Throughout all phases of the clinical development of trastuzumab, an IHC assay called the CTA developed by Genentech was used to detect HER2 tumor overexpression in possible eligible study patients with breast cancer. This assay utilized two different mouse monoclonal antibodies, the parent murine antibody to trastuzumab, clone 4D5, and anti HER2, clone CB11. The majority of tumor specimens in the different clinical trials with trastuzumab were tested using the 4D5 antibody. The CTA was a technically challenging IHC assay with more than 35 individual process steps, and it required four or more tissue sections per patient for completion (11).

In the early phase clinical trials with trastuzumab, different patient selection criteria for HER2 overexpression were used (8). However, with the conduct of the pivotal phase II trial in patients with metastatic breast cancer, where trastuzumab was given as monotherapy, the scoring criteria for the HER2 expression more or less found its final form. These scoring criteria were subsequently used throughout the clinical development program (16). In this phase II trial, the IHC staining was performed with either the 4D5 or CB11 antibody, and the HER2 tumor expression was scored as 0, 1+, 2+, or 3+ using standardized scoring criteria, as shown in Table 1 (17). Patients with weak (HER2 IHC2+) or complete (HER2 IHC3+) cell membrane staining of > 10% of the tumor cells were eligible for treatment with trastuzumab and were enrolled in the clinical trial. Among the 222 patients with metastatic breast cancer, the objective response rate (ORR) was 15%. Patients with HER2 IHC3+ tumors showed a somewhat higher ORR than those with HER2 IHC2+ tumors, 18% versus 6% (17).

The final phase III trial with trastuzumab in metastatic breast cancer also used the CTA for enrollment and here, 469 patients with either a HER2 IHC2+ or a HER2 IHC3+ score were randomized to chemotherapy (doxorubicin/epirubicin plus cyclophosphamide or paclitaxel alone) or chemotherapy plus trastuzumab (10). The study showed that addition of trastuzumab to chemotherapy was associated with a higher ORR (50% versus 32%; P<0.001), a longer time to disease progression (median, 7.4 versus 4.6 months; P<0.001), and a 20% reduction in the risk of death. Again, it was shown that patients with a score of HER2 IHC3+ benefitted more from treatment with trastuzumab than those with a score of HER2 IHC2+. During the clinical development of trastuzumab, more than 7,000 tumor samples were tested in a single centralized laboratory using the CTA (11).

Due to the complexity of the CTA, it was not considered robust enough to be developed for a broader use in the general pathology laboratories. In order to standardize the assay and make it more robust and reliable for use in routine diagnostic laboratories, Genentech began a cooperation with the Danish diagnostic company Dako in 1996 (11). During the initial development of the new HER2 IHC assay, selection of the optimal primary antibody, epitope retrieval and staining procedure took place. Different primary antibodies were assessed, which included the two mouse monoclonal antibodies to HER2, clone 4D5 and clone CB11 as well as an affinity-purified, polyclonal rabbit antibody. In order to enhance the performance of the assay, a two-step immunohistochemical staining procedure using the EnVision visualization system (Dako) was used. The development work with the different prototype assays showed that the Dako polyclonal antibody A0485 in combination with the EnVision visualization system was superior with regard to immunohistochemical sensitivity and specificity compared to assays based on the two mouse monoclonal antibodies utilized in the CTA. Based on the prototype testing, the polyclonal rabbit antibody was chosen as the primary antibody for the new assay, later named HercepTest™ (11).

Several concordance studies were performed with the HercepTest™, which also included a comparison to HER2 fluorescence in situ hybridization (FISH) using the PathVysion^® HER2 DNA Probe Kit (Abbott Diagnostics). In FISH, the HER2 gene copy number in relevant tumor cells was determined and reported relative to the copy number of a centromere 17 (CEN17) reference probe determined in the same cells. Using HER2/CEN17 ≥ 2.0 as cut-off value for HER2 amplification by FISH, the study showed a binary concordance of 86% (11). Other studies have subsequently estimated the concordance between HER2 FISH and HercepTest™ to be around 95% (18, 19). A final concordance study between the CTA and HercepTest™ was performed including 548 breast cancer specimens. These specimens were supplied from the Cooperative Breast Cancer Tissue Resource following approval of the concordance study by the National Cancer Institute. Approximately 50% of these specimens were HER2 positive as determined by the CTA, with a more or less equal distribution of HER2 IHC2+ and IHC3+ cases. The binary concordance between the two assays was 79% (95% CI; 76% - 82%), which was above the prespecified acceptance criteria of 75% (11).

Following positive evaluations at an Oncologic Drugs Advisory Committee Meeting and a Device Panel Meeting, the FDA approved both trastuzumab and HercepTest™ in September 1998, as shown in Figure 1 (11). It was important to have a robust and reliable assay available that was easy to distribute to routine diagnostic laboratories at the same time as trastuzumab was approved for clinical use. However, looking retrospectively at the approval of HercepTest™, there are some questionable aspects from today’s point of view. There were no prospective or retrospective clinical outcome data following treatments with trastuzumab that could be linked directly to HercepTest™. This deficiency was also noted in relation to the intended use for HercepTest™, where FDA stated: ‘The actual correlation of the HercepTest™ to Herceptin^® clinical outcome has not been established’ (17). However, what was missing with regard to data on the clinical utility of HercepTest™ at the time of approval came in abundance later on.

The intended use granted by FDA for HercepTest™ in 1998 stated that it is a semiquantitative IHC assay to determine HER2 overexpression in breast cancer tissues, routinely processed for histological evaluation. Furthermore, in relation to the predictive properties, it is stated that HercepTest™ is indicated as an aid in the assessment of patients for whom Herceptin^® (trastuzumab) treatment is being considered (11). Together with the approval of HercepTest™ the FDA also approved the scoring criteria, as shown in Table 1 . Based on the tumor cell membrane staining pattern, the level of HER2 expression at the time of biopsy is categorized into the four categories of: IHC0 (negative), IHC1+ (negative), IHC2+ (weakly positive), and IHC3+ (strongly positive). Figure 2 shows a breast carcinoma tissue section with a strong expression of HER2 (IHC3+) following staining with HercepTest™.

The FDA approval of HercepTest™ was based on analytical data and concordance to the CTA. Since the patients in all the trastuzumab clinical trials were selected using the CTA, a study to compare this assay to HercepTest™ was conducted. The study was, as previously described, performed on independent samples, not originating from the trastuzumab clinical trials. Thus, the actual correlation of HercepTest™ to patient outcome following treatment with trastuzumab was not initially established. The clinical utility of the assay was documented in the years following the approval, where data from a large number of prospective clinical trials were published. HercepTest™ has been used for patient selection in a number of significant breast cancer trials such as the HERA, CLEOPATRA, EMILIA, and more (4, 20–35). In these trials, HercepTest™ has demonstrated its clinical utility in the neoadjuvant, adjuvant, and metastatic setting as well as in relation to different types of HER2 targeted therapies.

One of the more significant clinical trials conducted with HER2 inhibitors in the past decade is CLEOPATRA (29). Here, 808 HER2-positive patients with metastatic breast cancer were randomized to receive placebo plus trastuzumab plus docetaxel or pertuzumab plus trastuzumab plus docetaxel as first-line treatment. Pertuzumab (Perjeta^®, Roche/Genentech) is a humanized monoclonal antibody that binds to subdomain II of the extracellular part of the HER2 receptor, in contrast to trastuzumab, which binds to subdomain IV. In the trial, HER2-positivity was defined as IHC3+ or amplification by FISH (HER2/CEN17 ≥ 2.0). The patients were selected based on results following testing with HercepTest™ and/or HER2 FISH pharmDx (Dako). The study showed a significant higher median progression-free survival (PFS) for the pertuzumab group compared to the control group, 18.5 months versus 12.4 months (HR 0.62; 95% CI, 0.51 - 0.75; P<0.001). Likewise, a statistical significantly higher ORR was observed for the pertuzumab group compared to the control group, 80.2% versus 69.3% (P<0.001). Based on data from the CLEOPATRA trial, a concomitant FDA approval of pertuzumab and HercepTest™ was granted in 2012 (see Figure 1 ). The intended use statement for HercepTest™ was expanded to include the use of the assay as an aid in the assessment of breast cancer patients for whom treatment with pertuzumab is being considered (17). Another significant clinical trial, where HercepTest™ was used, is EMILIA (31). In this trial, 990 metastatic breast cancer patients previously treated with trastuzumab and a taxane were randomized to ado-trastuzumab emtansine (T-DM1) or lapatinib plus capecitabine. Ado-trastuzumab emtansine (Kadcyla^®, Roche/Genentech) is an antibody-drug conjugate consisting of trastuzumab covalently linked to the cytotoxic agent DM1. The HER2 selection criteria were similar to those of CLEOPATRA and the trial showed that ado-trastuzumab emtansine was superior to lapatinib plus capecitabine with respect to median PFS, median overall survival (OS), and ORR. Based on data from the EMILIA trial, a concomitant FDA approval of ado-trastuzumab emtansine and HercepTest™ was granted in 2013 (see Figure 1 ). Again, the intended use statement for HercepTest™ was expanded to include ado-trastuzumab emtansine (17).

For some patients tested positive with a HER2 IHC or a HER2 in situ hybridization (ISH) assay, treatment with trastuzumab or other HER2 inhibitors does not always result in a positive clinical outcome and in these cases the test result must be regarded as ‘false positive’, which likely is due to involvement of other molecular pathways than HER2 in the tumor progression. For the CLEOPATRA trial and the use of HercepTest™, the group of patients who were tested positive and received treatment with pertuzumab plus trastuzumab plus docetaxel had an ORR of 80.2% (29). In the CLEOPATRA trial where molecular enrichment was used, ORR is identical to the positive predictive value of the assay.

In the CLEOPATRA trial, a wide range of different biomarkers related to the HER2 pathway were explored for their predictive value, such as PIK3CA, PTEN, pAKT, and more (33). Based on the results from this comprehensive analysis, it was concluded that HER2 protein overexpression or HER2 gene amplification are the only markers suited for patient selection for the trastuzumab plus pertuzumab–based regimen in patients with HER2-positive metastatic breast cancer. This conclusion again underlines the importance of HER2 protein overexpression or HER2 gene amplification as predictive biomarkers for HER2 targeted therapies in patients with breast cancer.

Both during the clinical development of trastuzumab and a few years after its approval, patients with both HER2 IHC2+ and IHC3+ tumors were offered treatment and enrolled in different clinical trials (10, 16, 21–23). However, several trials had showed that trastuzumab was more effective in patients with HER2 IHC3+ than HER2 IHC2+ tumors and the treatment criteria were gradually revised (10, 16, 21). When the different adjuvant trials with trastuzumab were set up, the selection criteria changed. In the adjuvant HERA trial, HER2 positivity was defined as either HER2 IHC3+ or HER2 amplification using FISH (HER2/CEN17 ≥ 2.0). Furthermore, for patients with HER2 IHC2+ tumors, a reflex test using FISH was performed and if the tumor had HER2 gene amplification, the patient could be enrolled in the trial (4). Another important aspect to mention, is the criteria for HER2 IHC tumor cell membrane staining. When FDA approved trastuzumab and HercepTest™, they also approved the IHC scoring criteria for the assay, which required a strong complete membrane staining in more than 10% of the tumor cells in order to be classified as HER2 IHC3+, as shown in Table 1 (17). This staining requirement has been used in all pivotal breast cancer trials with trastuzumab including the metastatic and adjuvant indications, and subsequently also in relation to other HER2 inhibitors, such as pertuzumab and ado-trastuzumab emtansine.

In order to improve the accuracy and to standardize HER2 testing, the ASCO and the College of American Pathologists (CAP) issued a clinical practice guideline for HER2 testing in breast cancer in 2007 (36). As part of the guideline, in order to improve the robustness of the HercepTest™ scoring, it was suggested to change the requirements for HER2 IHC3+ from a > 10% strong complete membrane staining of the tumor cells to > 30%. This change was highly questioned as no clinical outcome data was available to support the 30% requirement, as all pivotal clinical trials with trastuzumab had used the original FDA approved scoring criteria (37, 38). The use of the suggested ASCO/CAP scoring criteria would identify a slightly different patient population compared to the populations in the different clinical trials with trastuzumab (39). However, with the revision of the ASCO/CAP guideline in 2013 and 2018, the scoring criteria became similar to the original FDA approved criteria (40, 41).

Initially, when HercepTest™ was considered for testing of metastatic gastric tumor specimens, the interpretation and scoring criteria were based on the existing breast cancer criteria. However, due to marked differences in tumor biology between breast and gastric cancer tissue, modifications of the scoring criteria would be required before the assay could be used for enrollment of patients in clinical trials with trastuzumab. The modifications implemented in the scoring of gastric cancer specimens were based on input from a panel of international pathology and oncology experts, who recommended certain changes to the HercepTest™ scoring criteria used in breast cancer. The major difference between the interpretation and scoring criteria for the two indications was that specific criteria for biopsies were added for gastric cancer (46, 47). These changes were necessary to assess the tumor heterogeneity and incomplete HER2 cell membrane staining observed in gastric cancer tissues. The modifications agreed upon by the panel were incorporated in the scoring criteria for HercepTest™ as shown in Table 2 (17). Due to the differences between gastric cancer and breast cancer tissues, the panel also recommended that both IHC and FISH testing should be performed for the enrollment of gastric cancer patients in clinical trials with trastuzumab (48). Figure 3 shows a tissue section of an adenocarcinoma of the stomach with a strong expression of HER2 (IHC3+) following staining with HercepTest™.

Positive results from preclinical studies with trastuzumab in different HER2 in vitro and in vivo gastric cancer models as well as data from a few positive case studies lead to the initiation of the ToGA trial (42). The trial was designed as an open labeled, randomized multicenter phase III study in HER2-positive patients with histologically confirmed inoperable locally advanced, recurrent, or metastatic gastroesophageal junction adenocarcinoma (46). HER2 positivity was defined as either IHC3+ or positivity by HER2 FISH (HER2/CEN17 ≥ 2), using HercepTest™ and the HER2 FISH pharmDx assays, respectively. Based on the recommendations from the expert panel, both IHC and FISH tests were performed for almost all patients. After inclusion in the study, patients were randomized to receive chemotherapy (5-FU or capecitabine plus cisplatin) or chemotherapy plus trastuzumab. More than 3800 patients were screened for the study and 584 of these were enrolled in the ToGA trial. The primary endpoint was OS, with different secondary endpoints including ORR (46, 48). Overall, the ToGA trial showed that the combination of chemotherapy plus trastuzumab was superior to chemotherapy alone. The median OS increased from 11.1 to 13.8 months (HR 0.74; 95% CI, 0.60-0.91). For the group of patients who had received the combined treatment with chemotherapy and trastuzumab ORR was 47% and for the chemotherapy group 35%. A pre-planned exploratory analysis that looked at the efficacy in the different HER2 IHC scoring categories showed that the survival benefit provided by trastuzumab was dependent on the level of HER2 protein overexpression. The single subgroup of patients with the greatest survival benefit was the one with HER2 IHC3+. Here, the median OS increased to 17.9 months when trastuzumab was added to chemotherapy compared to chemotherapy alone with an OS of 12.3 months (HR 0.58; 95% CI, 0.41-0.81) (46, 48).

Based on the exploratory subgroup analyses from the ToGA trial, a specific HER2 testing algorithm was developed where the primary test was IHC with FISH reflex testing of HER2 IHC2+ patients. As more than 90% (533/584) of the patients enrolled in the ToGA trial had tumors with HER2 gene amplification, it was argued that the selection criteria for treatment with trastuzumab should be both IHC3+ and gene amplification. In relation to the approval of trastuzumab for treatment of gastric cancer, this was in fact the position taken by the FDA, who recommended that reflex testing with FISH should be done for both IHC2+ and IHC3+ (17, 48). Furthermore, in the indication for use statement for HercepTest™, it says that the ToGA trial demonstrated that HER2 amplification and protein overexpression are not as correlated as for breast cancer, thus a single method should not be used to determine HER2 status in gastric cancer patients (17). Looking specifically at the test results for the patients enrolled in the ToGA trial, the agreement between HER2 amplification and HER2 overexpression is somewhat lower in gastric cancer than typically observed in breast cancer (49, 50). A relatively high number of HER2 FISH positive patients were found among the HER2 IHC0 and IHC1+ tumors.

In 2010, the gastric cancer indication for trastuzumab was approved by the FDA together with HercepTest™ and the HER2 FISH pharmDx assays. As shown in Table 3 , these two assays are still the only available FDA approved companion diagnostics for this indication today. Similar to breast cancer, ASCO, CAP and the American Society for Clinical Pathology (ASCP) issued a practice guideline for HER2 testing in gastroesophageal adenocarcinoma in 2017 (52). As for the original ASCO/CAP breast cancer guideline, the ASCO/CAP/ASCP guideline for gastroesophageal adenocarcinoma deviates slightly from the original FDA approved scoring criteria. Also, here, the primary test is IHC, but FISH reflex testing should only be performed in HER2 IHC2+ patients, which is in contrast to the original FDA approved scoring criteria where a HER2 IHC3+ test result must also be confirmed by FISH testing.