Breast cancer classification

Breast cancers can be classified by different schemata. Each of these aspects influences treatment response and prognosis. Description of a breast cancer would optimally include all of these classification aspects, as well as other findings, such as signs found on physical exam. A full classification includes histopathological type, grade, stage (TNM), receptor status, and the presence or absence of genes as determined by DNA testing:

• Histopathology. Although breast cancer has many different histologies, the considerable majority of breast cancers are derived from the epithelium lining the ducts or lobules, and are classified as mammary ductal carcinoma. Carcinoma in situ is proliferation of cancer cells within the epithelial tissue without invasion of the surrounding tissue. In contrast, invasive carcinoma invades the surrounding tissue.[3] Perineural and/or lymphovascular space invasion is usually considered as part of the histological description of a breast cancer, and when present may be associated with more aggressive disease.
• Grade. Grading focuses on the appearance of the breast cancer cells compared to the appearance of normal breast tissue. Normal cells in an organ like the breast become differentiated, meaning that they take on specific shapes and forms that reflect their function as part of that organ. Cancerous cells lose that differentiation. In cancer, the cells that would normally line up in an orderly way to make up the milk ducts become disorganized. Cell division becomes uncontrolled. Cell nuclei become less uniform. Pathologists describe cells as well differentiated (low-grade), moderately differentiated (intermediate-grade), and poorly differentiated (high-grade) as the cells progressively lose the features seen in normal breast cells. Poorly differentiated cancers have a worse prognosis.
• Stage. The TNM classification for staging breast cancer is based on the size of the cancer where it originally started in the body and the locations to which it has travelled. These cancer characteristics are described as the size of the tumor (T), whether or not the tumor has spread to the lymph nodes (N) in the armpits, neck, and inside the chest, and whether the tumor has metastasized (M) (i.e. spread to a more distant part of the body). Larger size, nodal spread, and metastasis have a larger stage number and a worse prognosis. The main stages are:
  • Stage 0 which is in situ disease or Paget's disease of the nipple. Stage 0 is a pre-cancerous or marker condition, either ductal carcinoma in situ (DCIS) or lobular carcinoma in situ (LCIS).
  • Stages 1–3 are within the breast or regional lymph nodes.
  • Stage 4 is a metastatic cancer. Metastatic breast cancer has a less favorable prognosis.
• Receptor status. Cells have receptors on their surface and in their cytoplasm and nucleus. Chemical messengers such as hormones bind to receptors, and this causes changes in the cell. Breast cancer cells may or may not have many different types of receptors, the three most important in the present classification being: estrogen receptor (ER), progesterone receptor (PR), and HER2/neu. Cells with or without these receptors are called ER positive (ER+), ER negative (ER-), PR positive (PR+), PR negative (PR-), HER2 positive (HER2+), and HER2 negative (HER2-). Cells with none of these receptors are called basal-like or triple negative.
• DNA-based classification. Understanding the specific details of a particular breast cancer may include looking at the cancer cell DNA or RNA by several different laboratory approaches. When specific DNA mutations or gene expression profiles are identified in the cancer cells this may guide the selection of treatments, either by targeting these changes, or by predicting from these alterations which non-targeted therapies are most effective.
• Other classification approaches.
  • Computer models such as Adjuvant! can combine the various classification aspects according to validated algorithms and present visually appealing graphics that assist in treatment decisions.
  • The USC/Van Nuys prognostic index (VNPI) classifies ductal carcinoma in situ (DCIS) into dissimilar risk categories that may be treated accordingly.
  • The choice of which treatment to receive can be substantially influenced by comorbidity assessments.
  • Familial breast cancers may potentially undergo dissimilar treatment (such as mastectomy).

The 2012 World Health Organization (WHO) classification of tumors of the breast[10] which includes benign (generally harmless) tumors and malignant (cancerous) tumors, recommends the following pathological types:

Tubule formation score in the Nottingham system

This parameter assesses what percent of the tumor forms normal duct structures. In cancer, there is a breakdown of the mechanisms that cells use to attach to each other and communicate with each other, to form tissues such as ducts, so the tissue structures become less orderly.

Note: The overall appearance of the tumor has to be considered.

• 1 point: tubular formation in more than 75% of the tumor
• 2 points: tubular formation in 10 to 75% of the tumor
• 3 points: tubular formation in less than 10% of the tumor

This parameter assesses whether the cell nuclei are uniform like those in normal breast duct epithelial cells, or whether they are larger, darker, or irregular (pleomorphic). In cancer, the mechanisms that control genes and chromosomes in the nucleus break down, and irregular nuclei and pleomorphic changes are signs of abnormal cell reproduction.

Note: The cancer areas having cells with the greatest cellular abnormalities should be evaluated.

• 1 point: nuclei with minimal or mild variation in size and shape
• 2 points: nuclei with moderate variation in size and shape
• 3 points: nuclei with marked variation in size and shape

• 
  Ductal carcinoma with mild nuclear pleomorphism.
• 
  Invasive ductal carcinoma with moderate nuclear pleomorphism.
• 
  Invasive ductal carcinoma with marked nuclear pleomorphism.

Mitosis appearances in breast cancer

This parameter assesses how many mitotic figures (dividing cells) the pathologist sees in 10x high power microscope field. One of the hallmarks of cancer is that cells divide uncontrollably. The more cells that are dividing, the worse the cancer.

Note: Mitotic figures are counted only at the periphery of the tumor, and counting should begin in the most mitotically active areas.

The scores for each of these three criteria are added together to give a final overall score and a corresponding grade as follows:

• 3-5 Grade 1 tumor (well-differentiated). Best prognosis.
• 6-7 Grade 2 tumor (moderately differentiated). Medium prognosis.
• 8-9 Grade 3 tumor (poorly differentiated). Worst prognosis.

Lower-grade tumors, with a more favorable prognosis, can be treated less aggressively, and have a better survival rate. Higher-grade tumors are treated more aggressively, and their intrinsically worse survival rate may warrant the adverse effects of more aggressive medications.

Tumor – The tumor values (TX, T0, Tis, T1, T2, T3 or T4) depend on the cancer at the primary site of origin in the breast, as follows:[22]

• TX: inability to assess that site
• Tis: ductal carcinoma in situ (DCIS), lobular carcinoma in situ (LCIS) or Paget's disease
• T1: Less than 2 cm

• T1a: 0.1 to 0.5 cm
• T1b: 0.5 to 1.0 cm
• T1c: 1.0 to 2.0 cm

• T2: 2 to 5 cm
• T3: Larger than 5 cm
• T4

• T4a: Chest wall involvement
• T4b: Skin involvement
• T4c: Both 4a and 4b
• T4d: Inflammatory breast cancer, a clinical circumstance where typical skin changes involve at least a third of the breast.

Lymph Node – The lymph node values (NX, N0, N1, N2 or N3) depend on the number, size and location of breast cancer cell deposits in various regional lymph nodes, such as the armpit (axillary lymph nodes), the collar area (supraclavicular lymph nodes), and inside the chest (internal mammary lymph nodes.)[23][24] The armpit is designated as having three levels: level I is the low axilla, and is below or outside the lower edge of the pectoralis minor muscle; level II is the mid-axilla which is defined by the borders of the pectoralis minor muscle; and level III, or high (apical) axilla which is above the pectoralis minor muscle. Each stage is as follows:[22]

• N0: There is some nuance to the official definitions for N0 disease, which includes:

• N0(i+) : Isolated Tumor Cell clusters (ITC), which are small clusters of cells not greater than 0.2 mm, or single tumor cells, or a cluster of fewer than 200 cells in a single histologic cross-section, whether detected by routine histology or immunohistochemistry.[25]
• N0(mol-): regional lymph nodes have no metastases histologically, but have positive molecular findings (RT-PCR).[25]

• N1: Metastases in 1-3 axillary lymph nodes and/or in internal mammary nodes; and/or in clinically negative internal mammary nodes with micrometastasis, or macrometastasis on sentinel lymph node biopsy.[25]

• N1mi: Micrometastasis, that is, lymph node clusters at least 2 mm or 200 cells, but less than 2.0 mm.[25] At least one carcinoma focus over 2.0 mm is called "Lymph node metastasis". If one node qualifies as metastasis, all other nodes even with smaller foci are counted as metastases as well.

• N2: Fixed/matted ipsilateral axillary nodes.
• N3

• N3a – Ipsilateral infraclavicular nodes
• N3b – Ipsilateral internal mammary nodes
• N3c – Ipsilateral supraclavicular nodes

• M0: No clinical or radiographic evidence of distant metastases
• M0(i+): Molecularly or microscopically detected tumor cells in circulating blood, bone marrow or non-regional nodal tissue, no larger than 0.2 mm, and without clinical or radiographic evidence or symptoms or signs of metastases, and which, perhaps counter-intuitively, does not change the stage grouping, as staging for in M0(i+) is done according to the T and N values
• M1: Distant detectable metastases as determined by classic clinical and radiographic means, and/or metastasis that are histologically larger than 0.2 mm.

A combination of T, N and M, as follows:[22]

• Stage 0: Tis
• Stage I: T1N0
• Stage II: T2N0, T3N0 T0N1, T1N1, or T2N1
• Stage III: Invasion into skin and/or ribs, matted lymph nodes, T3N1, T0N2, T1N2, T2N2, T3N2, AnyT N3, T4 any N, locally advanced breast cancer
• Stage IV: M1, advanced breast cancer

The impact of different stages on outcome can be appreciated in the following table, published in a 2007 textbook,[26] which shows the observed 5-year overall survival of over 50,000 patients from 1989 who were reclassified using the AJCC 5th edition criteria; the data is also available in the AJCC source,[27] which also gives the relative survival rate in comparison to an age-matched (actually, age- sex- and race-matched) population. This data is historical, does not show the influence of important additional factors such as estrogen receptor (ER) or HER2/neu receptor status, and does not reflect the impact of newer treatments.

Although TNM classification is an internationally agreed system, it has gradually evolved through its different editions; the dates of publication and of adoption for use of AJCC editions is summarized in the table in this article; past editions are available from AJCC for web download.[28]

Several factors are important when reviewing reports for individual breast cancers or when reading the medical literature, and applying staging data.

It is crucial to be aware that the TNM system criteria have varied over time, sometimes fairly substantially, according to the different editions that AJCC and UICC have released.[28] Readers are assisted by the provision in the table of direct links to the breast cancer chapters of these various editions.

As a result, a given stage may have quite a different prognosis depending on which staging edition is used, independent of any changes in diagnostic methods or treatments, an effect that can contribute to "stage migration".[35] For example, differences in the 1998 and 2003 categories resulted in many cancers being assigned differently, with apparent improvement in survival rates.[36]

As a practical matter, reports often use the staging edition that was in place when the study began, rather than the date of acceptance or publication. However, it is worth checking whether the author updated the staging system during the study, or modified the usual classification rules for specific use in the investigation.

A different effect on staging arises from evolving technologies that are used to assign patients to particular categories, such that increasingly sensitive methods tend to cause individual cancers to be reassigned to higher stages, making it improper to compare that cancer's prognosis to the historical expectations for that stage.

Finally, of course, a further important consideration is the effect of improving treatments over time as well.

Previous editions featured three metastatic values (MX, M0 and M1) which referred respectively to absence of adequate information, the confirmed absence, or the presence of breast cancer cells in locations other than the breast and regional lymph nodes, such as to bone, brain, lung.

AJCC has provided web accessible poster versions of the current versions of these copyrighted TNM descriptors and groups,[29] and readers should refer to that up to date, accurate information[29] or to the National Cancer Institute (NCI)[21] or National Comprehensive Cancer Network[37] sites which reprints these with AJCC permission.

For accurate, complete, current details refer to the accessible copyrighted documentation from AJCC,[29] or to the authorized documentation from NCI[21] or NCCN;[37] for past editions refer to AJCC.[28]

Receptor status was traditionally considered by reviewing each individual receptor (ER, PR, her2) in turn, but newer approaches look at these together, along with the tumor grade, to categorize breast cancer into several conceptual molecular classes[44] that have different prognoses[37] and may have different responses to specific therapies.[45] DNA microarrays have assisted this approach, as discussed in the following section. Proposed molecular subtypes include:

• Basal-like: ER-, PR- and HER2-; also called triple negative breast cancer (TNBC).[46] Most BRCA1 breast cancers are basal-like TNBC.
• Luminal A: ER+ and low grade
• Luminal B: ER+ but often high grade
• Luminal ER-/AR+: (overlapping with apocrine and so called molecular apocrine) - recently identified androgen responsive subtype which may respond to antihormonal treatment with bicalutamide[42]
• ERBB2/HER2-amplified: has overexpressed HER2/neu[38]
• Normal breast-like[37][44][47]
• Claudin-low: a more recently described class; often triple-negative, but distinct in that there is low expression of cell-cell junction proteins[46] including E-cadherin[46] and frequently there is infiltration with lymphocytes.[44][48][49]

Traditional DNA classification was based on the general observation that cells that are dividing more quickly have a worse prognosis, and relied on either the presence of protein Ki67 or the percentage of cancer cell DNA in S phase. These methods, and scoring systems that used DNA ploidy, are used much less often now, as their predictive and prognostic power was less substantial than other classification schemes such as the TNM stage. In contrast, modern DNA analyses are increasingly relevant in defining underlying cancer biology and in helping choose treatments.[51][52][53][54]

HER2/neu status can be analyzed by fluorescent in-situ hybridization (FISH) assays.[37] Some commentators prefer this approach, claiming a higher correlation than receptor immunohistochemistry with response to trastuzumab, a targeted therapy, but guidelines permit either testing method.[37]

Molecular classification of breast cancer from mRNA expression profiles

DNA microarrays have compared normal cells to breast cancer cells and found differences in the expression of hundreds of genes. Although the significance of many of those genetic differences is unknown, independent analyses by different research groups has found that certain groups of genes have a tendency to co-express. These co-expressing clusters have included hormone receptor-related genes, HER2-related genes, a group of basal-like genes, and proliferation genes. As might therefore be anticipated, there is considerable similarity between the receptor and microarray classifications, but assignment of individual tumors is by no means identical. By way of illustration, some analyses have suggested that approximately 75% of receptor classified triple-negative breast cancers (TNBC) basal-like tumors have the expected DNA expression profile, and a similar 75% of tumors with a typical basal-like DNA expression profile are receptor TNBC as well. To say this in a different way to emphasize things, this means that 25% of triple-negative breast cancer (TNBC) basal-like tumors as defined by one or other classification are excluded from the alternative classification's results. Which classification scheme (receptor IHC or DNA expression profile) more reliably assorts particular cancers to effective therapies is under investigation.

Several commercially marketed DNA microarray tests analyze clusters of genes and may help decide which possible treatment is most effective for a particular cancer.[55] The use of these assays in breast cancers is supported by Level II evidence or Level III evidence. No tests have been verified by Level I evidence, which is rigorously defined as being derived from a prospective, randomized controlled trial where patients who used the test had a better outcome than those who did not. Acquiring extensive Level I evidence would be clinically and ethically challenging. However, several validation approaches[56][57] are being actively pursued.

Numerous genetic profiles have been developed.[58][59] The most heavily marketed are:

• Oncotype DX is supported by Level II evidence, and was originally designed for use in estrogen receptor (ER) positive tumors,[60] and has been endorsed by the American Society of Clinical Oncology (ASCO)[58][61] and the NCCN.[37]
• MammaPrint is supported only by Level III evidence, can be performed on estrogen receptor (ER) positive and negative tumors, and has FDA approval.
• Two other tests also only have Level III evidence: Theros and MapQuant Dx.[62][63][64]

These multigene assays, some partially and some completely commercialized, have been scientifically reviewed to compare them with other standard breast cancer classification methods such as grade and receptor status.[47][59] Although these gene-expression profiles look at different individual genes, they seem to classify a given tumor into similar risk groups and thus provide concordant predictions of outcome.[37][60]

Although there is considerable evidence that these tests can refine the treatment decisions in a meaningful proportion of breast cancers[58][59] they are fairly expensive; proposed selection criteria for which particular tumors may benefit by being interrogated by these assays[37] remain controversial, particularly with lymph node positive cancers.[37] One review characterized these genetic tests collectively as adding "modest prognostic information for patients with HER2-positive and triple-negative tumors, but when measures of clinical risk are equivocal (e.g., intermediate expression of ER and intermediate histologic grade), these assays could guide clinical decisions".[38]

Oncotype DX assesses 16 cancer-related genes and 5 normal comparator reference genes, and is therefore sometimes known as the 21-gene assay. It was designed for use in estrogen receptor (ER) positive tumors. The test is run on formalin fixed, paraffin-embedded tissue. Oncotype results are reported as a Recurrence Score (RS), where a higher RS is associated with a worse prognosis, referring to the likelihood of recurrence without treatment. In addition to that prognostic role, a higher RS is also associated with a higher probability of response to chemotherapy, which is termed a positive predictive factor.

These results suggest that not only does Oncotype stratify estrogen-receptor positive breast cancer into different prognostic groups, but also suggest that cancers that have a particularly favorable Oncotype DX microarray result tend to derive minimal benefit from adjuvant chemotherapy and so it may be appropriate to choose to avoid side effects from that additional treatment. As an additional example, a neoadjuvant clinical treatment program that included initial chemotherapy followed by surgery and subsequent additional chemotherapy, radiotherapy, and hormonal therapy found a strong correlation of the Oncotype classification with the likelihood of a complete response (CR) to the presurgical chemotherapy.[65]

Since high risk features may already be evident in many high risk cancers, for example hormone-receptor negativity or HER-2 positive disease, the Oncotype test may especially improve the risk assessment that is derived from routine clinical variables in intermediate risk disease.[66] Results from both the US[67] and internationally[68] suggest that Oncotype may assist in treatment decisions.[69]

Oncotype DX has been endorsed by the American Society of Clinical Oncology (ASCO)[58][61] and the NCCN.[37] The NCCN Panel considers the 21-gene assay as an option when evaluating certain tumors[37] to assist in estimating likelihood of recurrence and benefit from chemotherapy, emphasizing that the recurrence score should be used along with other breast cancer classification elements when stratifying risk.[37] Oncotype fulfilled all California Technology Assessment Forum (CTAF) criteria in October 2006.[70] The U.S. Food and Drug Administration (FDA) does not mandate approval of this class of tests if they are performed at a single, company-operated laboratory[71] Genomic Health, which developed Oncotype DX, offers the test under these so-called home brew rules and, accordingly, to that extent the Oncotype DX assay is not specifically FDA approved.[71]

The MammaPrint gene pattern is a commercial-stage 70-gene panel marketed by Agendia,[72] that was developed in patients under age 55 years who had lymph node negative breast cancers (N0).[70] The commercial test is marketed for use in breast cancer irrespective of estrogen receptor (ER) status.[70] The test is run on formalin fixed, paraffin-embedded tissue. MammaPrint traditionally used rapidly frozen tissue[37] but a room temperature, molecular fixative is available for use within 60 minutes of obtaining fresh tissue samples.[73] MammaPrint categorizes tumors as either high or low risk.

A summary of clinical trials using MammaPrint is included in the MammaPrint main article. The available evidence for Mammaprint was reviewed by California Technology Assessment Forum (CTAF) in June 2010; the written report indicated that MammaPrint had not yet fulfilled all CTAF criteria.[70] MammaPrint has 5 FDA clearances and is the only FDA cleared microarray assay available. To be eligible for the MammaPrint gene expression profile, a breast cancer should have the following characteristics: stage 1 or 2, tumor size less than 5.0 cm, estrogen receptor positive (ER+) or estrogen receptor negative (ER-). In the US, the tumor should also be lymph node negative (N0), but internationally the test may be performed if the lymph node status is negative or positive with up to 3 nodes.[74]

One method of assessing the molecular subtype of a breast cancer is by BluePrint,[75] a commercial-stage 80-gene panel marketed by Agendia, either as a standalone test, or combined with the MammaPrint gene profile.

The choice of established chemotherapy medications, if chemotherapy is needed, may also be affected by DNA assays that predict relative resistance or sensitivity. Topoisomerase II (TOP2A) expression predicts whether doxorubicin is relatively useful.[76][77] Expression of genes that regulate tubulin may help predict the activity of taxanes.

Various molecular pathway targets and DNA results are being incorporated in the design of clinical trials of new medicines.[78] Specific genes such as p53, NME1, BRCA and PIK3CA/Akt may be associated with responsiveness of the cancer cells to innovative research pharmaceuticals. BRCA1 and BRCA2 polymorphic variants can increase the risk of breast cancer, and these cancers tend to express a pr ofile of genes, such as p53, in a pattern that has been called "BRCA-ness." Cancers arising from BRCA1 and BRCA2 mutations, as well as other cancers that share a similar "BRCA-ness" profile, including some basal-like receptor triple negative breast cancers, may respond to treatment with PARP inhibitors[79] such as olaparib. Combining these newer medicines with older agents such as 6-Thioguanine (6TG) may overcome the resistance that can arise in BRCA cancers to PARP inhibitors or platinum-based chemotherapy.[80] mTOR inhibitors such as everolimus may show more effect in PIK3CA/Akt e9 mutants than in e20 mutants or wild types.[81]

DNA methylation patterns can epigenetically affect gene expression in breast cancer and may contribute to some of the observed differences between genetic subtypes.[82]

Tumors overexpressing the Wnt signaling pathway co-receptor low-density lipoprotein receptor-related protein 6 (LRP6) may represent a distinct subtype of breast cancer and a potential treatment target.[83]

Numerous clinical investigations looked at whether testing for variant genotype polymorphic alleles of several genes could predict whether or not to prescribe tamoxifen; this was based on possible differences in the rate of conversion of tamoxifen to the active metabolite, endoxifen. Although some studies had suggested a potential advantage from CYP2D6 testing, data from two large clinical trials found no benefit.[84][85] Testing for the CYP2C19*2 polymorphism gave counterintuitive results.[86] The medical utility of potential biomarkers of tamoxifen responsiveness such as HOXB13,[87] PAX2,[88] and estrogen receptor (ER) alpha and beta isoforms interaction with SRC3[89][90] have all yet to be fully defined.

Computer models consider several traditional factors concurrently to derive individual survival predictions and calculations of potential treatment benefits. The validated algorithms can present visually appealing graphics that assist in treatment decisions. In addition, other classifications of breast cancers do exist and no uniform system has been consistently adopted worldwide.

Adjuvant![91] is based on US cohorts[92] and presents colored bar charts that display information that may assist in decisions regarding systemic adjuvant therapies. Successful validation was seen with Canadian[93] and Dutch[94] cohorts. Adjuvant! seemed less applicable to a British cohort[95] and accordingly PREDICT is being developed in the United Kingdom.[96]

Among many immunohistochemical tests that may further stratify prognosis, BCL2 has shown promise in preliminary studies.[97]

The USC/Van Nuys prognostic index (VNPI) classifies ductal carcinoma in situ (DCIS) into dissimilar risk categories that may be treated accordingly.[98]

The choice of which treatment to receive can be substantially influenced by comorbidity assessments.

There is some evidence that breast cancers that arise in familial clusters, such as Hereditary breast—ovarian cancer syndrome, may have a dissimilar prognosis. Also potentially dissimilar treatment.