Molecular subtyping of breast cancer: ready to use?

Learning vector quantized representation for cancer subtypes identification

BACKGROUND AND OBJECTIVE

Defining and separating cancer subtypes is essential for facilitating personalized therapy modality and prognosis of patients. The definition of subtypes has been constantly recalibrated as a result of our deepened understanding. During this recalibration, researchers often rely on clustering of cancer data to provide an intuitive visual reference that could reveal the intrinsic characteristics of subtypes. The data being clustered are often omics data such as transcriptomics that have strong correlations to the underlying biological mechanism. However, while existing studies have shown promising results, they suffer from issues associated with omics data: sample scarcity and high dimensionality while they impose unrealistic assumptions to extract useful features from the data while avoiding overfitting to spurious correlations.

METHODS

This paper proposes to leverage a recent strong generative model, Vector-Quantized Variational AutoEncoder, to tackle the data issues and extract discrete representations that are crucial to the quality of subsequent clustering by retaining only information relevant to reconstructing the input.

RESULTS

Extensive experiments and medical analysis on multiple datasets comprising 10 distinct cancers demonstrate the proposed clustering results can significantly and robustly improve prognosis over prevalent subtyping systems.

CONCLUSION

Our proposal does not impose strict assumptions on data distribution; while, its latent features are better representations of the transcriptomic data in different cancer subtypes, capable of yielding superior clustering performance with any mainstream clustering method.

Prognosis of breast cancer molecular subtypes in routine clinical care: A large prospective cohort study

BACKGROUND

In Germany, most breast cancer patients are treated in specialized breast cancer units (BCU), which are certified, and routinely monitored. Herein, we evaluate up-to-date oncological outcome of breast cancer (BC) molecular subtypes in routine clinical care of a specialized BCU.

METHODS

The study was a prospectively single-center cohort study of 4102 female cases with primary, unilateral, non-metastatic breast cancer treated between 01 January 2003 and 31 December 2012. The five routinely used molecular subtypes (Luminal A-like, Luminal B/HER2 negative-like, Luminal B/HER2 positive-like, HER2-type, Triple negative) were analyzed. The median follow-up time of the whole cohort was 55 months. We calculated estimates for local control rate (LCR), disease-free survival (DFS), distant disease-free survival (DDFS), overall survival (OS), and relative overall survival (ROS).

RESULTS

Luminal A-like tumors were the most frequent (44.7 %) and showed the best outcome with LCR of 99.1 % (95 % CI 98.5; 99.7), OS of 95.1 % (95 % CI 93.7; 96.5), and ROS of 100.0 % (95 % CI 98.5; 101.5). Triple negative tumors (12.3 %) presented the poorest outcome with LCR of 89.6 % (95 % CI 85.8; 93.4), OS of 78.5 % (95 % CI 73.8; 83.3), and ROS of 80.1 % (95 % CI 73.8; 83.2).

CONCLUSIONS

Patients with a favorable subtype can expect an OS above 95 % and an LCR of almost 100 % over 5 years. On the other hand the outcome of patients with HER2 and Triple negative subtypes remains poor, thus necessitating more intensified research and care.

Prognosis of breast cancer molecular subtypes in routine clinical care: A large prospective cohort study

Background

In Germany, most breast cancer patients are treated in specialized breast cancer units (BCU), which are certified, and routinely monitored. Herein, we evaluate up-to-date oncological outcome of breast cancer (BC) molecular subtypes in routine clinical care of a specialized BCU.

Methods

The study was a prospectively single-center cohort study of 4102 female cases with primary, unilateral, non-metastatic breast cancer treated between 01 January 2003 and 31 December 2012. The five routinely used molecular subtypes (Luminal A-like, Luminal B/HER2 negative-like, Luminal B/HER2 positive-like, HER2-type, Triple negative) were analyzed. The median follow-up time of the whole cohort was 55 months. We calculated estimates for local control rate (LCR), disease-free survival (DFS), distant disease-free survival (DDFS), overall survival (OS), and relative overall survival (ROS).

Results

Luminal A-like tumors were the most frequent (44.7 %) and showed the best outcome with LCR of 99.1 % (95 % CI 98.5; 99.7), OS of 95.1 % (95 % CI 93.7; 96.5), and ROS of 100.0 % (95 % CI 98.5; 101.5). Triple negative tumors (12.3 %) presented the poorest outcome with LCR of 89.6 % (95 % CI 85.8; 93.4), OS of 78.5 % (95 % CI 73.8; 83.3), and ROS of 80.1 % (95 % CI 73.8; 83.2).

Conclusions

Patients with a favorable subtype can expect an OS above 95 % and an LCR of almost 100 % over 5 years. On the other hand the outcome of patients with HER2 and Triple negative subtypes remains poor, thus necessitating more intensified research and care.

Electronic supplementary material

The online version of this article (doi:10.1186/s12885-016-2766-3) contains supplementary material, which is available to authorized users.

Molecular classification of breast carcinomas with particular emphasis on "basal-like" carcinoma: a critical review

During the last 11 years, 5 molecular subtypes of breast carcinoma (luminal A, luminal B, Her2-positive, basal-like, and normal breast-like) have been characterized and intensively studied. As genomic research evolves, further subtypes of breast cancers into new "molecular entities" are expected to occur. For example, a new and rare breast cancer subtype, known as claudin-low, has been recently found in human carcinomas and in breast cancer cell lines. There is no doubt that global gene expression analyses using high-throughput biotechnologies have drastically improved our understanding of breast cancer as a heterogeneous disease. The main question is, however, whether new molecular techniques such as gene expression profiling (or signature) should be regarded as the gold standard for identifying breast cancer subtypes. A critical review of the literature clearly shows major problems with current molecular techniques and classification including poor definitions, lack of reproducibility, and lack of quality control. Therefore, the current molecular approaches cannot be incorporated into routine clinical practice and treatment decision making as they are immature or even can be misleading. This review particularly focuses on the "basal-like" breast cancer subtype that represents one of the most popular breast cancer "entities". It critically shows major problems and misconceptions with and about this subtype and challenges the common claim that it represents a "distinct entity". It concludes that the term "basal-like" is misleading and states that there is no evidence that expression of basal-type cytokeratins in a given breast cancer, regardless of other established prognostic factors, does have any impact on clinical outcome.

Characteristics and outcomes according to molecular subtypes of breast cancer as classified by a panel of four biomarkers using immunohistochemistry

To investigate the significance of immunohistochemical molecular subtyping, we evaluated outcomes of subtypes based on estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), and Ki-67. Using tissue microarrays, 1006 breast cancer patients between November 1999 and August 2005 were categorized into four subtypes: luminal A (ER+ and/or PR+, HER2-, Ki-67 < 14%), luminal B (ER+ and/or PR+, HER2-, Ki-67 ≥ 14% or ER+ and/or PR+, HER2+), HER2-enriched (ER-, PR-, HER2+), and triple-negative breast cancer (TNBC) (ER-, PR-, HER2-). Demographics, recurrence patterns, and survival were retrospectively analyzed using uni-/multivariate analyses. Luminal A, luminal B, HER2-enriched, and TNBC accounted for 53.1%, 21.7%, 9.0%, and 16.2% of cases, respectively. Luminal A presented well-differentiation and more co-expression of hormone receptors comparing to luminal B. HER2-enriched showed larger size and higher nodal metastasis. TNBC demonstrated younger age at diagnosis, larger size, undifferentiation, higher proliferation, and frequent visceral metastases. The peak of recurrence for luminal A was at 36 months postoperatively, while that for HER2-enriched and TNBC peaked at 12 months. The relapse risk of luminal B was mixed. Luminal A showed the best survival, but no difference was observed between the other three subtypes. When matched by nodal status, however, TNBC showed the worst outcomes in node-positive patients. In multivariate analyses, luminal A remained a positive prognostic significance. Immunohistochemically-defined subtypes showed different features, recurrence patterns, and survival. Therefore, molecular subtypes using four biomarkers could provide clinically useful information of tumor biology and clinical behaviors, and could be used for determining treatment and surveillance strategies.

Prediction of outcome of early ER+ breast cancer is improved using a biomarker panel, which includes Ki-67 and p53

BACKGROUND

The aim of this study is to determine whether immunohistochemical (IHC) assessment of Ki67 and p53 improves prognostication of oestrogen receptor-positive (ER+) breast cancer after breast-conserving therapy (BCT). In all, 498 patients with invasive breast cancer from a randomised trial of BCT with or without tumour bed radiation boost were assessed using IHC.

METHODS

The ER+ tumours were classified as 'luminal A' (LA): ER+ and/or PR+, Ki-67 low, p53-, HER2- or 'luminal B' (LB): ER+ and/or PR+and/or Ki-67 high and/or p53+ and/or HER2+. Kaplan-Meier and Cox proportional hazards methodology were used to ascertain relationships to ispilateral breast tumour recurrence (IBTR), locoregional recurrence (LRR), distant metastasis-free survival (DMFS) and breast cancer-specific survival (BCSS).

RESULTS

In all, 73 patients previously LA were re-classified as LB: a greater than four-fold increase (4.6-19.3%) compared with ER, PR, HER2 alone. In multivariate analysis, the LB signature independently predicted LRR (hazard ratio (HR) 3.612, 95% CI 1.555-8.340, P=0.003), DMFS (HR 3.023, 95% CI 1.501-6.087, P=0.002) and BCSS (HR 3.617, 95% CI 1.629-8.031, P=0.002) but not IBTR.

CONCLUSION

The prognostic evaluation of ER+ breast cancer is improved using a marker panel, which includes Ki-67 and p53. This may help better define a group of poor prognosis ER+ patients with a greater probability of failure with endocrine therapy.

Molecular pathology of breast cancer: the journey from traditional practice toward embracing the complexity of a molecular classification

CONTEXT

Adenocarcinoma of the breast is the most frequent cancer affecting women in both developed and developing regions of the world. From the moment of clinical presentation until the time of pathologic diagnosis, patients affected by this disease will face daunting questions related to prognosis and treatment options. While improvements in targeted therapies have led to increased patient survival, these same advances have created the imperative to accurately stratify patients to achieve maximum therapeutic efficacy while minimizing side effects. In this evolving era of personalized medicine, there is an ever-increasing need to overcome the limitations of traditional diagnostic practice.

OBJECTIVE

To summarize the molecular diagnostics traditionally used to guide prognostication and treatment of breast carcinomas, to highlight published data on the molecular classification of these tumors, and to showcase molecular assays that will supplement traditional methods of categorizing the disease.

DATA SOURCES

A review of the literature covering the molecular diagnostics of breast carcinomas with a focus on the gene expression and array studies used to characterize the molecular signatures of the disease. Special emphasis is placed on summarizing evolving technologies useful in the diagnosis and characterization of breast carcinoma.

CONCLUSIONS

Available and emerging molecular resources will allow pathologists to provide superior diagnostic, prognostic, and predictive information about individual breast carcinomas. These advances should translate into earlier identification and tailored therapy and should ultimately improve outcome for patients affected by this disease.

Metastatic breast cancer cells in the bone marrow microenvironment: novel insights into oncoprotection

Among all cancers, malignancies of the breast are the second leading cause of cancer death in the United States after carcinoma of the lung. One of the major factors considered when assessing the prognosis of breast cancer patients is whether the tumor has metastasized to distant organs. Although the exact phenotype of the malignant cells responsible for metastasis and dormancy is still unknown, growing evidence has revealed that they may have stem cell-like properties that may account for resistance to chemotherapy and radiation. One process that has been attributed to primary tumor metastasis is the epithelial-to-mesenchymal transition. In this review, we specifically discuss breast cancer dissemination to the bone marrow and factors that ultimately serve to shelter and promote tumor growth, including the complex relationship between mesenchymal stem cells (MSCs) and various aspects of the immune system, carcinoma-associated fibroblasts, and the diverse components of the tumor microenvironment. A better understanding of the journey from the primary tumor site to the bone marrow and subsequently the oncoprotective role of MSCs and other factors within that microenvironment can potentially lead to development of novel therapeutic targets.