Assessment of Predictive Biomarkers in Breast Cancer: Challenges and Updates

The management of patients with breast cancer (BC) relies on the assessment of a defined set of well-established prognostic and predictive markers. Despite overlap, prognostic markers are used to assess the risk of recurrence and the likely benefit of systemic therapy, whereas predictive markers are used to determine the type of systemic therapy to be offered to an individual patient. In this review, we provide an update and present some challenges in the assessment of the main BC-specific molecular predictive markers, namely hormone receptors (oestrogen receptor [ER] and progesterone receptor [PR]), human epidermal growth factor receptor 2 (HER2), and KI67. As the main platform for assessing these markers in BC is immunohistochemistry (IHC), we address the cut-off values used to define positivity, the ER-low subgroup, the existence and significance of the ER−/PR+ phenotype, the use of PR in routine practice, and the role of hormone receptors in ductal carcinoma in situ. We discuss the newly introduced HER2-low class of BC and the clinical/biological difference between different HER2 groups (e.g., HER2 IHC score 3+ BCs vs. those with a HER2 IHC score 2+ with HER2 gene amplification). The review concludes with an update on the applications of KI67 assessment in BC and observations on the role of immune checkpoint identification in BC.

© 2022 S. Karger AG, Basel

IntroductionHormone Receptors

The oestrogen receptor (ER) was the first predictive marker to be introduced in cancer and remains the most commonly used and well-established predictive marker for consideration of endocrine therapy in the management of breast cancer (BC). ER is also considered as a major classifier of BC, and existing data indicate that ER-positive and -negative BCs have different clinical, molecular, and epidemiological risk factor profiles [1]. The outcome for patients with ER-positive BC is significantly better than for those with ER-negative tumours [2], with high levels of ER having the most favourable outcome [3]. Despite this, the prognostic value of ER per se is weak and wanes over time, and its most powerful value is as a predictor of response to endocrine therapy. ER-negative BC is unlikely to respond to endocrine therapy, whereas ER-positive tumours respond to such therapy although to a variable degree [4]. At least 70% of patients with strongly ER-positive BC demonstrate response to endocrine therapy while low ER-positive tumours disclose less response. However, there is no evidence that the correlation between the degree of response and the level of expression of ER in BC is linear. As such, the determination of the lowest cut-off that should be used to define ER positivity is difficult.

Immunohistochemistry (IHC) is the recommended assay for assessment of ER as well as for the progesterone receptor (PR). A number of gene expression assays are used as tools in clinical practice to predict the risk of recurrence and/or benefit from chemotherapy for ER-positive BC. Some of these, including mRNA-based assays, provide quantitative results for ER mRNA with good levels of concordance with ER measured by IHC and can serve as quality assurance for ER IHC, but there are no data that these assays can predict response to endocrine therapy. As such, current guidelines do not recommend the use of mRNA-based assays to identify patients for endocrine therapy [5-7]. However, discordant results between ER status by IHC and mRNA-based assays in routine practice should raise concern of a false IHC test (Table 1), and repeating the IHC test using a different antibody clone or different tumour block should be considered.

Table 1.

Causes of false ER test results

/WebMaterial/ShowPic/1439539

Using IHC, approximately 80% of BCs express ER [8-10]; however, the published percentage of positivity shows variation based on the study cohort, whether screen detected or symptomatic, patient ethnicity, age, and tumour types. There are also data to support the bimodal expression of ER in BC with >90% of the cases showing negative (<1%) or highly positive (≥70%) expression, while weak positive cases (1–10%; ER low) and intermediately positive (11–69%) cases are infrequent (2–3% and 5–9%, respectively) [9].

Defining ER Positivity

There has been considerable discussion regarding the threshold used to define an ER-positive tumour, which is currently defined as immunoreactivity in ≥1% of invasive carcinoma cells [11]. The rationale for this threshold is that tumours with this low level of ER positivity show response to endocrine therapy (despite being limited compared to tumours with stronger ER positivity) [12, 13], and it is logical not to deny patients access to potentially beneficial endocrine therapy. However, there is debate regarding the clinical significance of very low levels of expression. This uncertainly stems partly from the IHC methodology used in the initial validation studies which differed from current methods [4, 14]. IHC was carried out retrospectively and performed manually on historical surplus frozen tissue from samples taken initially for biochemical ligand-binding hormone receptor assay (LBA), the results of which were used for validation [4]. Although concordance between the two assays was good, discordance was observed close to the threshold. The fact that the ER expression profile differs between the assays is also hard to explain and further adds to the uncertainty. ER expression is linear by LBA in contrast to the well-established bimodal pattern observed by IHC [9, 10], a pattern that is also seen at the mRNA level [9].

The practical dilemma for the small proportion (4%) of tumours with test results close to the threshold cut-off of 1% and up to 10% of positive cells [15] (now categorized as ER low positive) [16] is whether to regard these as ER positive, requiring long-term endocrine therapy, or as ER-negative tumours in patients who would be eligible for other treatment options. There are data suggesting that BCs with ER expression in 1–10% of the tumour cells are biologically more similar to ER-negative BC than to strong ER-positive tumours on many levels [17-21]. ER-low positive tumours are molecularly heterogeneous and include all intrinsic subtypes by the PAM50 RT-qPCR assay with a predominance of the basal subtype and have ESR mRNA levels and ER gene signatures more similar to ER-negative than ER-positive tumours [17, 22]. The pathological features and response to neoadjuvant chemotherapy (NACT) for ER-low positive and ER-negative tumours are also similar [23].

Reproducibility issues with the interpretation of IHC results at this low level add to the difficulty [15, 24] and sampling issues are also relevant because a significant proportion of ER-low positive cases diagnosed on core needle biopsy (CNB) (∼45%) convert to ER negative on resection [9, 15, 25]. Therefore, if ER is 1–10% on CNB, repeat staining of the surgical excision specimen may be recommended. If the final score is <1%, the tumour is considered negative, if the score is >10%, it is considered positive, and if it is 1–10%, it is considered low positive. In these low positive (1–10%) results for ER, a comment may be included to reference the fact that there are limited data on the most appropriate treatment for patients with these tumours [26]. Therefore, the current guidance from American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) focuses attention on standardizing test results for ER-low positive cases which will facilitate studies aimed at understanding the biology and clinical behaviour of these tumours [16]. Whether or not more detailed molecular assays are required to guide treatment decisions in these uncommon cases remains to be determined.

ER status also informs clinical decisions other than predicting response to endocrine therapy. ER is used to build prognostic models; stratify patients into different treatment streams, e.g., HR-positive cancer versus TNBC; predict the likelihood of pathological complete response (pCR) to NACT; and to select patients for genetic testing [27, 28]. Notably, the ≥1% threshold has not been validated for these scenarios, and emerging data suggest that a higher cut-off closer to 10% may be more appropriate for many of these purposes [19]. It is worth mentioning that some European countries have already endorsed the 10% cut-off for ER and PR [29, 30].

In addition to participation in national external quality assurance schemes and in other quality assurances initiatives such as yearly/periodically monitoring of positivity rate per institution [30, 31], the use of external and internal positive and negative controls, and monitoring overall positive and negative rates collectively and for individual pathologists, further investigation of borderline cases, such as ER low positive, ER−/PR+ cases, and tumours with unusual staining patterns (potential false-positive cases including strong background staining and granular nuclear staining with cytoplasmic spillover [15]) can identify any drift in the sensitivity of the assays. Measures that can be taken include repeat staining on the same or subsequent specimen, repeat staining using a different clone or in a different laboratory, and correlation with results of the multigene assays, if available. If the results are discordant, further workup should be carried out to investigate the actual status of the tumour and the performance in the local laboratory.

In addition to the percentage of staining of ER, several scoring systems for ER exist that have their own definitions for positivity and are included in the report in many laboratories. Their scoring systems assess the quantitative proportion of stained nuclei and a measure of staining intensity and combine these to give an overall score for a tumour. Single use of these scoring systems without stating the percentage of positive cells should be discouraged as discrepancies at the low cut-off of positivity exist. Each system classifies tumours in a slightly different way which makes direct comparisons difficult. The most commonly used scoring systems are the “Allred” [4] or “modified quick” score and the H-score (Table 2). It should be noted that the cut-off for positivity using the Allred score is ≥3. In a small number of tumours, a score of 3 can be obtained if < 1% of cells show moderate or strong staining. Such tumours are positive by Allred system but are in conflict with the ≥1% cut-off and should be classified as HR negative regardless of the Allred score [11]. The response of these tumours to hormonal therapy is unknown. It is our practice to assess and report the percentage of positive cells and report as positive or negative using the ≥1% cut-off, with some laboratories also providing the H-score and Allred score.

Table 2.

The most commonly used scoring systems: the Allred/quick score and the H-score

/WebMaterial/ShowPic/1439537Progesterone Receptor

Molecular cross talk between ER and PR occurs and the PR gene is regulated by ER as an ER-dependent gene product. PR expression is generally regarded as a surrogate for a functional ER pathway. As such, loss of PR expression in some ER+ tumours has been explained by a non-functional ER pathway; hence, the presence of PR usually indicates that ER pathway is intact and functional [32]. However, the fact that some ER-positive/PR-negative tumours respond to ER antagonists suggests that the ER pathway remains functional and that other regulatory mechanisms are involved [33]. PR can also regulate ER, and PR expression in ER+ tumours is associated with a good prognosis [34, 35]. To reflect the prognostic role of PR in BC, several multigene signatures contain PR as one of the main component genes in their testing gene set. PR levels by IHC have been used by various prognostic tools, such as Magee Equations, IHC4 score, the AJCC eighth edition prognostic stage groupings, and nomograms that predict the 21-gene recurrence score results [26], in addition to various predictors of response to neoadjuvant therapy [36-38].

However, the predictive role of PR independent of ER in the adjuvant setting is less well established. Several studies have reported a more favourable outcome for patients with ER-positive/PR-positive tumours following endocrine therapy compared to those with ER-positive/PR-negative tumours and an inverse relationship between the level of PR expression and outcome following endocrine therapy [39]. However, meta-analyses of randomized clinical trials have not confirmed a predictive value for PR for response to endocrine therapy that is independent of ER [40].

As the primary purpose of HR assessment in BC is to identify patients who are likely to benefit from endocrine therapy, testing for PR in BC remains optional in some countries and centres. The 2010 American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines [41] recommended that ER and PR status be determined on all BC while the 2020 guidelines [26] stated that PR is used primarily for prognostic purposes in the setting of an ER-positive BC but also noted that the principles of ER testing apply to PR, which implies that PR testing in BC is mandatory. Similar recommendations are made by the UK National Institute of Health and Care Excellence (NICE) guidelines [42], the European Group on Tumour Markers [43], the European Society for Medical Oncology (ESMO) [44], the Spanish Society of Medical Oncology (SEOM) [45], the Chinese Society of Clinical Oncology (CSCO), and others [46]. However, the UK Royal College of Pathologists guidelines [47] and some others regard PR testing in BC as optional. It is our experience and opinion that PR should be assessed routinely in BC for the following reasons: (1) PR status not only provides prognostic significance [48, 49] but also provides information about the ER pathway and the level of likely clinical response to endocrine therapy [5, 16]. (2) PR status can be used for the molecular classification of BC into luminal A and B subtypes. (3) PR testing provides prognostic significance equivalent to Ki67 expression in BC and is included as one of the component genes in many multigene prognostic signatures that are increasingly used in the evaluation of BC. (4) PR staining can provide a quality assurance measure for ER staining particularly to minimize ER false-negative cases. (5) PR testing follows the same procedures as ER and its testing is inexpensive compared to molecular assays.

PR positivity is seen in 65–75% of BC and in approximately 80–90% of ER-positive cases, with the proportion varying depending on the cut-off point used for positivity [32, 34]. Compared with ER, there are fewer studies standardizing and validating IHC assays for PR and focussing on the cut-off of positivity. Although the low 1% cut-off for ER positivity is also recommended for PR [26], this seems very low in view of the different clinical indications and utility of ER and PR status. There is no evidence to suggest that PR-low positive tumours (1–10%) show response to endocrine therapy if the BC is ER negative, or that the predictive response of ER-positive tumours is significantly different if PR is expressed at such a low level [27, 50]. In addition, the previous studies on the prognostic value of PR mainly used a cut-off of 10%, 20%, or greater [34, 46, 48, 50]. Here, we propose that BCs with >10% PR immunostaining are PR positive, tumours with 1–10% are PR poor, and these should be combined with tumours lacking PR staining (0%) or showing <1% expression (PR negative). However, in line with the current recommendations [26], we also recommend reporting the percentage and intensity of cells staining, which provide information about the degree of PR positivity and subsequently the magnitude of the prognostic and predictive value of PR status.

Although the expression of PR is highly correlated with ER, the correlation is imperfect, resulting in four combination phenotypes with ER positive/PR positive being the most common. The ER-positive/PR-negative and ER-negative/PR-negative combinations are reported to each account for 15–20% of BCs [10]. There is debate about the validity of the ER-negative/PR-positive category, which is reported to account for 0.3–2% of BCs in different series [10, 25, 51, 52]. Approximately 1% of a series of around 200,000 BCs diagnosed between 2009 and 2016 in routine practice were ER negative/PR positive suggesting that the category is exceptionally rare using current assay methods [10]. Expression of PR in the absence of ER is not easily explained biologically because PR is a downstream product of ER regulation. An explanation for this anomaly is that a conformational change in ER, possibly as a result of an ER mutation, prevents antigen binding but retains constitutive activity; this remains unconfirmed. Other explanations include downregulation of ER expression by epigenetic regulation, HER2 receptor, or by high oestrogen levels [53]. However, some large series reported a response to endocrine therapy in ER-negative/PR-positive tumours, which may be substantial, but which again is not easily explained in the absence of a functional ER pathway [54] but rather by enrichment of this subgroup with tumours showing false-negative ER staining [25, 55]. It is noteworthy that several studies have found it difficult to reproduce the ER-negative/PR-positive category and a significant proportion of these tumours are reclassified as either dual receptor positive or negative on re-testing [9, 56]. In a previous study of 9844 BC [56], 27 cases were initially diagnosed as ER negative/PR positive (0.3%). Following retesting, the ER-negative/PR-positive phenotype was confirmed in 7 cases only (0.1% of all BC) [56]. We (ER) have observed similar findings in Nottingham when we reviewed and re-evaluated the ER-negative/PR-positive cases over the last 10 years. When we encounter an ER-negative/PR-positive BC, particularly in a CNB sample, we strongly recommend repeat testing on additional blocks, optimally fixed, from the surgical resection specimen [25, 26, 56] (see above). Confirmed ER negative/PR positive following repeat staining may be offered endocrine therapies, although there are limited data to support this [26].

Reporting Hormone Receptor Status in Ductal Carcinoma in situ

ER status is used in the setting of non-invasive disease to select patients with pure ductal carcinoma in situ (DCIS) for potential treatment with endocrine therapy. Determination of ER status in pure DCIS is increasingly recommended and used in when endocrine therapy is being considered [16, 57] and as alternative to surgery particularly during the COVID-19 pandemic and the need to delay surgery deemed to be non-urgent. The adjuvant use of ER in DCIS is based on clinical trial data that demonstrate a significant reduction in ipsilateral and contralateral BC following adjuvant endocrine therapy in women with DCIS [58], although an impact on survival has not yet been demonstrated [59, 60]. It is recommended that the quantification method, reporting and scoring systems that are used for invasive tumours are applied to DCIS [6, 60]. However, there are no studies evaluating different cut-off levels for positivity in DCIS and most studies have applied a ≥1% cut-off or an Allred score of >2 [58], but care should be taken when interpreting the expression of ER in DCIS and a cut-off of >10% should preferably be used if the results are to be considered for further management of patients. There is no evidence that patients with DCIS showing 1–10% ER positivity will benefit from endocrine therapy. This is particularly important in the context of DCIS heterogeneity where low-grade low-risk DCIS is more likely to be ER positive, while high-grade, high-risk DCIS that may require greater intervention is the component that may lack ER expression. In such lesions with heterogeneous expression of ER, a comment on the grade of DCIS and location of ER expression may help in guiding further management decision-making. There are no data to support either a prognostic or predictive role for PR in DCIS and we do not recommend its assessment in DCIS as a routine unless used as a diagnostic marker.

Human Epidermal Growth Factor Receptor 2

A proportion of BCs overexpresses HER2, which is mainly due to ERBB2 (HER2) gene amplification [61, 62]. Earlier studies have shown that the HER2 gene is amplified or the protein is overexpressed in up to 30% of BCs. More recent data show a rate that is closer to 15%, probably reflecting the adoption of strict guidelines for its assessment with reduction of false-positive results, reported in up to 19% of cases [63, 64], and the frequency of screening mammography detecting early-stage BC in published series [32]. The HER2 gene encodes a tyrosine-kinase receptor residing on the surface membrane of breast epithelial cells and forms complexes with similar proteins (such as erbB1, erbB3, and erbB4) which act as receptors for several ligands that regulate many normal cellular functions. Clinically, overexpression of HER2 is associated with aggressive histological features and shorter survival [34, 65]. Importantly, HER2 is the main predictive factor for treatment with anti-HER2-targeted therapy such as trastuzumab, pertuzumab, lapatinib, tucatinib, and neratinib, which significantly improve the outcome for patients with HER2-positive BC [66]. Trastuzumab was the first anti-HER2-targeted agent developed and is a humanized monoclonal antibody that works by binding to the HER2 receptor. Subsequently, several biosimilar versions of trastuzumab have been developed. This field continues to expand with several anti-HER2 antibody-drug conjugates (ADCs) currently in development. Many HER2-targeted therapies have been Food and Drug Administration (FDA)-approved including ado-trastuzumab emtansine (T-DM1) and the fam-trastuzumab deruxtecan (T-DXd). ADCs incorporate the HER2-targeted anti-tumour properties of trastuzumab with the cytotoxic activity of the potent agent, allowing intra-cellular drug delivery specifically to HER2-overexpressing cells. The use of ADC targeting HER2 protein in BC has expanded the use of anti-HER2 therapy to BC displaying varying levels of HER2 membrane expression without the need to have evidence of HER2 gene amplification (the HER2-low group; see below).

IHC for protein overexpression and in situ hybridization (ISH) for gene amplification are the techniques recommended for determining HER2 status. Currently, other available HER2 testing techniques including mRNA-based assays are not recommended for diagnostic purposes. High concordance between IHC and gene amplification status is reported [67]. The current UK recommendations for HER2 testing advocate a two-tier system using IHC with reflex ISH testing in equivocal cases, but this does not preclude laboratories from using primary HER2 ISH testing particularly if the quality of tissue fixation is questionable [68].

Only membrane staining of invasive tumour cells is considered when scoring HER2. The HER2 IHC scoring method is a semi-quantitative system that assesses the completeness and intensity of staining of the reaction product and percentage of membrane-positive cells, giving a score range of 0–3+. Tumour samples scoring 3+ are regarded as unequivocally positive while those scoring 0/1+ are negative as they are unlikely to show HER2 gene amplification or respond to anti-HER2 therapy. Borderline scores (2+) are regarded as equivocal and mandate further assessment using ISH as only 15–30% of these cases show HER2 gene amplification.

HER2-Low BC

The traditionally dichotomous separation of HER2 into HER2-positive and HER2-negative BC was based on firm clinical data demonstrating that only HER2-positive tumours benefit from anti-HER2-targeted therapy. However, with the development of ADCs that are designed to target HER2 protein and deliver chemotherapy inside cancer cells, there is increased interest in BCs that express HER2 without evidence of gene amplification or of HER2 being an oncogenic driver of BC. Some ADCs, such as trastuzumab deruxtecan and trastuzumab duocarmazine [69, 70], have demonstrated encouraging response rates not only in HER2-positive but also in the so-called HER2-low BC patients [71]. HER2-low BC is defined as those tumours with HER2 IHC scores of 1+ or 2+ without HER2 gene amplification and comprises about 50–55% of BC [71, 72]. These tumours are more frequently hormone receptor positive compared to HER2 score 0 and HER2-positive tumours (64% compared to 37% and 50%, respectively [72]) and show clinical differences compared to HER2 score 0 and HER2-positive tumours both in outcome and response to NACT [72]. These results suggest that HER2-low positive tumours can be identified as a new clinically significant subgroup of BC that are distinct from HER2 score 0 and HER2-positive tumours. In general, these tumours do not respond to trastuzumab and other drugs that target the HER2 oncogenic pathways, but this HER2 profile may be considered as a potential target of the trastuzumab containing ADCs [66, 70-73].

The introduction of the HER2-low class of BC will have an impact on the pathological assessment of HER2 status as the existing guidelines focus on the identification of HER2-positive tumours and little or no clinical data available to inform cut-off values for distinguishing HER2 IHC scores 0 and 1+. Unlike the criteria for IHC score 2+, detailed in the existing guidelines, the definition of HER2 IHC score 1+ BC is less clear [74, 75]. The current definition of score 1+ specifies incomplete membrane staining that is faint/barely perceptible and in >10% of tumour cells while score 0 cases are defined as either no staining or the presence of membrane staining that is incomplete and is faint/barely perceptible and in ≤10% of tumour cells [74, 75]. In the context of using HER2 protein as a target of ADC therapy, future studies to investigate the clinical significance of these arbitrary criteria and supportive evidence for the 10% cut-off and the character and intensity of membrane staining, currently used to define the different HER2 categories, are required. In addition, an important effort will be necessary to address the reproducibility issue of the HER2-low category. Some national guidelines have already begun to alert pathologists to this new category and on the necessity to precisely assess and document the score 1+ category in their pathology reports [76].

Current definitions may need adjustment to ensure adequate response to such therapy, including consideration of potential side effects, if we consider the differential expression of HER2 protein compared to normal tissues (bystander effect). The HER2 receptor has an important role in normal cell growth and differentiation and HER2 is a membrane constituent of a variety of epithelial cell types. Low level of HER2 protein expression is identified on cell membranes of epithelial cells in the gastrointestinal, respiratory, reproductive, and urinary tract as well as in the skin, breast, and placenta in a level which is similar to the levels found in non-amplified, non-overexpressing BC [77, 78].

Differential Response of HER2-Positive BC Based on Protein Expression

Another recently reported observation is the differential response of HER2-positive BC to neoadjuvant HER2-targeted therapy based on the level of HER2 protein expression. The magnitude of response, measured by the rate of pCR, of HER2 IHC score 3+ cases ranges from 55 to 70% compared to 17–20% in HER2 IHC score 2+ with HER2 gene amplification [79, 80]. Such low response rates of BC with equivocal HER2 protein expression (IHC score 2+) appear not to differ from the response rates of HER2-negative BC to chemotherapy (15–25%) and may support the increasing use of ADC in these tumours.

The biological functions of the HER2 receptor are related to the level of protein expression (e.g., number of HER2 membrane receptors per cell) rather than to HER2 gene amplification per se. Previous studies have demonstrated that the level of HER2 gene amplification is not a prognostic factor in patients with HER2-positive BC treated with adjuvant trastuzumab-based targeted therapy [81, 82]. In clinical practice, HER2 gene amplification testing is mainly used to confirm the presence of HER2 protein overexpression [83] and these are strongly correlated [84]. In HER2 2+ tumours, the level of HER2 protein expression is equivocal and the level of HER2 gene amplification is low compared to tumours with definite HER2 overexpression (score 3+). The relatively low level of HER2 protein expression may explain the limited response of the HER2 2+ tumours that are classified as HER2 positive based on the demonstration of HER2 gene amplification [81].

Some authors have suggested that the level of HER2 IHC staining in HER2-positive tumours has no role in clinical management with adjuvant trastuzumab [84]. However, their analysis was based on a spectrum of HER2 staining rather than on a dichotomized classification (IHC 3+ vs. 2+). Others have proposed that assessment of HER2 gene amplification by FISH testing is the preferred method to select patients for trastuzumab therapy. Their studies included IHC 3+ and 2+ tumours [85, 86] such that the FISH-positive group included IHC 3+ and 2+ (hence overall high benefits) while the FISH-negative group included only IHC 2+ tumours (hence low benefits), resulting in overemphasis of the value of FISH testing. In another retrospective study of 364 HER2-positive BC patients [87], the authors concluded that FISH was superior to IHC-based assessment of HER2 status. In that study, 22 of the IHC 3+ tumours were FISH negative and 12 tumours were reclassified into IHC 0/1+ which makes interpretation of these results difficult [75]. Our clinical and research experience supports the cumulative evidence that IHC is the preferred method of HER2 assessment in BC. Although the role of ER in the response rate of HER2-positive BC is recognized, more data are needed to determine whether ER status plays a role in the low response rate of the IHC score 2+ HER2 gene amplified category compared to the IHC score 3+ tumours or if the impact is solely dependent on the activation of the HER2 pathways and the oncogenic effect of the HER2 protein.

Unusual Patterns of HER2 IHC Staining

For IHC scoring of HER2, only complete membranous staining is considered with intensity varying from strong to weak in >10% of cases for scoring 2+ and 3+ cases. Cases with incomplete membrane staining are typically considered in the HER2-negative category. However, some BC may show incomplete membrane staining with evidence of HER2 gene amplification. Invasive micropapillary carcinoma (IMPC) is a unique type of BC that clusters of tumour cells surrounded by clear spaces with inverted polarization of cells (inside out pattern with luminal surface of the cells facing the stroma). This morphology of IMPC often impacts the staining of HER2 with basolateral staining in a cup or U shape which spares the luminal aspect of the cell [88]. This can result in “incomplete” membranous staining similar to the HER2 staining pattern that is observed in gastric carcinomas [89]. Therefore, IMPC with moderate to intense but incomplete membrane staining in >10% of tumour cells is classified as HER2 IHC equivocal (2+) instead of 1+ [90] with approximately 20% of these cases showing HER2 gene amplification [88].

Modification of ER, PR, and HER2 Status during BC Evolution

Changes in the status of ER, PR, and HER2 status between CNB and excision specimens, between primary and metastatic tumours, and following neoadjuvant therapy have been the subject of many reviews and meta-analyses [91-93] including a review in this issue of the journal. Loss of ER or PR in recurrent tumours is associated with shorter outcome compared with receptor-positive concordance. Contrasting this, gain of ER is associated with longer survival compared with receptor-negative concordance [93]. Despite the clinical evidence of the impact of the receptor discordance and the high rate of phenotypic drift reported in some studies, we advise investigating such discordant cases to exclude false-positive or -negative results. This may include review of the previous biopsy, interrogation of the clinical history, and even repeating the staining in some cases (e.g., strongly ER-positive primary tumour and ER-negative recurrence). Loss of ER and PR following endocrine therapy and loss of HER2 expression following anti-HER2 therapy can be explained biologically by therapy effect. In addition to the impact of intratumour heterogeneity, other factors that may contribute to receptor discordance include technical issues (e.g., metastatic tumour may be tested on cytological material or on a bone biopsy with the impact of decalcification), the presence of more than one primary tumour at time of initial diagnosis with different phenotypes, a weakly positive primary tumour with receptor positivity around the threshold of positivity, a new primary tumour with a different receptor status as distinct from a true recurrence or a metastasis originating from a different, possibly occult, tumour in the ipsilateral or contralateral breast.

Assessment of KI67 in BC

Ki67 is a non-histone nuclear protein that was first detected in 1980s [94]. Its expression is correlated with the proliferation status of the cells and BC demonstrating the highest levels of Ki67 is associated with features of biological aggressiveness. Levels are higher in grade 3 tumours, ER negative, HER2 positive, and TNBC and are associated with shorter survival outcomes [95, 96]. Several methods have been described to assess the proliferative activity of tumours. Ki67 has attracted the most attention in the oncology community as a potential prognostic marker in BC, especially in HR-positive disease, as a discriminator between the two subtypes of luminal tumours and as a marker of response to neoadjuvant hormone and chemotherapy [97]. Unlike mitotic counts which identify mitotically active cells committed for cell division (M phase of the cell cycle), Ki67 IHC staining not only identifies cells committed to cell division but also demonstrates cells in the interphase stage of the cell cycle [98, 99]. This may question the biological value of Ki67 as a surrogate marker for cell division in BC. In addition, the added value of the Ki67 score as assessed using IHC over adequately measured mitotic counts in well-fixed BC tissue remains to be confirmed [100, 101]. The analytical validity of Ki67 is also questionable due to the lack of consensus on scoring methods and cut-off values, the reliability of various antibodies, and the actual clinical utility [102-104]. Until now, the traditional visual morphological detection of mitotic figures using well-established methods in H&E sections remains the most widely used and accepted method in routine practice in most laboratories as it is an easy, rapid, and simple technique for assessment of growth rate and tumour behaviour [101, 105, 106].

Neoadjuvant Therapy

A pCR is generally used as a surrogate endpoint for survival in NACT trials [107], but the precise relationship between pCR and overall survival is uncertain [108]. For ER-positive tumours, the likelihood of achieving a pCR and its prognostic significance are less than for the more biologically aggressive tumour subtypes [109]. Consequently, the potential prognostic role of other biomarkers such as Ki67 is being explored for ER-positive disease. Based on data from 1,166 participants, increasing pre-treatment Ki67 levels, in tertiles of <15%, 15–35%, and >35%, were a significant independent predictor of adverse outcome at 6 years in the ER-positive subgroup [110]. In ER-positive patients, Ki67 measured after treatment provided more prognostic information than the baseline Ki67 level or pCR on multivariable analysis [109]. High post-treatment Ki67 scores identified patients who were at significant risk of relapse compared to those with intermediate Ki67 levels, and the outcome for patients with low Ki67 in the residual disease was no different to that for those who had achieved a pCR. Thus, tumours in which the proportion of proliferating cells decreased following NACT appeared to have a better prognosis than tumours in which proliferation remained high.

Substantial evidence supports Ki67, measured following a short period of neoadjuvant endocrine therapy (NeoET), as a strong prognostic marker and Ki67 is the most widely used surrogate for outcome in these NeoET studies [111]. One of the earliest NeoET studies was the IMPACT trial that investigated if short-term changes following anastrozole or tamoxifen, alone or in combination, were predictive of long-term recurrence and its design mirrored the larger ATAC adjuvant trial that had the same treatment arms [112]. Analysis showed that both the baseline and 2-week Ki67 levels were predictive of recurrence, but only the level at 2 weeks was an independent predictor of survival on multivariable analysis after a median follow-up of 37 months [113].

An advantage of NeoET is that it allows the biological response to therapy to be examined, and a change in Ki67 levels after a short period of treatment is the most widely used metric to assess response. Aromatase inhibitor (AI) therapy has been used in most NeoET studies and the results support Ki67 as a reliable indicator of the biological activity of AI therapy [111, 114, 115]. In the NeoET IMPACT trial, suppression of Ki67 occurred after 2 and 12 weeks of therapy and was significantly greater following anastrozole (76% and 82%, respectively) compared to tamoxifen (60% and 62%, respectively) or the combination of both agents (64% and 61%, respectively) [116]. Thus, a change in Ki67 may identify cases that respond poorly to AI therapy and could be used to target therapeutic.

Conclusion on the Prognostic and Predictive Role of Ki67

Despite strong evidence supporting Ki67 as a prognostic marker and as a predictor of response to therapy, Ki67 is not widely used in routine clinical practice. The main barrier to its use is the well-documented concern regarding the analytic validity of the assay that has been difficult to overcome. In a recent consensus meeting of the International Ki67 in Breast Cancer Working Group (IKWG) [97], current evidence for Ki67 IHC analytical validity and clinical utility in BC was presented, including considerations for pre-analytical handling and participation in quality assurance and quality control programmes. A standardized visual scoring method was recommended for adoption. It was concluded that Ki67 IHC as a prognostic marker in BC has clinical validity, but the clinical utility is evident only for prognosis estimation in ER-positive and HER2-negative patients to identify those who are unlikely to benefit from chemotherapy. Cut-offs of 5% or less or 30% or more are now recommended to estimate prognosis [97] rather than using a single cut-off of 10% or 14%, cut-offs which have been commonly used to dichotomize BC into two groups resulting in extensive debate regarding the optimal single cut-off and its reproducibility.

In 2021, the US FDA approved abemaciclib (CDK4/6 inhibitor) in combination with endocrine therapy for the adjuvant treatment of adult patients with ER-positive, HER2-negative, node-positive, early BC at high risk of recurrence, and a Ki67 score of ≥20% as determined by an FDA-approved test [117, 118]. This perspective of using Ki67 to predict response to CDK4/6 inhibitors might considerably change our clinical practice with regard to use of Ki67.

Importantly, Ki67, which reflects the proliferative activity of BC, is a continuous variable. A pragmatic approach to the use of a cut-off should be adopted, appreciating that a cut-off derived from clinical trials inclusion criteria or retrospective studies based on an association with specific end points is related to the variables tested and to the context in which Ki67 was used. Unlike ER and HER2, which are mainly used in the clinical setting as predictive markers for specific therapy with standardized thresholds to determine positivity, Ki67 is used in a variety of clinicopathological scenarios with varying cut-offs emphasizing the importance of high levels of concordance and adherence to the published evidence in the relevant context.

Diagnostic Applications of Ki67 in Breast Pathology

Ki67 can be used as an adjunctive diagnostic aid for certain breast lesions. Ki67 is typically low in low-nuclear grade intraductal proliferative lesions such as atypical ductal hyperplasia, flat epithelial atypia, and lobular neoplasia but is high in intermediate and high-grade DCIS and pleomorphic LCIS. Epithelial hyperplasia of usual type may show scattered positive cells but not very high Ki67 expression. Assessment of Ki67 expression in stromal cells can also be helpful in some cases such as in the multinucleated multilobated giant stromal cells that can be seen in fibroepithelial lesions. Ki67 is typically low in these cells when present in benign lesions, whereas in malignant giant cells, it is typically positive in the majority of these cells reflecting the proliferative activity rather than the degenerative nature of these cells in benign lesions. The frequency of Ki67 expression in the stromal cells can help grade phyllodes tumours if mitotic figures cannot be accurately assessed due to poor fixation. Ki67 can also help to determine tumour grade when mitotic counts are borderline as in grade 2 tumours and in tumours with poor fixation [100].

Predictive Role of Multiple IHC Biomarkers

Evaluation of aforementioned 4 breast IHC biomarkers alone provides prognostic and predictive information; however, combinatorial expression can offer prognostic classification of BC much better than any individual biomarker [119] and can predict response to chemotherapy in the neoadjuvant settings. In a different approach, the combined expression of these 4 biomarkers, with or without clinicopathological variables, were used to triage cases for the Oncotype DX test. Combining H-score of ER and PR, status of HER2, percentage of Ki67, tumour size (in centimetres) with the Nottingham grade score were demonstrated to predict the Oncotype DX recurrence score [120, 121]. This is the first of three models of multiple linear regression analysis, known as Magee EquationsTM (ME1, ME2, ME3), which were designed based on different data availability (http://path.upmc.edu/onlineTools/MageeEquations.html). The second model (ME2) was based on the same data except Ki67. MEs 1 and 2 require tumour size and Nottingham grade score in addition to semi-quantitative receptor results for calculation, so ME1 and ME2 scores are not available on all cases [122]. The third model (ME3) utilizes semi-quantitative IHC expression levels for ER, PR, HER2, and Ki67. ME3 is useful in the neoadjuvant setting as pathological tumour size and Nottingham scores are not available or reliable in core biopsy before therapy. For cases with residual disease, residual cancer burden score/category (whenever available) was correlated with ME3 scores [122]. The estimation of the recurrence score using Magee Equations and Magee Decision AlgorithmTM has been validated retrospectively [123, 124]. Magee Equation 3 score ≥31 predicts pCR in the neoadjuvant setting and for tumour recurrence when pCR is not achieved [125]. Patients with high ME3 scores were 13 times more likely to achieve pathologic complete response compared to those with ME3 scores less than 31 [125]. ME3 is a simple, fast, robust, and reliable tool to select ER+/HER2−patients who will benefit the most from NACT [122]. High ME2 score was also associated with tumour size reduction after NACT what was proved by Saigosoom et al. [126] in retrospective study. The authors applied pre-treatment status of ER, PR, HER2, and Ki67 from core biopsy pathology reports and estimated tumour size reduction based on the difference between pre-treatment clinical size and post-treatment pathology size [126]. According to the authors, this finding might suggest the application of Magee Equation for prediction of NAC in HR+, HER2− BC patients who will receive NACT with the aim of downsizing for conversion from total mastectomy to breast conserving surgery.

As last year oncologists and pathologists from Leuven, Belgium, showed in retrospective analysis that ME can be useful in selecting ER+, HER2−-early BC patients may need Oncotype DX testing [127]. Concordances of 100% were observed between risk results obtained by Oncotype DX and ME2 or average of the three Magee equations. The using of such cost-effective method as ME may reduce delay in therapeutic decision-making.

Immune Checkpoint Molecules

The central role of the immune system in cancer biology and therapeutics is being increasingly recognized and studied. Immune checkpoints are regulators of the immune system and are crucial for self-tolerance, which prevents the immune system from attacking cells indiscriminately. However, some cancers can protect themselves from attack by stimulating immune checkpoint targets. Immune checkpoint molecules include stimulatory and inhibitory factors. Stimulatory checkpoint molecules include members of the tumour necrosis factor (TNF) receptor superfamily – CD27, CD40, CD137, OX40, and GITR – and the B7-CD28 superfamily – CD28 itself and ICOS. Inhibitory checkpoints include several molecules such as PD-1, CTLA-4, A2AR, B7-H4, BTLA, IDO, LAG3, and B7-H3 (CD276). Inhibitory checkpoint molecules are targets for cancer immunotherapy due to their potential for use in multiple types of cancers. The checkpoint inhibitors approved at present block CTLA4 and PD-1 and PD-L1. The programmed cell death 1 (PD1) protein, a member of the B7/CD28 family of co-stimulatory receptors present mainly on T cells, sends inhibitory signals to T cells to suppress the anti-tumour response [128]. PD1 functions mainly through its ligand PDL1, which is expressed on a variety of immune cells such as antigen-presenting cells and activated T and B lymphocytes and tumour cells. CTLA-4 (or CD152) expression on Treg cells serves to control T-cell proliferation.

A number of immune checkpoint inhibitors are used in the treatment of several types of cancer including BC. These include antibodies against CTLA-4 such as ipilimumab and tremelimumab and against PD-1 such as pembrolizumab and its ligand PD-L1 such as atezolizumab. To date, pembrolizumab and atezolizumab have been the most extensively studied in BC. Atezolizumab is the first immunotherapy to be approved for PD-L1-positive, advanced TNBC and is administered in combination with the chemotherapy agent nab-paclitaxel with PD-L1 positivity defined as expression in >1% of immune cells [129]. If atezolizumab is planned, PD-L1 should be tested using the SP142 antibody on the immune cells of the tumour stroma and the cut-off value for positivity is 1%. On the contrary, if pembrolizumab is planned, PD-L1 should be tested by the 22C3 pharmDx test, in the tumour and immune cells (combined score – CPS), and the cut-off is ≥ 10 [130].

In 2019, the FDA approved atezolizumab in combination with chemotherapy for adult patients with unresectable locally advanced or metastatic PD-L1-positive TNBC. In 2021, FDA approved pembrolizumab for high-risk, early-stage TNBC in combination with chemotherapy as neoadjuvant treatment, followed by single-agent adjuvant treatment post-surgery. Currently, PD-L1 testing is not performed in routine clinical practice for TNBC unlike the situation for other cancers such as non-small-cell lung cancer and urothelial carcinoma. More widespread PD-L1 testing in BC is likely to be limited to central reference laboratories or those in hospitals offering immunotherapy to patients with BC. At present, each agent is licenced with specific companion diagnostics. In the future, however, it is likely that additional validated immunohistochemical assays to assess the target immune checkpoint expression will be available and approved.

Other Predictive Markers in BC

Other established and novel predictive markers in the field of BC therapy include BRCA1/BRCA2 gene mutation detection in the setting of hereditary BC, ESR1 gene mutations in recurrent tumours following endocrine treatment, and PIK3CA mutation testing to predict response to PI3K inhibitors. These are not discussed in this review as testing is not performed using IHC. The role of androgen receptor IHC testing in TNBC remains to be defined and other IHC predictive markers in BC are expected to emerge in the near future.

Acknowledgments

Emad Rakha is part of the PathLAKE digital pathology consortium. These new centres are supported by a £50m investment from the Data to Early Diagnosis and Precision Medicine strand of the government’s Industrial Strategy Challenge Fund, managed and delivered by UK Research and Innovation (UKRI).

Conflict of Interest Statement

The author declares that there is no conflict of interest.

Funding Sources

No funds.

Author Contributions

Emad Rakha drafted the manuscript and reviewed and approved the submission. Ewa Chmielik reviewed the manuscript and approved the submission. Fernando C Schmitt reviewed the manuscript and approved the submission. Puay Hoon Tan reviewed the manuscript and approved the submission. Cecily M Quinn reviewed the manuscript and approved the submission. Grace Gallagy reviewed the manuscript and approved the submission.

Copyright: All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and retrieval system, without permission in writing from the publisher.
Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug.
Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.

留言 (0)

沒有登入
gif