Breast cancer

Assessment of breast cancer risk

Healthcare professionals may assess BC risk using these approaches^8,14,35:

• Identify patients at risk of a germline mutation and offer them formal genetic testing
• For patients who do not meet the criteria for genetic testing, or who test negative for germline mutations, quantify the risk of developing cancer over a specified length of time

With the resulting assessment information, surveillance or lifestyle, pharmacological or surgical interventions can be administered to improve a patient’s risk profile (Figure 6)³⁵.

Figure 6. Assessment in women at risk of breast cancer (Adapted³⁵).

BC risk is multifactorial. Hence, optimal risk assessment of BC involves consideration of clinicopathological, molecular and genetic factors.

Risk assessment of breast cancer: clinicopathological factors

Important clinicopathological factors used in BC risk assessment are^3,4:

• expression of ER/PR, HER2 and proliferation markers, such as Ki-67
• gene expression assay
• number of involved regional lymph nodes
• tumour histology
• tumour stage (tumour size, presence of peritumoural vascular invasion)
• histological grade

Immunohistochemically detected tumour markers are incorporated into the eighth edition of the American Joint Committee on Cancer (AJCC) Tumour, Node, Metastasis (TNM) to improve risk assessment, which also uses genomic assays to downstage some ER+, lymph node-negative tumours^3,4.

As asymptomatic distant metastases are rare, comprehensive laboratory testing, including tumour markers and radiological staging, is not necessary for all patients. Minimum blood work-up, including a full blood count, liver and renal function tests, alkaline phosphatase and calcium levels, is advised before surgery and systemic (neo)adjuvant therapy³.

In patients at higher risk of BC—high tumour burden, aggressive biology, signs, symptoms or laboratory values suggesting the presence of metastases—imaging of chest, abdomen and bone is recommended. ¹⁸F-fluorodeoxyglucose (FDG)-positron emission tomography (PET)/computed tomography (CT) scanning may be useful when conventional methods are inconclusive, and may replace traditional imaging for staging in high-risk patients³.

At a general level, clinicopathological parameters can fail to characterise the biological heterogeneity of tumours, which has important implications for treatment benefit. Traditional clinicopathological-based adjuvant chemotherapy (ChT) may not permit accurate individualised treatment. ChT overexposure is possible in low-risk groups, or in those who are already cured, and it can limit optimal treatment planning for groups at high risk of recurrence⁷.

Patients with low-risk HR+/HER2- breast cancers can be spared the toxicities of cytotoxic chemotherapies without compromising cancer prognosis^7,9,11.

Most recurrences after 5 years of endocrine therapy are not preventable by extending treatment beyond the initial 5 years. Ongoing adjuvant trials for the cyclin-dependent kinases 4 and 6 (CDK4/6) inhibitors may improve outcomes⁷

Several gene expression profiles are available to complement the limitations of clinicopathological factors in BC risk assessment. Used together, they are more accurately predictive of BC risk and benefit of adjuvant ChT³.

Risk assessment of breast cancer: molecular factors

Gene-based assays deploy gene-expression profiling in determining the molecular subtype of the cancer, including traditional molecular biomarkers, such as ER, PR or HER2. Traditional biomarkers are described in more detail in the breast cancer stratification section.

Genetic assays supplement the traditional histopathological markers in BC risk assessment and can help identify patients at high risk of recurrence. They also provide a more quantitative approach to risk assessment and facilitate individualisation of treatment.

Such advantages can benefit quality of life and healthcare costs, because patients who will not benefit from a certain treatment can be spared both the toxicity and the expense.

Gene-based assays include MammaPrint^®, Oncotype DX Recurrence Score^®, Prosigna^®, EndoPredict^® and Breast Cancer Index^® (BCI)³. Apart from MammaPrint^®, these assays were designed for patients with ER+ BC only³.

Mostly, the genes assessed in these assays are involved in the regulation of the cell cycle and ER responses.

The Oncotype DX Recurrence Score^® is a molecular tool that uses reverse transcription polymerase chain reaction (RT-PCR) assays, and incorporates 21 genes (including six housekeeping genes) related to proliferation, survival, invasion and oestrogen signaling. Expression levels of these genes equal a low, intermediate or high recurrence score (RS), thereby evaluating the likelihood of distant metastasis within the subsequent 10 years. It has shown prognostic value for the risk of recurrence 0–10 years after diagnosis, and may enable the prediction of benefit from adjuvant chemotherapy^3,7.

The clinical utility of MammaPrint^® and Oncotype DX^® has been, or is still being prospectively evaluated, in large randomised clinical trials such as MINDACT for MammaPrint^®, West German Study Group (WSG) PLAN B trial, TAILORx and RxPONDER (SWOG 1007) for Oncotype DX^®38–40.

A level of evidence (LoE) and grade of recommendation (GoR) of [I, A] (Table 2) have been achieved through the MINDACT trial for the prognostic value and clinical utility (for ChT decision making) of MammaPrint (for clinical high risk, low genomic score), and for Oncotype DX® through the TAILORx and PLAN B trials³. MammaPrint® and Oncotype DX® are able to identify patients with an ultra-low risk of death from BC at 10 or 20 years^41,42.

At the European Society for Medical Oncology congress in 2021, results from a single centre retrospective cohort study were revealed, showing that patients with low risk Oncotype DX scores^® who had high Ki67 tumours have a greater risk of relapse. Therefore, Ki67 status may be a helpful indicator of low risk Oncotype DX^®-scoring patients in which adjuvant chemotherapy may be of benefit⁴³.

Table 2. Levels of evidence and grades of recommendation (Adapted⁴⁴).

The PAM50 score combines the expression levels of 50 genes and tumour size, which are used to define the intrinsic subtypes of BC, and information from a set of proliferation-related genes for prognostic purposes. PAM50 has prognostic value for the prediction of recurrence risk between year 0 and year 10, with similar performance for the prediction of risk at years 0–5 and 5–10^3,7.

EndoPredict^® is a 12-gene expression panel that is combined with clinical parameters to generate an EPClin score. EPClin has good prognostic value for the prediction of BC recurrence 0–10 years after diagnosis and of late recurrence in a combined cohort from the ABCSG 6 and ABCSG 8 studies^3,7.

A score of [I, B] (Table 2) according to biomarker LoE guidelines has been achieved from retrospective analyses of data from prospective trials regarding the prognostic value of Prosigna^® and Endopredict^®, in ER+ BC^45,46. In addition, the prognostic value of MammaPrint^® has been validated in the RASTER trial, a prospective but non-randomised, clinical trial⁴⁷.

The Breast Cancer Index^® (BCI) is a molecular score that incorporates the ratio of HOXB13:IL17BR levels and a proliferation-related five-gene module referred to as Molecular Grade Index^3,7. In the Stockholm TAM data set, a heterogeneous cohort (in terms of BC risk) of 1,374 postmenopausal women treated with tamoxifen, BCI had prognostic value for recurrence beyond 5 years⁴⁸. In the TransATAC study, BCI had prognostic value for late recurrence in women with no lymph node involvement^3,7.

The OPTIMA Prelim Trial has recently shown that the use of molecular tests has an 86% probability of being cost-effective⁴⁹.

Clinical validity has been shown for PAM50 and BCI, but clinical utility and the predictive benefit of extended ET have not yet been established^3,7.

Risk assessment modelling

Predictive models have been validated to assess BC risk in populations and individuals. Individualised risk assessment is an essential part in the effective assessment of women at high risk of BC^35,50.

Identifying women at high-risk before breast cancer develops enables this group to consider more frequent screening with mammography and breast magnetic resonance imaging, genetic testing, and chemoprevention

Assessing the risk of carrying a germline mutation

Germline mutations associated with familial BC have been identified in other genes, including TP53, PTEN, ATM, CHEK2, NBS1, RAD50, BRIP, and PALB2, and others are suspected^3,8,35.

Several predictive models and scoring systems have been designed to assess the presence of a BRCA1 mutation or a BRCA2 mutation in a given individual based on family history^35,50. These models are empirical models and the genetic risk prediction models^35,50.

Empirical models estimate the probability that genetic testing will detect a BRCA1 or BRCA2 mutation. They do not make assumptions about the underlying genetic risks such as penetrance, mutation frequencies, and method of inheritance^35,50.

Empirical models include the Couch model, Myriad II, and the Australian LAMBDA model.

Genetic risk prediction models make direct assumptions about the number of susceptibility genes involved, the allele frequencies in the general population, and the cancer risks that are conferred by these alleles. These models use pedigree analysis methods, which are based on information about the exact relationships among individuals within a family^35,50.

The advantage of genetic risk prediction models is that they can compute cancer risks and mutation carrier probabilities regardless of the family structure and disease pattern. However, their accuracy depends on the accuracy of their underlying assumptions^35,50.

Examples include BRCAPRO, the Yale University model, and the International Breast Cancer Intervention Study (IBIS) model.

Comparison of germline mutation risk prediction models

In one UK study, data from six genetics clinics were used to compare the most commonly used models in the UK: Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA), BRCAPRO, IBIS, the Myriad II, and the Manchester model⁵¹. Only BOADICEA accurately predicted the overall observed number of mutations that were detected⁵¹. BOADICEA also provided the best discrimination between mutation carriers and noncarriers, and was statistically significantly better than all of the other models, except BRCAPRO (area under the receiver operating characteristic curve [AUC]: BOADICEA = 0.77; BRCAPRO = 0.76; IBIS = 0.74; Manchester = 0.75; Myriad II = 0.72)⁵¹. An AUC of 0.5 identifies a model whose discriminatory accuracy is no better than random chance. An AUC of 1.0 identifies a model with perfect discriminatory accuracy. Realistically, an AUC of 0.7 or 0.8 is consistent with good discrimination.

Incorporation of risk factors into risk assessment models

Current EBC risk prediction models are based on combinations of risk factors. In general, their outputs include a BC risk estimate over a specific time and/or over the lifetime of the patient.

Several predictive models have been designed for this purpose (Table 3).

Table 3. Known risk factors and their incorporation into current risk models (Adapted³⁵).

Gail model

The most widely used model for EBC risk assessment is the Gail model^52–55, which has been validated in three large population-based database studies^52–54. However, a systematic review reported that, although eight studies comprising almost 13,000 patients had shown that the Gail model is well calibrated, it has limited discriminatory accuracy⁵⁶. This limitation could explain the poor individualised risk assessment of the Gail model when tested in higher risk populations³⁵.

Claus model

The Claus model, developed using data from the Cancer and Steroid Hormone Study, uses family history to estimate risk in EBC. It incorporates a more comprehensive history than the Gail model and includes affected first- and second-degree relatives and the age at which cancers in those relatives were diagnosed. However, the Claus model has never been validated in an independent dataset³⁵. A further limitation of the Claus model is that it does not include any nonhereditary risk factors, such as hormonal or reproductive factors³⁵.

BRCAPRO model

In addition to assessing the likelihood of carrying a BRCA gene mutation, the computerised BRCAPRO model includes an extension software package that calculates overall BC risk. An advantage of this model is that it includes information on both affected and unaffected relatives. However, no nonhereditary risk factors have been included. This model may therefore underestimate BC risk in women who have nonhereditary risk factors³⁵.

Jonker model

Jonker et al.⁵⁷ published a genetic model to predict BC risk based on the family history of breast and ovarian cancers. In this model, familial clustering of breast and ovarian cancers is explained by three genes: BRCA1, BRCA2 and a hypothetical third gene called BRCAu. The hypothetical gene was modelled to explain familial clustering of BC that was not accounted for by the BRCA1 and BRCA2 genes. The original Jonker model does not include data on personal risk factors for BC.

The original Jonker model developed into the validated Claus extended model⁵⁸. This model includes estimates of the risk of bilateral BC, and of having three or more affected relatives. A limitation of the extended model is its inability to estimate risk in women with complex family histories and its validation in individual families rather than in an independent series⁵⁹.

IBIS model

The IBIS model⁶⁰ attempts to integrate family history, proxy measures of endogenous oestrogen exposure, and benign breast disease. Unlike the Claus and BRCAPRO models, the IBIS model allows for the presence of multiple genes of differing penetrance. Like the Jonker model, IBIS’s algorithm includes the likelihood of BRCA1 and BRCA2 mutations while allowing for a lower penetrance of BRCAu.

Practical limitations of risk models

Adoption, small family size, and lack of information about family history reduce the usefulness of all predictive risk models for EBC. Given the reluctance of people to discuss their medical conditions, particularly those involving BC, generations of family medical history may be lost to present-day patients who are receiving care in the era of genetic testing³⁵.

The foremost limitation of early breast cancer risk models is their reliance on known risk factors, despite data that show that up to 60% of breast cancers can arise in the absence of any known risk factors³⁵

Another limitation is that current risk models include information only about first- or second-degree relatives of the person who is being assessed. This may underestimate BC risk if there are many third-degree or higher relatives with BC³⁵.

Risk models for BC can be improved by incorporating population-specific risks, mutation prevalence, or tumour-specific characteristics. For example, BRCA1 mutations are associated with triple-negative grade 3 BC histology. The presence of this phenotype could allow for an increase in the estimated risk of carrying a BRCA1 mutation³⁵.

While numerous risk models are available in the BC setting, they have not been widely implemented to guide screening decisions in routine clinical decision-making. This is partly due to lack of clarity on which risk model to use, limited accuracy of risk models, and the time needed to perform risk assessment and interpret results⁶¹.

Future developments in breast cancer risk assessment

Advances in breast imaging that focus on reproducible measures of breast density may permit mammographic density to be included in future risk assessment modeling^35,62.

Genome-based research may add valuable data to BC risk prediction methods. One common variant in the CASP8 gene with a minor allele frequency greater than 5% is associated with BC risk⁶³. As a result, investigation of a range of high-risk genes and single-nucleotide polymorphisms (SNPs) in several genes that are associated with low risk of BC have been conducted to add knowledge of cancer genetics^29,64,65. However, because of the variable presence of some SNPs in the general population, large datasets are needed to identify, at genome-wide statistical significance levels, loci with small per-allele effect sizes³⁵.

While BOADICEA and IBIS show similar performance in assessing for BC risk, improvements in the accuracy of predictions are possible with hybrid models that incorporate the polygenic risk component of BOADICEA and the non-family-history risk factors included in IBIS⁶⁶.

Models that incorporate additional genetic and nongenetic risk factors and estimate risk of tumour subtypes may further improve BC risk prediction⁶¹.

Review

• Women at high-risk of breast cancer are those with strong family history of breast, ovarian or pancreatic cancer, diagnosis of breast cancer before the age of 50, diagnosis of triple-negative breast cancer before the age of 60, and personal history of ovarian cancer or second breast cancer^4,8,35
• It is critical that accurate and individualised risk assessment is conducted by the patient’s multidisciplinary team to select women most likely to benefit from prevention strategies or treatment⁴
• Gene-expression profiles, such as MammaPrint®, Oncotype DX Recurrence Score^®, Prosigna^®, Endopredict^® and Breast Cancer Index^®, can complement clinicopathological assessment of breast cancer risk and recurrence⁴
• To date, no risk assessment model can perfectly discriminate between families that do and do not have mutations, or between women who will and will not develop breast cancer³⁵

Risk stratification tools using clinicopathological features of the patient’s disease and its molecular profile are used to prevent overtreatment of low-risk patients and to plan early treatment or prevention strategies for those at high risk of cancer recurrence

Stratification could also be important for selecting patients for clinical trials. It could facilitate the discovery of novel drivers, the study of tumour evolution, and the identification of mechanisms of treatment resistance.

While only 6–10% of the women are diagnosed with de novo metastatic breast cancer (mBC), most women with mBC have been previously diagnosed with locally advanced disease and subsequently have cancer recurrence in the form of metastasis⁶. Approximately 30% of these patients will develop incurable cancer recurrence^6–8.

Risk of cancer recurrence is high in patients with hormone receptor-positive (HR+) and oestrogen receptor-positive (ER+) BC. Relapse of BC may occur as late as >20 years after the initial diagnosis, particularly in patients with ER/progesterone receptor-positive (PR+) BC³.

The higher the risk of cancer recurrence, the more aggressive the therapy. Thus, stratification can prevent chemotherapeutic overtreatment in some patients at low risk of recurrence, or in those who are cured, and it can enable planning suitable treatments or prevention strategies in those at high risk of recurrence^3,7.

A clinical decision-making tool has been proposed for the combined use of clinical and genomic tests for risk stratification of patients upon completion of 5 years of endocrine therapy (ET) without distant recurrence (Figure 7)⁷.

Figure 7. Decision-making aid for clinical and genomic testing (Adapted from Richman & Dowsett et al.⁷). BCI, Breast Cancer Index; CTS5, Clinical Treatment Score at 5 Years; PAM50, Prediction Analysis of Microarray 50; EPClin, EndoPredict^®.

In Figure 7, the clinical treatment score at 5 years (CTS5) is calculated for all women upon completion of 5 years of adjuvant ET. For most women with a low CTS5 score, ET can be discontinued because extended therapy is unlikely to benefit them. For most women with a high risk of recurrence, extended ET up to 10 years is recommended if the toxicity profile is favourable. In both situations, genomic testing is unlikely to add further prognostic information and is not recommended. Women with an intermediate clinical risk and those at borderline low-intermediate or high-intermediate clinical risk should receive a genomic test to enable integrated clinical-genomic stratification of their risk of late recurrence. Following genomic testing, the following scenarios are possible: discontinuation of ET for patients with a low risk of recurrence or recommendation of extended ET for up to 10 years in patients with a high risk. Women who remain at an intermediate level of risk should discuss toxicities and personal preferences with their clinician.

Role of hormone receptors and biomarkers in breast cancer stratification

Oestrogen receptor

ER is arguably the most important prognostic biomarker in BC, because of the development of targeted therapy with tamoxifen or aromatase inhibitors.

The application of 5 years of adjuvant tamoxifen-based therapy in ER+ BC showed a 29% reduction of the risk for death from the disease^36,67,68. In approximately 30–40% of patients with advanced ER+ BC, response to the treatment is likely to be positive. In approximately 20% of patients, stable disease could be achieved. Hormone therapy is generally free of toxicity, permitting long-lasting use⁶⁹.

Progesterone receptor

In EBC, PR expression is associated with tumour grade, ER expression, Nottingham prognostic group and HER2- status⁷⁰. Evidence shows a better prognosis in PR+ cancers⁷⁰. The evaluation of PR expression does not appear to have a role in the ET choice in both locally advanced and mBC⁷¹.

PR is expressed in 60–70% of invasive ductal carcinomas. The correlation between ER and PR expression is high, but 10% of ER+ cancer can be PR-. In these patients, the risk of recurrence and mortality compared to ER+/PR+ cancer are higher⁶⁹.

High expression of PR protein is more frequently observed in tumours with a better baseline prognosis (luminal A) than in tumours with a poor baseline prognosis (luminal B).

Ki-67

No significant relationship between proliferation index (Ki-67) and prognostic factors such as hormone receptors and HER2 has been found⁶⁹. Furthermore, no significant correlation was observed between Ki-67 and disease-free survival (DFS) at 3 and 5 years or with overall survival (OS) at 5 years⁷³. Therefore, the role of Ki-67 in patient stratification is limited.

Although Ki-67 may be used to discriminate higher risk groups in the context of ER+/PR+ BC, no consensus has been agreed on the cut-off⁶⁹.

HER2

HER2+ is more frequently found in ER- tumours than in ER+ tumours. Only 12 % of ER+ BCs are positive for HER2 by amplification/overexpression of the gene⁷⁴.

Tumours with HER2 amplification are associated with a less favourable prognosis when compared with tumours having similar morphological features, but lacking amplification of the gene or the overexpression of the protein⁶⁹.

Although HER2 was originally proposed as a prognostic biomarker for BC, currently its usefulness lies in predicting response to anti-HER2 therapy in neoadjuvant and adjuvant settings⁶⁹.

Amplification/overexpression of HER2 is a necessary condition for the administration of anti-HER2 therapies. Of the anti-HER2 treatments available, trastuzumab is the most studied.

HER2-low

Almost half of the currently defined HER2-negative BCs show some degree of HER2 immunohistochemistry (IHC) expression and have been recently renamed as HER2-low BC⁷⁵. This is a recent designation defined as tumours with IHC staining of 1+, or 2+ without HER2 gene amplification⁷⁶. These tumours are targets for antibody–drug conjugates (ADCs), one of which, trastuzumab deruxtecan (T-DXd), was approved by the FDA in August 2022 and the EMA in January 2023 for treatment of metastatic BC in patients with HER2-low cancers who have received one or two chemotherapies or developed recurrent disease during or within 6 months of completing chemotherapy⁷⁷. This was based on results from the DESTINY-Breast04 clinical trial, where treatment with T-DXd resulted in significantly longer progression-free and overall survival than the physician’s choice of chemotherapy^77,78.

Although hormone receptors and HER2 overexpression currently represent the main protagonists of targeted therapy for BC, chemotherapy (ChT) remains the standard of care for tumours lacking ER, PR and HER2 expression. Consequently, much effort in BC research involves study of targetable molecules with predictive purposes in triple negative tumours⁶⁹.

Triple-negative breast cancer

A meta-analysis of triple-negative cancer (TNBC) defined seven clusters with different molecular profiling⁷⁹:

• Basal-like 1 (BL1), basal-like 2 (BL2)
• Immunomodulatory (IM)
• Mesenchymal (M)
• Mesenchymal stem-like (MSL)
• Luminal androgen receptor (LAR)
• Unstable cluster (UNS)

BL1 tumours are characterised by genetic alteration of genes involved in cell-cycle regulation, including the higher rate of mutations of TP53, amplifications of MYC, CDK6 and CCNE1, and deletions of BRCA2, PTEN, MDM2 and RB1⁸⁰.

BL2 tumors are characterised by genetic alterations involving growth factor signaling pathways and metabolic pathway activity, including EGF, NGF, MET, Wnt/β-catenin and IGF1R⁸¹.

As a group, basal-like tumours can have high proliferative activity, as shown by high Ki-67 mRNA expression or nuclear Ki-67 immunohistochemical staining, high metabolic activity, and genetic instability⁷². Most BRCA-mutated BCs are part of the basal-like groups⁸².

IM tumors are characterised by mutation of genes involved in antigen processing and cytokine signaling, such as JAK/STAT, TNF and NFkB. The prevalent expression of genes related to cell motility and epithelial mesenchymal transition are the genetic clues of mesenchymal-like subtypes, including M and MSL subtypes⁸³.

LAR subtype is characterised by the high expression of Androgen receptor (AR) with frequent mutations involving genes like PIK3CA, KMT2C, CDH1, NF1 and AKT1⁸⁰.

Emerging biomarkers

Programmed cell death receptor ligand 1 (PD-L1) is considered the most relevant predictive biomarkers of response to anti-PD1/PD-L1 immune checkpoint inhibitors. High levels of PD-L1 expression have been found in several types of cancers, and are generally associated with a poor prognosis. PD-L1 is overexpressed in a subset of basal-like and TNBCs, and its expression mainly occurs on tumour-infiltrating immune cells than on tumour cells⁸⁴.

Vascular endothelial growth factor (VEGF) seems to be the most important angiogenetic factor in BC, promoting growth of blood vessels that support the neoplasm and metastatic dissemination. High levels of VEGF have been correlated to poor prognosis and TP53 mutation in BC. Consequently, immunohistochemical evaluation of VEGF expression may have a prognostic role in the near future⁶⁹.

Microsatellite instability (MSI) is a phenotype defined by defects in mismatch repair proteins MLH1, PMS2, MSH2 and MSH6, leading to replication errors in segments of DNA characterised by short nucleotide repeat motifs known as microsatellites. MSI is an opportunity to select patients who may benefit from immune checkpoint inhibitors, especially among TNBC⁸⁵.

AXL, deriving from the Greek “anexelekto”, meaning “uncontrolled”, is a gene located in the q13.2 locus of chromosome 19, and encodes a membrane-located receptor with tyrosine kinase activity, belonging to the TAM receptors family. AXL deregulation has been shown in several neoplasms, including breast carcinoma; therefore, it is a potential target for targeted therapy⁶⁹.

ADAPTlate

ADAPTlate is a prospective, multicentre, interventional, two-arm, open, randomised, controlled adjuvant phase 3 trial, to investigate the additional benefit from 2 years of the CDK4/6 inhibitor abemaciclib combined with ET, compared with ET alone, in patients with high-risk HR+/HER2- EBC⁸⁶.

The study seeks to improve adjuvant therapy for patients at high risk for late-disease recurrence who completed definite locoregional therapy, with or without (neo)adjuvant ChT, and are under adjuvant ET. This high-risk population does not benefit as much from standard ET and can go on to develop secondary ET resistance and late recurrences⁸⁶.

Translational research aims to evaluate the potential mechanisms of resistance to ET and/or CDK4/6 targeted therapy. Exploratory tissue biomarker research will assess alterations in molecular markers and liquid biopsies, such as circulating tumour cells (CTC), circulating tumour DNA (ctDNA), and circulating tumour RNA (ctRNA), to study mutations and gene expression relevant for HR+/HER2- EBC⁸⁶.

Recruitment is ongoing as of 2021. Eligible patients will be pre-/post- menopausal patients with histologically confirmed invasive HR+/HER2- EBC, 2–6 years after primary diagnosis, with either⁸⁶:

• a known high clinical risk (clinical node staging / pathological assessment of the regional lymph nodes [c/pN] 2–3 or high clinical treatment score (CTS) score in pN 0–1 or non-pathological complete response (pCR) after neoadjuvant CT in cN 1 or G3 tumours or G3 and Ki-67 ≥40% in pN 0–1)
• a known high genomic risk (RS >25 in c/pN 0, RS >18 in c/pN 1 or high-risk Prosigna, EPclin, or Mammaprint in pN 0–1)
• an intermediate clinical, but unknown, genomic risk (luminal B-like (G3 or Ki-67 ≥40%) in c/pN 0–1 and either RS >25 in c/pN 0, or RS >18 in c/p N1 in screening)

The duration of treatment will be two years for abemaciclib + ET (premenopausal: aromatase inhibitor [AI] + gonadotropin-releasing hormone [GnRH]) followed by 3–6 years ET alone. Patients in control arm will receive 5–8 years of ET at the investigator’s choice⁸⁶.

Review

• Knowledge of the molecular mechanisms underlying BC has led to the identification of different biomarkers^12–14
• Based on their prognostic value, traditional biomarkers associated with BC, such as oestrogen receptor (ER), progesterone receptors (PR), or HER2, enable clinicians to stratify patients who could be eligible for adjuvant therapy, or at low or high-risk of cancer recurrence^3,4,12–14
• Several emerging biomarkers may further improve prognostic stratification^69,84,85

Further investigation is needed to determine if patients are stratified into clinically meaningful groups, and whether the resulting treatment changes ultimately improve clinical outcomes

Treatment options for early breast cancer

Treatment of EBC is complex and involves multiple options. These include local modalities (surgery, radiotherapy (RT)], systemic anticancer treatments (chemotherapy [ChT], endocrine therapy [ET], molecularly targeted therapies) and supportive measures, delivered in diverse sequences³.

Prognostic and predictive biomarkers such as oestrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), Ki-67, and approved genomic signatures can help determine suitable treatments (Figure 8)³.

Figure 8. Early breast cancer treatment algorithm (Adapted⁴). BCS, breast-conserving surgery; ChT, chemotherapy; ET, endocrine therapy; HER2, human epidermal growth factor receptor 2; RT, radiotherapy; TNBC, triple-negative breast cancer. To view levels of evidence and grades of recommendation, see Table 2.

Surgery

Table 4 provides an overview of current local surgical treatments for EBC³.

Table 4. An overview of current local surgical treatments for early breast cancer (Adapted³).

Adjuvant systemic treatment

The decision on adjuvant/neoadjuvant systemic treatment is based on the predicted patient sensitivity to particular treatment types, the benefit from their use, and the patient’s risk of relapse (Figure 9)⁴. Relapse relies on the patient’s tumour burden and tumour biology.

Figure 9. Adjuvant systemic treatment choices for early breast cancer by marker expression and intrinsic phenotype (Adapted⁴). ChT, chemotherapy; ER, oestrogen receptor; ET, endocrine therapy; HER2, human epidermal growth factor receptor2; N0, node-negative; PR, progesterone receptor; TNBC, triple-negative breast cancer. To view levels of evidence and grades of recommendation, see Table 2.

The decision on adjuvant systemic treatment also incorporates the predicted short- and long-term toxicities of treatment, the patient’s age, general health status, comorbidities and preferences⁴.

Decision-making tools such as PREDICT Plus^®, NPI^® and Adjuvant! Online^® can help predict recurrence risk and potential benefit from systemic treatments (Table 5)⁴

Table 5. Summary of biomarkers for treatment decision making (Adapted⁴).
ChT, chemotherapy; ER, oestrogen receptor; ET, endocrine therapy; GoR, grade of recommendation; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridisation; LoE, level of evidence; PR, progesterone receptor; RT-PCR, reverse transcription polymerase chain reaction. To view levels of evidence and grades of recommendation, see Table 2.

Genomic tests are not recommended for patients with clinicopathological low-risk tumours (pT1a, pT1b, G1, ER high, pN0) or patients with comorbid health conditions who are not candidates for adjuvant ChT³.

Premenopausal and postmenopausal women

For premenopausal women, tamoxifen for 5–10 years is a standard of care. For postmenopausal women, aromatase inhibitors (AIs) and tamoxifen are considered standard treatments³.

Chemotherapy

Chemotherapy (ChT) is advised for most patients with most triple-negative, HER2+ BC and those with high-risk luminal-like HER2- tumours. Sequential anthracycline/taxane-based regimen is the standard for most patients with EBC. In some lower risk patients, 4 cycles of anthracycline- or taxane-based ChT or cyclophosphamide/methotrexate/5-fluorouracil (CMF) may be used³.

Taxanes can improve the efficacy of ChT, independently of age, nodal status, tumour size or grade, steroid receptor expression or tamoxifen use. Taxanes allow for the use of a lower total dose of anthracyclines through the use of sequential regimens³.

Anti-HER2 therapy

Trastuzumab combined with ChT in patients with HER2 overexpression/amplification roughly halves recurrence and mortality risk, compared with ChT alone, signifying a 10% improvement in long-term DFS and 9% increase in 10-year OS. One year of adjuvant trastuzumab remains a standard for the vast majority of HER2+ EBC (Figure 10).

Figure 10. HER2-positive breast cancer treatment (Adapted⁴). 
ChT, chemotherapy; EMA, European Medicines Agency; ER, oestrogen receptor; ESMO-MCBS, ESMO-Magnitude of Clinical Benefit Scale; HER2, human epidermal growth factor receptor 2; N-positive, node-positive; pCR, pathological complete response; T-DM1, trastuzumab emtansine. To view levels of evidence and grades of recommendation, see Table 2. 

Personalised medicine

ER, PR and HER2 status are routinely used as prognostic/predictive factors to select patients for targeted ET or anti-HER2 treatment³.

More recently, surrogate intrinsic tumour phenotypes, based on biomarker expression, have been used for personalised treatment. Urokinase plasminogen activator-plasminogen activator inhibitor 1 (uPA-PAI1), a biomarker of tumour invasiveness, has been validated in prospective clinical trials as a prognostic marker for both node-negative and node-positive BC and can be used in treatment decision-making for EBC⁹³.

Molecular signatures for ER+ BC such as MammaPrint^®, Oncotype DX^®, EndoPredict^®, Prosigna^® and Breast Cancer Index^® may help with adjuvant ChT decision making, together with clinicopathological features, in challenging cases of BC, such as luminal B-like/HER2- or node-negative/nodes 1–3+ BC³.

PARP inhibitors in early breast cancer

The OlympiA study was the first randomised, double-blind, phase 3 study, to investigate the use of the poly (ADP-ribose) polymerase (PARP) inhibitor olaparib as adjuvant therapy following adjuvant or neoadjuvant ChT in patients with gBRCA mutations and HER2− (TNBC or HR+) high-risk EBC⁹⁴. Results from the OlympiA study showed that adjuvant treatment with olaparib, after completion of neoadjuvant or adjuvant ChT, significantly improved  invasive disease-free survival (iDFS) and distant disease-free survival (DDFS)⁹⁴. Notably, these results have now been incorporated into the June 2021 American Society of Clinical Oncology (ASCO) rapid guidelines recommendation update. Specifically, guidelines now recommend offering 1 year of adjuvant olaparib to patients with high-risk early-stage HER2− BC and germline BRCA mutations after completion of (neo)adjuvant chemotherapy and local treatment⁹⁵.

In March 2022, olaparib received FDA approval for adjuvant treatment of adult patients with deleterious or suspected deleterious germline BRCA-mutated (gBRCAm) HER2−, high-risk EBC after treatment with neoadjuvant or adjuvant chemotherapy. The approval stipulates that patients must be selected for therapy using an FDA-approved companion diagnostic for olaparib⁹⁶.

CDK4/6 inhibitors in early breast cancer

Following the success of cyclin-dependent kinases 4 and 6 inhibitors (CDK4/6 inhibitors) in the advanced breast cancer (ABC) setting for HR+/HER2- BC, and evidence in EBC from the phase 2 NeoPalAna trial (palbociclib plus anastrozole), several phase 2/3 clinical trials have assessed CDK4/6 inhibitor efficacy in combination with ET in the adjuvant setting^97,98. These trials include monarchE for abemaciclib⁹⁹, PALLAS and PENELOPE B for palbociclib^100,101, and NATALEE and Ear-LEE-1 for ribociclib^102-103.

The monarchE trial showed that adding abemaciclib to standard-of-care adjuvant ET, including ovarian function suppression for premenopausal women, reduced the risk of cancer recurrence⁹⁹. This was the first clinical trial to establish a role for CDK4/6 inhibition in the adjuvant treatment of EBC.

The outcome of the monarchE trial led to approval by the US Food and Drug Administration and the European Medicines Agency for the use of abemaciclib in combination with ET as adjuvant treatment of adult patients with HR+/HER2−, node-positive, early breast cancer (EBC) at high risk of recurrence^104,105

Learn more about clinical trials for CDK4/6 inhibitors in early and advanced breast cancer

Neither the PALLAS study nor the PENELOPE B trial demonstrated a statistically significant improvement in iDFS with palbociclib + ET, compared with ET alone, therefore failing to meet their primary endpoints^100,106,107. By contrast, at a pre-planned interim analysis in March 2023, it was announced that the NATALEE trial had met its primary endpoint, achieving a statistically significant improvement in iDFS in people with HR+/HER2– EBC treated with ribociclib + ET when compared with ET alone¹⁰⁸. Further results from the NATALEE clinical trial are expected to be released in late 2023.

Review

• Treatment of BC in specialised breast units or centres, provided by specialist multidisciplinary teams, improves patient breast cancer outcomes^3,4
• The treatment landscape for early BC is complex and comprises local modalities, systemic anticancer and supportive measures³
• Prognostic and predictive biomarkers such as ER, PR, human epidermal growth factor receptor2, and Ki-67 and approved genomic signatures, can help determine suitable personalised therapies for targeted endocrine therapy or anti-human epidermal growth factor receptor2 treatments³
• Breast-conserving surgery is the preferred local treatment for many patients with early breast cancer³
• Oncoplastic techniques are used when needed, and breast reconstruction should be available and proposed to all women requiring mastectomy³
• Chemotherapy is recommended in most triple-negative, human epidermal growth factor receptor 2+ breast cancers, and in high-risk luminal-like human epidermal growth factor receptor 2- Sequential anthracycline/taxane-based regimen is standard-of-care for most patients with early breast cancer^3,4
• The PARP inhibitor olaparib is approved for adjuvant treatment of deleterious or suspected deleterious germline BRCA-mutated (gBRCAm) HER2−, high-risk EBC after treatment with neoadjuvant or adjuvant chemotherapy
• The CDK4/6 inhibitor abemaciclib in combination with ET is now an option for adjuvant treatment of patients with HR+/HER2−, node-positive early breast cancer at high risk of recurrence

The final decision on adjuvant systemic treatment is based on the patient’s expected sensitivity to specific treatment types, treatment benefits and toxicities, and the patient’s risk of relapse (comprising the patient’s tumour burden and tumour biology), age, overall health status, comorbidities and personal preferences³

Treatment options for advanced breast cancer

Advanced breast cancer (ABC) comprises both inoperable locally advanced breast cancer (LABC) and metastatic breast cancer (mBC)⁴.

Treatment of ABC is complex. Involvement of all appropriate specialties in a multidisciplinary team (MDT), not restricted to medical, radiation and surgical oncologists, imaging experts, pathologists, gynaecologists, psycho-oncologists, social workers, nurses and palliative care specialists, is essential⁴.

Following a thorough assessment and confirmation of mBC, the potential treatment goals of care are discussed with the patient and family. Treatment should be personalised for the individual patient⁴.

mBC remains largely incurable; however, as survival is improving in some patients with ABC, consideration of survivorship issues is part of routine care. Sensitive healthcare professionals are ready to change and adapt treatment strategies to disease status, treatment of adverse effects and quality of life, patients’ priorities and life plans⁴.

With some patients achieving long-term relapse-free survival, discussion is now ongoing as to whether a subset of patients may be curable with a more aggressive MDT approach and curable intent¹⁰⁹.

In the absence of a cure, the goal of treatment must be to extend life and delay disease progression, while ensuring that patients continue to have the best quality of life for as long as possible¹¹⁰.

Patient involvement in their care has been associated with positive patient outcomes¹¹¹

Treatment choice for ABC should take at least these factors into consideration⁴:

• Hormone receptor (HR), human epidermal growth factor receptor 2 (HER2), and germline BRCA status
• Phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) in hormone receptor-positive (HR+) and programmed death-ligand-1 (PD-L1) in triple-negative breast cancer (TNBC), if targeted therapies are accessible
• Previous therapies and their toxicities
• Disease-free interval (DFI)
• Tumour burden (defined as number and site of metastases)
• Age of the patient
• Performance status (PS)
• Comorbidities (including organ dysfunctions)
• Menopausal status (for ET)
• Need for rapid disease/symptom control
• Socio-economic and psychological factors
• Available therapies in the patient’s country
• Patient’s preference

Most cases of mBC are treated using systemic therapy, which includes endocrine therapy (ET), chemotherapy (ChT) and various targeted therapies, such as the oral cyclin dependent kinase 4/6 inhibitors (CDK4/6 inhibitors), abemaciclib, palbociclib and ribociclib^3,112. These may be used as single agents or in combination with one another.

Current targets of systemic therapies include a range of key pathways, such as¹¹²:

• oestrogen receptor (ER) signalling
• HER2 signalling
• phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) signalling

HR and HER2 status of the cancer is a particularly important consideration when selecting a treatment strategy⁴.

Hormone receptor-positive, human epidermal growth factor receptor 2-negative metastatic breast cancer

ET is considered a mainstay of HR+/HER2- mBC treatment, and is internationally recommended as a first-line systemic treatment for this disease subtype^3,113. This is in preference to chemotherapy (ChT), which is usually reserved for those with rapidly progressing visceral disease^3,114.

A range of differentETs is currently available for the treatment of HR+ mBC (Table 6). Ensuring that patients receive the treatment best suited to their individual needs requires careful consideration of the following factors^3,114:

• Menopausal status
• Comorbidities
• Prior adjuvant therapy
• Drug availability
• Patient preference
• Drug safety profile

Table 6. Summary of endocrine therapies available to treat hormone receptor-positive metastatic breast cancer (Adapted⁸⁴).

Continuation of ET is an option following disease progression on first-line ET, provided the patient derived some benefit from their initial therapy, and unless they are experiencing symptoms of the underlying disease or the disease is progressing rapidly¹¹⁴.

Failure of three lines of ET usually necessitates the use of ChT, although continuation of ET may be considered reasonable if disease progression remains slow and the patient is asymptomatic^113,114.

Premenopausal and postmenopausal women

In postmenopausal women, an aromatase inhibitor (AI) such as anastrozole or letrozole may be used as a form of endocrine monotherapy in the first line of treatment. Similarly, tamoxifen, a selective oestrogen receptor modulator (SERM), or fulvestrant, a selective oestrogen receptor degrader (SERD), may initially be used to treat this group of patients¹¹².

In premenopausal women, treatment options mirror those available to postmenopausal women, but additional treatment with ovarian suppression drugs such as leuprolide or goserelin, or ovarian ablation are required in order to reduce oestrogen levels^4,6. These patients may also be treated with a selective oestrogen receptor modulator (SERM)^3,115.

CDK4/6 inhibitors as a treatment option in advanced breast cancer

Three oral CDK4/6 inhibitors, abemaciclib, palbociclib and ribociclib, are available for the treatment of HR+/HER2- mBC following the success of a series of trials investigating their safety and efficacy¹¹⁶. All three CDK4/6 inhibitors are indicated in patients with local ABC^104,117,118.

The CDK4/6 inhibitors are now widely recommended for use in first-line treatment in combination with either an AI or fulvestrant. This recommendation applies to postmenopausal women, and to premenopausal women in combination with an luteinising hormone-releasing hormone (LHRH) agonist⁴.

Abemaciclib is also approved by the US Food and Drug Administration (FDA) as a monotherapy in HR+/HER2- mBC, in adults who progressed following ET and ChT. Regimens that include CDK4/6 inhibitors may also be used as second-line treatment in cases where they have not been used as first-line treatment and have shown promising results in the later line setting^112,114.

It is not known if CDK4/6 inhibitors should be preferably administered in the first- or second-line ABC setting. However, the majority of panellists preferred giving a CDK4/6 inhibitor in the first-line setting for most patients⁴.

No data currently support the use of a combination of CDK4/6 inhibitor and ET as maintenance therapy following ChT. Maintenance therapy, in this case, is delivered with ET alone⁴.

Using artificial intelligence to predict CDK4/6 treatment response in advanced breast cancer

Although treatment with CDK4/6 inhibitors is supported by evidence in ER+/HER2− BC, side effects and financial considerations have contributed to limited administration. In addition, predictive biomarkers for CDK4/6-targeted therapies are currently lacking, despite ongoing research in the field¹¹⁹.

A retrospective analysis was recently conducted that examined CDK4/6 inhibitor response using a database of over 1,000 genomes from patients with BC (171 of whom were treated with palbociclib and ribociclib). The analysis incorporated genomics, molecular biology and machine learning algorithms. The results indicated that the genomic signature of CDK4/6 inhibitors, comprising nine biomarkers, improved the predictive stratification of patients with advanced ER+/HER2− BC. This is promising research that may help clinicians identify patients most likely to benefit from CDK4/6 inhibitor treatment¹¹⁹.

Endocrine resistance

Endocrine resistance is common among HR+ mBC patients receiving ET and results in disease progression¹²⁰.

Dysregulation of the phosphatidylinositol-3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway frequently occurs in BC, and hyperactivation of the PI3K pathway is associated with acquired resistance to endocrine therapy^120,121. mTOR, which comprises two multi-protein complexes, is a downstream effector of the PI3K/AKT pathway¹²².

Everolimus restricts the growth of tumour cells via inhibition of mTOR complex 1 and is often given in combination with exemestane as second- or later-line treatment in postmenopausal patients who have developed resistance to non-steroidal AIs¹²².

In addition, a second molecule targeting a different part of the PI3K/AKT/mTOR pathway, alpelisib, was recently approved by the US Food and Drug Administration (FDA)^3,123. This PI3Kα-specific inhibitor can be given in combination with fulvestrant to postmenopausal patients who have undergone unsuccessful treatment with an AI and have HR+/HER2-, PI3Kα mutated mBC^3,124.

The HER2- status of these patients means that among those with a germline BRCA1/2 mutation, poly-ADP ribose polymerase (PARP) inhibitors may also prove an effective treatment option¹¹².

A recent retrospective study evaluated the distribution and incidence of de novo PIK3CA mutations in HR+/HER2- metastatic breast cancer, enrolling patients on endocrine therapy being treated with CDK4/6 inhibitors, whose initial formalin-fixed paraffin-embedded tissue samples were available. Results indicate that a worse clinical outcome is associated with HR+/HER2- metastatic breast cancer patients who exhibit the presence of de novo PI3K mutation. This is despite the use of CDK4/6 inhibitors. This study concludes that clinicians should consider assessing the presence of de novo PIK3CA mutation, upfront¹²⁵.

Hormone receptor-positive, human epidermal growth factor receptor 2-positive metastatic breast cancer

It has been estimated that overexpression or gene amplification of HER2 occurs in 20–30% of BC cases^4,85. Of these, around half are also HR+^4,126. Overexpression of HER2 promotes tumour cell proliferation and is associated with significantly shortened disease-free survival and overall survival in patients with BC^4,127. HER2 signalling pathways, therefore, represent an important treatment target in patients with HER2+ cancers. The addition of multiple anti-HER2 targeted therapies to oncologists’ treatment arsenal has resulted in significant improvements to patient outcomes in this group, even in the advanced setting. However, despite the benefits of HER2-targeted therapies, de novo and acquired resistance to these therapies is very common^4,128.

Currently available HER2-targeted therapies include trastuzumab, pertuzumab, lapatinib, ado-trastuzumab emtansine (T-DM1) and T-DXd. Standard first-line therapy in patients with HER2+ mBC is trastuzumab and pertuzumab taken in combination with a taxane^3,114,129. This is generally followed by trastuzumab emtansine in the second line of treatment^3,114,129. If the disease continues to progress, a new ChT or ET with continued HER2 blockade can be used, even if previous HER2-targeted therapy has failed to prevent disease progression^4,6.

Hormone receptor-negative, human epidermal growth factor receptor 2-positive metastatic breast cancer

HR−/HER2+ mBC is the least common of the four subtypes¹³⁰. Treatment options for patients with this type of breast cancer are very similar to those available to patients with the HR+/HER2+ subtype, starting with dual blockade or HER2 in combination with ChT. However, later-line therapy usually comprises a ChT, rather than ET, in combination with an HER2-targeted agent⁶.

Triple-negative breast cancer (TNBC)

Metastatic TNBC has the worst prognosis of the four BC subtypes, lacking the molecular targets of both endocrine and HER2-targeted therapies1^4,131. For this reason, chemotherapeutic approaches are widely recommended for the treatment of this group of patients. Monotherapy in the form of an anthracycline, a taxane, capecitabine or eribulin is common in both first- and later-line treatment^4,6.

PARP inhibitors may also be used to treat patients with this disease subtype, provided they have a germline BRCA1/2 mutation^4,94,131,132.

Sacituzumab govitecan is an ADC that targets Trop-2 for the selective delivery of SN-38, the active metabolite of irinotecan. It is approved by the EMA and FDA for the treatment of adult patients with unresectable locally advanced or metastatic TNBC who have received at least two prior therapies, at least one of them for advanced disease^133,134.

The immune checkpoint inhibitor pembrolizumab is approved by the EMA and FDA for the treatment of high-risk early-stage TNBC as neoadjuvant therapy in combination with chemotherapy and as a single agent for adjuvant therapy. Pembrolizumab, in combination with chemotherapy, is also approved for treatment of locally recurrent unresectable or metastatic TNBC in adults whose tumours express programmed death ligand 1 (PD-L1) with a combined positive score (CPS) of ≥10 and who have not received prior chemotherapy for metastatic disease^135,136.

Erβ is another potential treatment target for TNBC and a phase 2 trial for oestradiol is currently in progress^24,137.

Review

• Treatment of ABC is complex, and involves all suitable specialties in a multidisciplinary team⁴
• Treatment should be personalised for each patient. As survival is improving in some patients, integrating survivorship issues into optimal care is essential⁴
• Treatment choice for ABC should consider a range of complex, interacting factors, including hormone receptor, human epidermal growth factor receptor 2 status, and germline BRCA status, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha in hormone receptor-positive and programmed death-ligand-1 in TNBC, disease-free interval, comorbidities, and patient preferences⁴
• Most patients with mBC are treated with systemic therapy, comprising ET, ChT, or targeted therapies^4,112
• Continuation of ET is an option following disease progression on first-line ET^4,109
• In premenopausal women, treatment options are the same as those available to postmenopausal women, but further treatment with ovarian suppression drugs may be required in this group^4,6,112
• Everolimus with exemestane as second- or later-line treatment in postmenopausal patients is recommended in patients who develop endocrine resistance^3,122
• HER2 signalling pathways are an important treatment target in patients with HER2+ cancers^3,126,128
• For TNBC, chemotherapeutic approaches are advised^4,131,132

CDK4/6 pathway

CDK4/6 and cell proliferation

The mammalian cell cycle is a tightly regulated, ordered process that is active through four distinct phases termed G1, S, G2 and M (Figure 11)¹⁴⁸.

Figure 11. The classical model of the mammalian cell cycle (Adapted¹⁴⁸). CDK 1,2,4 or 6, cyclin dependent kinase 1,2,4 or 6; G1, gap 1 phase; G2, gap 2 phase; M, mitosis; S, DNA synthesis.

Between each phase is a controlled checkpoint that is regulated by cyclins and cyclin-dependent kinases (CDKs)¹⁴⁸. Cyclin-dependent kinases (CDKs) belong to a family of serine/threonine protein kinases that interact with specific cyclins to promote cell cycle progression (as well as other functions). Cyclins are a diverse family of proteins divided into four classes (A, B, D and E) with multiple members within each class, such as cyclin D1, D2 and D3.

The role of CDK4/6 in the cell cycle

During the G1 phase of the cell cycle, cells must overcome the restriction point, which is a checkpoint that, once passed, commits a cell to the S phase and another cycle of cell division¹⁴⁸. Progression through G1 and the restriction point requires synthesis of cyclin D and the formation of a cyclin D:CDK4/6 complex¹⁴⁸. A key function of the cyclin D:CDK4/6 complex is to overcome the inhibitory function of the retinoblastoma (RB) protein that binds and sequesters the transcription factor E2F¹⁴⁸.

Cyclin D:CDK4/6 complexes phosphorylate RB reducing its ability to bind E2F and enabling E2F to drive the transcription of genes required for the S phase of the cell cycle (Figure 12)^149,150. These include cyclin E, which binds CDK2 and promotes progress through the S phase and genes involved in DNA replication.

CyclinD:CDK4 has been shown to directly promote other pathways involved in cell proliferation, migration and the DNA damage response, while CDK6 may play a kinase-independent role in angiogenesis^149,151.

Figure 12. The role of cyclin D:CDK4/6 in breast cell proliferation (Adapted¹⁴⁹). CDK, cyclin-dependent kinase; E2; oestrogen; E2F, E2 factor; ER, oestrogen receptor; HER, human epidermal growth factor receptor; P, phosphorylation; RB, retinoblastoma protein.

Role of CDK4/6 in breast cancer pathogenesis

Escape of senescence and progression through the cell cycle is essential for cancer development and progression. A number of strategies are utilised by BC cells to achieve this, many resulting in an increase in cyclin D-dependent activity¹⁴⁹.

Amplification of CDK4 and cyclin D1 has been observed in 15–25% of BCs, while overexpression of cyclin D1 has been reported in over half of all BCs¹⁴⁹. Interestingly, amplification of cyclin D1 and CDK4 genes differs between BC subtypes. Cyclin D1 amplification was seen in 29%, 58% and 38% of luminal A, luminal B and HER2 subtypes, respectively¹⁴⁹. Meanwhile, CDK4 amplification was seen in 14%, 25% and 24% of luminal A, luminal B and HER2 subtypes, respectively¹⁴⁹. However, reduced expression/mutation/loss of Rb protein is more common in the basal subtype where changes in cyclin D and CDK4 were less common¹⁴⁹.

Upstream signalling is also capable of driving increased cyclin D:CDK4/6 activity. Oestrogen activation of ER upregulates cyclin D1 levels culminating in increased activation of CDK4/6 and cell proliferation¹⁴⁹. Mitogenic signalling pathways can also promote cyclin D1:CDK4/6 signalling.

Interestingly, cyclin D1 can independently activate ER suggesting a possible mechanism for ER+ cells becoming resistant to hormone therapy and the oestrogen-independence of some ER+ cancers¹⁴⁹.

CDK4/6 inhibition

For many years, it was believed that the cyclins and CDKs would be essential to cellular proliferation and normal embryonic development. However, this changed with the discovery that knockout mice lacking individual cyclins or CDKs were viable and double knockouts of CDK4/6 and triple knockouts of cyclin D1, D2 and D3 develop normally until mid-late gestation¹⁴⁹. Further, BC cells that become oestrogen-independent and resistant to hormone therapy are still reliant on cyclin D:CDK4/6 signalling to drive proliferation^149,152.

With these insights and the likelihood that CDK4/6 may be dispensable for some normal cell function, in addition to existing knowledge on inhibiting kinase activity, CDK4/6 became an increasingly interesting target for drug discovery¹⁴⁹.

Many of the initial attempts at CDK inhibition resulted in poor activity and an undesirable adverse event profile. These first generation of CDK inhibitors, including flavopiridol and UCN-01, were essentially pan-CDK inhibitors^149,151. Following the poor outcomes associated with pan-CDK inhibition, second generation CDK-specific inhibitors were developed.

While CDK2 was an initial target for many groups looking to develop inhibitors, the focus shifted to inhibition of CDK4/6. In recent years, the first CDK inhibitors targeting CDK4/6, have reached the market (Figure 13)¹⁵¹.

Figure 13. Chemical structures of CDK4/6 inhibitors used to treat HR+/HER2- metastatic breast cancer (Adapted^{105,118,153,154}).

Abemaciclib, palbociclib and ribociclib are all CDK4/6 inhibitors that differ structurally, abemaciclib more so than palbociclib and ribociclib.

Ribociclib and abemaciclib show stronger activity towards CDK4 than CDK6 in in vitro studies whereas palbociclib has similar activity against both kinases¹⁵⁵. Abemaciclib is the most potent of the three inhibitors and shows modest activity against a range of other CDKs, including CDK9¹⁵⁵. It seems unlikely that this has a significant in vivo effect, given the similar specificity for Rb+ tumour cells versus Rb− tumour cells seen with all the CDK4/6 inhibitors¹⁵⁶. However, some differences in adverse events between abemaciclib and palbociclib or ribociclib, utility of abemaciclib as a monotherapy, and the possible future utility of abemaciclib following progression on palbociclib may be explained by this difference in specificity^155,157,158. For instance, neutropenia is a similarly common adverse event for both palbociclib and ribociclib; however, it is less common in patients who received abemaciclib. This has ultimately led to differences in dosing between the three CDK4/6 inhibitors, with palbociclib and ribociclib dosed once daily in cycles of three weeks ON and 1 week OFF, whereas abemaciclib can be dosed continuously, twice daily^{104,105,117,118,154}.

Another difference that is worth noting is central nervous system bioavailability, with a higher efflux noted for palbociclib compared with abemaciclib and ribociclib, indicating that the latter two may be better suited to patients with central nervous system involvement¹⁵⁶.

Find out more on the differences between the available CDK4/6 inhibitors, and how this could impact on selecting the best CDK4/6 inhibitor for your patient

CDK4/6 inhibitors mode of action

The role of CDK4/6 in cell proliferation has been extensively studied and it was assumed that inhibition of CDK4/6 would simply prevent cell cycle progression. However, the mechanism of action of CDK4/6 inhibitors may be more complicated than this, involving additional non-cell cycle related effects, including quiescence and senescence, and alterations to metabolism and the immune system¹⁵⁶.

Cell cycle inhibition

The classical mechanism of action of CDK4/6 inhibitors is through the inhibition of the cell cycle. Inhibition of CDK4/6 prevents phosphorylation and inactivation of the Rb protein resulting in continued inhibition of the E2F transcription factor^149,156. This prevents the transcription of cyclin E and cyclin A which normally promote G1-S progression.

Quiescence versus senescence

The response of tumour cells to CDK4/6 inhibition does not seem to be uniform. Depending on the cell type and transforming event, some Rb-positive cells undergo quiescence while others become senescent. The advantage of creating senescent cells is that they will not return to the cell cycle following removal of the inhibitory signal and are generally resistant to proliferative signals¹⁵⁶. Achieving senescence requires cells to transition from quiescence and is termed geroconversion or senescence after growth arrest (SAGA).

CDK4 inhibitor-induced SAGA in breast cancer cells involves the downregulation of mouse double minute 2 homolog 2 (MDM2), redistribution of the chromatin remodeller ATRX and repression of HRAS transcription^159,160.

Metabolism and immune system

In vitro experiments have shown CDK4/6 inhibitors, often in combination with other therapies, are capable of manipulating metabolic pathways and driving cells towards apoptosis or senescence. However, these changes are likely to be context-dependent and more research is required to fully understand this mechanism¹⁵⁶.

Finally, CDK4/6 inhibition can affect maturation of sentinel cells of the immune system and expansion of regulatory T-cells¹⁵⁶. This increase in immune activity is thought to be at least partly related to reduced E2F transcriptional activity¹⁶¹.

In this clip, we asked Dr Laura Spring from the Massachusetts General Hospital Cancer Centre and Harvard Medical School, to tell us when she uses CDK4/6 inhibitors, and for which breast cancer patient groups.

CDK4/6 inhibitor safety and tolerability

CDK4/6 inhibitors are well tolerated, with any side-effects and adverse events resolving with dose modification and supportive care (table 7)¹⁶². Haematological toxicities, primarily neutropenia, are the most common side-effect observed with all three inhibitors¹⁶². Cytopenias occurring due to CDK4/6 inhibitor treatment are considered an on-target effect due to the role CDK6 plays in the proliferation of haematological precursors¹⁶².

Table 7. Adverse event frequency of CDK4/6 inhibitors in phase 3 trials (%) (Adapted¹⁶²)

A frequently reported serious adverse reaction to CDK4/6 inhibitor therapy is venous thromboembolism (VTE)¹⁶³. The incidence of VTE reported in MONARCH 2 and 3 (4.8% and 6.1%), MONALESSA-3 (4.8%), and PALOMA-1, 2 and 3 (1.6%) has led to warnings about increased VTE risk^{140,147,164,165}. However, as many of these clinical trials excluded patients with pre-existing cardiovascular conditions, the results regarding VTE risk were not an accurate depiction of overall VTE risk among the general real-world population.

This is demonstrated in results from a real-world clinical practice study assessing the rate of VTE following abemaciclib, which differs significantly from the rates seen among the highly selected clinical trial populations. In a multicentre observational cohort study of 364 patients with mBC, 7.4% developed thromboembolic complications over a median treatment duration of 5.5 months, with the median time to thrombosis only 2.8 months after initiating abemaciclib therapy¹⁶⁶.

Additionally, it was reported that development of VTE while on abemaciclib therapy was associated with a twofold increase in mortality risk and a much lower overall survival¹⁶⁶.

These findings are of particular importance because abemaciclib is the only approved CDK 4/6 inhibitor for use in early breast cancer¹⁶⁶.

Similarly, in real-world populations of people treated with palbociclib, the VTE incidence is 6–10%, significantly higher than the 1.6% reported in the clinical trials^163,167.

Given the considerable rates of VTE observed in both the clinical trial and the real-world settings, low-risk thromboprophylaxis could be an important consideration for certain patient subgroups receiving abemaciclib¹⁶⁶. However, further investigation to quantify the risk and benefits of such an intervention is still required.

Review

• Cyclin-dependent kinases (CDKs) comprise a family of serine/threonine protein kinases that interact with specific cyclins to promote cell cycle progression¹⁴⁸
• HR+/HER2− BC cells often overexpress cyclin D, which activates cell cycle progression through cyclin-dependent kinases 4 and 6 (CDK4/6)
• HR+/HER2− BC cells are sensitive to CDK4/6 inhibitors, such as oral abemaciclib, palbociclib and ribociclib¹³⁸
• CDK4/6 inhibitors are well tolerated, with any side-effects and adverse events resolving with dose modification and supportive care¹⁶²
• Haematological toxicities, primarily neutropenia, are the most common side-effect observed with all three inhibitors¹⁶²
• The most frequently reported serious adverse event of CDK4/6 inhibitor therapy is VTE¹⁶³

Learn more about the evidence for the use of CDK4/6 inhibitors for the treatment of HR+/HER2− breast cancer