories worldwide. Moreover, another limitation of this method is that it doesn't necessarily re ect target protein expression and counting FISH spots is wearisome and can be biased by tumor heterogeneity [9,19,20].
Nevertheless, these causes of assay variability may explain the differences in ER, PR, HER2, and Ki67 IHC results in breast carcinomas reported previously.
Currently, treatment of invasive breast carcinoma relies essentially upon ER, PR, HER2 and Ki67 status [33,34] and accuracy of assays is critical. Hence, to overcome limitations of IHC and HER2 FISH, there have been efforts to establish alternative methods to assess the 4 biomarkers of interest as accurately as possible [35]. One of the options is to use RT-qPCR. Reverse transcription quantitative PCR (RT-qPCR) represents a sensitive, e cient, and reliable approach for analyzing RNA. The initial step in RT-PCR is the production of a single-strand complementary DNA copy (cDNA) of the RNA through the action of the retroviral enzyme, reverse transcriptase, to amplify that part of this cDNA by PCR. RT-PCR is used to analyze differential gene expression or cloned cDNAs. The results showed a moderate Kappa correlation agreement for PGR/PR (using PR IHC+ 1%) and MKi67/Ki67 (excluding equivocal cases) between both assays (κ "PR"= 0.565 ; κ "Ki67"= 0.458). Xpert® Breast Cancer STRAT4, however, demonstrated a signi cant overall concordance with IHC for PGR (83.5%) and MKi67 (81%). The concordance rates observed in other studies vary from 81-92% for PR and from 78-89% for Ki67, in accordance with agreement percentages obtained in our analysis [16-18, 20, 38].
Discordance between assay methods can be attributed to several factors, including the tissue xation, antibody clone used in IHC, and scoring methods used. Preanalytical factors are essential to monitor and a quality assessment scheme should be put in place in any laboratory routinely performing the assessment of the 4 biomarkers [39]. Particular attention should be paid to the impact of sample handling, time of xation, duration of tissue xation, antibody selection, control samples and interpretation of assay on Xpert® Breast Cancer STRAT4 results [8, 10,11,14]. In spite of the systematic practice of immunohistochemistry methods, procedural inconsistency remains elevated in clinical settings, leading to interlaboratory and intralaboratory variations and to high false-negative (for ER and PR) and false-positive (for HER2) [15]. This inconsistency emphasizes the importance of a standardized retrieval method in the performance of reliable IHC and/ or FISH for the four markers routinely screened in breast cancer diagnosis. countries. It is a sensitive method that can replace IHC and FISH in remote areas where IHC cannot be performed, because it is a nonoperator -dependent technique and does not require an equipped molecular laboratory [9,11,[16][17][18][19][20]38]. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Competing interests J.W. was an employee of Cepheid at the time of this study.
The funders had no role in the selection of samples used in the study, in the collection of data from participating sites, the interpretation of the nal data analyses in the study, or the decision to publish the results.
The corresponding authors had sole nal responsibility for preparing the original draft manuscript, nalizing data analyses, or interpretation of data, in the writing of the manuscript, and in the decision to publish the results.
The other authors declare that they have no competing interests.

Funding
This study was funded by Cepheid®. All Xpert® Breast Cancer STRAT4 (CE-IVD) kits were provided by Cepheid. Cepheid was not involved in sample selection, nor the nal data analysis.
Authors' contributions RE Realised the technical part, analyzed and interpreted the data and was a major contributor in writing the manuscript. AN Study monitoring and corrected the manuscript. AK selected samples with the pathologist and Contributed to the technical analysis. YZ contributed to the statistical study. MO corrected the manuscript. JW corrected the manuscript. MK the study supervisor, selected eligible cases for the study, reviewed the H&E stained slide, Interpreted the data and Corrected the manuscript. All authors read and approved the nal manuscript.    Comparison of HER2/ERBB2 determined by RT-qPCR and FISH with FISH assessment of IHC2+. Graph of mRNAexpression ERBB2 dCt determined with Xpert® Breast Cancer STRAT4 test by HER2 FISH ampli cation result categorized as negative « Not ampli ed » (<2) or positive « Ampli ed » (≥2).

Figure 5
Comparison of HER2/ERBB2 determined by either RT-qPCR or by immunohistochemistry with or without FISH assessment of IHC2+. Graph of mRNAexpression ERBB2 dCt determined with Xpert® Breast Cancer STRAT4 test by IHC/FISH Her2 results including all sample size.