Prediction Of Breast Cancer Outcome Research Articles

Personalized circulating tumor DNA for minimal residual disease and dynamic assessment in patients undergoing neoadjuvant chemotherapy for breast cancer: Preliminary analysis from MSK-LINC.

Personalized circulating tumor DNA for minimal residual disease and dynamic assessment in patients undergoing neoadjuvant chemotherapy for breast cancer: Preliminary analysis from MSK-LINC.

Spatial Distribution and Densities of CD103+ and FoxP3+ Tumor Infiltrating Lymphocytes by Digital Analysis for Outcome Prediction in Breast Cancer.

The evaluation of tumor-infiltrating lymphocytes (TILs) for breast cancer prognosis is now established. However, the clinical value for their spatial distributions of specific immune subsets, namely CD103+ tissue-resident memory T cells FoxP3+ regulatory T ells, have not been thoroughly examined. Representative whole sections of breast cancers were subjected to CD103 and FoxP3 double staining. Their density, ratio, and spatial features were analyzed in tumor area and tumor-stromal interface. Their associations with clinicopathological parameters and patient's prognosis were analyzed. CD103 TILs were closer to tumor nests than FoxP3 TILs in the tumor-stromal interface. Their densities were associated with high-grade disease, TNBC, and stromal TILs. High stromal FoxP3 (sFoxP3) TILs and close proximity of sCD103 TILs to tumor were independently associated with better survival at multivariate analysis. Subgroup analysis showed the high FoxP3 TILs density associated better survival was seen in HER2-OE and TNBC subtypes while the proximity of CD103 TILs to tumor nests associated better survival was seen in luminal cancers. The prognostic impact of CD103 and FoxP3 TILs in breast cancer depends on their spatial localization. High sFoxP3 TIL density and the lower distance of CD103 TILs from the tumor nests had independent favorable prognostic values.

Outcome and biomarker supervised deep learning for survival prediction in two multicenter breast cancer series

BackgroundPrediction of clinical outcomes for individual cancer patients is an important step in the disease diagnosis and subsequently guides the treatment and patient counseling. In this work, we develop and evaluate a joint outcome and biomarker supervised (estrogen receptor expression and ERBB2 expression and gene amplification) multitask deep learning model for prediction of outcome in breast cancer patients in two nation-wide multicenter studies in Finland (the FinProg and FinHer studies). Our approach combines deep learning with expert knowledge to provide more accurate, robust, and integrated prediction of breast cancer outcomes. Materials and methodsUsing deep learning, we trained convolutional neural networks (CNNs) with digitized tissue microarray (TMA) samples of primary hematoxylin-eosin-stained breast cancer specimens from 693 patients in the FinProg series as input and breast cancer-specific survival as the endpoint. The trained algorithms were tested on 354 TMA patient samples in the same series. An independent set of whole-slide (WS) tumor samples from 674 patients in another multicenter study (FinHer) was used to validate and verify the generalization of the outcome prediction based on CNN models by Cox survival regression and concordance index (c-index). Visual cancer tissue characterization, i.e., number of mitoses, tubules, nuclear pleomorphism, tumor-infiltrating lymphocytes, and necrosis was performed on TMA samples in the FinProg test set by a pathologist and combined with deep learning-based outcome prediction in a multitask algorithm. ResultsThe multitask algorithm achieved a hazard ratio (HR) of 2.0 (95% confidence interval [CI] 1.30–3.00), P < 0.001, c-index of 0.59 on the 354 test set of FinProg patients, and an HR of 1.7 (95% CI 1.2–2.6), P = 0.003, c-index 0.57 on the WS tumor samples from 674 patients in the independent FinHer series. The multitask CNN remained a statistically independent predictor of survival in both test sets when adjusted for histological grade, tumor size, and axillary lymph node status in a multivariate Cox analyses. An improved accuracy (c-index 0.66) was achieved when deep learning was combined with the tissue characteristics assessed visually by a pathologist. ConclusionsA multitask deep learning algorithm supervised by both patient outcome and biomarker status learned features in basic tissue morphology predictive of survival in a nationwide, multicenter series of patients with breast cancer. The algorithms generalized to another independent multicenter patient series and whole-slide breast cancer samples and provide prognostic information complementary to that of a comprehensive series of established prognostic factors.

The implementation of a commercially available multi-gene profile test for breast cancer characterization in a department of pathology: what have we learned from the first 100 cases?

Analysis of breast cancer prognostic and predictive factors is still nowadays poorly accurate and standardized. The advent of multi-gene expression profiles (MGEPs) has improved the prediction of breast cancer outcome, particularly regarding early luminal breast cancers (LBCs). The availability in our Institute of EndoPredict® (EP), a last-generation prognostic gene signature assay, has prompted us to study a series of LBCs, firstly verifying its reproducibility on six routine representative cases, either presenting non-optimal preanalytical conditions or different tumor samples from the same patient; secondly, correlating EP results on 8 retrospectively recruited samples with patients' follow-up; thirdly, applying prospectively EP on 100 routinely diagnosed cases, assessing the oncologists' and pathologists' attitude toward it. The complete reproducibility of EP on all the samples investigated in the first phase allowed to state that EP overcomes the detrimental effects of an inaccurate pre-analytic phase, determining the most appropriate prognostic and predictive parameters of breast cancer. The second phase confirmed EP as a fundamental tool in guiding therapeutic decision, improving the classical bio-pathological characterization and recovering 38% patients' inadequately managed. Finally, the study disclosed how oncologists sometimes inadequately requested EP, but also how it allows a better stratification of breast cancer otherwise considered poorly aggressive and not requiring an EP test, such as G1 neoplasms or tubular histotype. In conclusion, the introduction of EP test in an Anatomic Pathology Department emerges as a useful tool in routine breast cancer diagnosis, both for the characterization of individual cases and, as a result, for more appropriate therapeutic choices.

A Pipeline for Integrated Theory and Data-Driven Modeling of Biomedical Data.

Genome sequencing technologies have the potential to transform clinical decision making and biomedical research by enabling high-throughput measurements of the genome at a granular level. However, to truly understand mechanisms of disease and predict the effects of medical interventions, high-throughput data must be integrated with demographic, phenotypic, environmental, and behavioral data from individuals. Further, effective knowledge discovery methods must infer relationships between these data types. We recently proposed a pipeline (CausalMGM) to achieve this. CausalMGM uses probabilistic graphical models to infer the relationships between variables in the data; however, CausalMGM's graphical structure learning algorithm can only handle small datasets efficiently. We propose a new methodology (piPref-Div) that selects the most informative variables for CausalMGM, enabling it to scale. We validate the efficacy of piPref-Div against other feature selection methods and demonstrate how the use of the full pipeline improves breast cancer outcome prediction and provides biologically interpretable views of gene expression data.

FUNMarker: Fusion Network-Based Method to Identify Prognostic and Heterogeneous Breast Cancer Biomarkers.

Breast cancer is a heterogeneous disease with many clinically distinguishable molecular subtypes each corresponding to a cluster of patients. Identification of prognostic and heterogeneous biomarkers for breast cancer is to detect cluster-specific gene biomarkers which can be used for accurate survival prediction of breast cancer outcomes. In this study, we proposed a FUsion Network-based method (FUNMarker) to identify prognostic and heterogeneous breast cancer biomarkers by considering the heterogeneity of patient samples and biological information from multiple sources. To reduce the affect of heterogeneity of patients, samples were first clustered using the K-means algorithm based on the principal components of gene expression. For each cluster, to comprehensively evaluate the influence of genes on breast cancer, genes were weighted from three aspects: biological function, prognostic ability and correlation with known disease genes. Then they were ranked via a label propagation model on a fusion network that combined physical protein interactions from seven types of networks and thus could reduce the impact of incompleteness of interactome. We compared FUNMarker with three state-of-the-art methods and the results showed that biomarkers identified by FUNMarker were biological interpretable and had stronger discriminative power than the existing methods in differentiating patients with different prognostic outcomes.

Transcriptional Signatures Derived from Murine Tumor-Associated Macrophages Predict Outcome in Breast Cancer Patients

Tumor-associated macrophages (TAMs) are frequently the most abundant immune cells in murine and human cancers and are associated with poor survival. Here we generated TAM molecular signatures from K14cre;Cdh1flox/flox;Trp53flox/flox (KEP) and MMTV-NeuT (NeuT) transgenic mice which resemble human invasive lobular carcinoma (ILC) and HER2+ tumors, respectively. Determination of TAM-specific signatures in breast cancer required relationship analysis with healthy mammary tissue macrophages, since comparison with other macrophage populations overestimated TAM-specific gene expression. TAMs from the two models featured a distinct transcriptomic profile and KEP-derived signatures reliably predicted outcome in ILC patients, indicating that translation of murine TAM signatures to patients warrants consideration of the cancer subtype. Collectively, we show that a transgenic mouse tumor model can be utilized to derive a TAM signature for human breast cancer outcome prediction and we provide a generalizable strategy for determining and applying immune cell signatures provided the murine model reflects the human disease.

Breast cancer outcome prediction with tumour tissue images and machine learning

PurposeRecent advances in machine learning have enabled better understanding of large and complex visual data. Here, we aim to investigate patient outcome prediction with a machine learning method using only an image of tumour sample as an input.MethodsUtilising tissue microarray (TMA) samples obtained from the primary tumour of patients (N = 1299) within a nationwide breast cancer series with long-term-follow-up, we train and validate a machine learning method for patient outcome prediction. The prediction is performed by classifying samples into low or high digital risk score (DRS) groups. The outcome classifier is trained using sample images of 868 patients and evaluated and compared with human expert classification in a test set of 431 patients.ResultsIn univariate survival analysis, the DRS classification resulted in a hazard ratio of 2.10 (95% CI 1.33–3.32, p = 0.001) for breast cancer-specific survival. The DRS classification remained as an independent predictor of breast cancer-specific survival in a multivariate Cox model with a hazard ratio of 2.04 (95% CI 1.20–3.44, p = 0.007). The accuracy (C-index) of the DRS grouping was 0.60 (95% CI 0.55–0.65), as compared to 0.58 (95% CI 0.53–0.63) for human expert predictions based on the same TMA samples.ConclusionsOur findings demonstrate the feasibility of learning prognostic signals in tumour tissue images without domain knowledge. Although further validation is needed, our study suggests that machine learning algorithms can extract prognostically relevant information from tumour histology complementing the currently used prognostic factors in breast cancer.

Association between Ki-67 expression and clinicopathological features in prognosis of breast cancer: A retrospective cohort study.

Background:Breast cancer is the most common diagnosed female cancer. Breast cancer is also the leading cause of cancer death in females accounting for 13.7% of female cancer-related mortality globally. Variable known prognostic factors such as histological tumor type, tumor size, nodal status, grade, age, and estrogen receptor (ER) status and the proliferation marker – Ki-67 influence the type of treatment decision. The purpose of this present study is to investigate the association between Ki-67 expression with several clinicopathological variables and patients’ outcome.Materials and Methods:This is a retrospective cohort study from September 2008 to March 2017; 165 newly diagnosed breast cancer patients were enrolled in the study. Ki67 levels were measured using immunohistochemistry and compared with clinicopathological variables. The relation of Ki67 expression with disease-free survival (DFS) and overall survival (OS) was also analyzed.Results:The result of this study revealed that age, tumor size, menopausal status, and human epidermal growth factor receptor 2 (HER2) status had no effect on the patients’ outcome. Patients with ER-positive, progesterone receptor (PR)-positive, and HER2-negative tumors expressed a higher rate of Ki-67 (>10%) than patients with ER-negative, PR-negative, and HER2-positive tumors, respectively. However, we found that Ki-67 levels were not significantly increased statistically with ER, PR, and HER2 statuses. There was a statistically significant correlation between Ki-67 expression and with higher stages of the disease. Multivariate analysis showed that Ki-67 expression could not to be an independent prognostic factor for 5-year OS and DFS. Furthermore, p53 status was only prognostic factor for 5-year OS whereas higher stages of disease and p53 status were prognostic factors for 5-year DFS.Conclusion:Ki67 could not be an independent variable for prediction of breast cancer outcome.

BLASSO: integration of biological knowledge into a regularized linear model

BackgroundIn RNA-Seq gene expression analysis, a genetic signature or biomarker is defined as a subset of genes that is probably involved in a given complex human trait and usually provide predictive capabilities for that trait. The discovery of new genetic signatures is challenging, as it entails the analysis of complex-nature information encoded at gene level. Moreover, biomarkers selection becomes unstable, since high correlation among the thousands of genes included in each sample usually exists, thus obtaining very low overlapping rates between the genetic signatures proposed by different authors. In this sense, this paper proposes BLASSO, a simple and highly interpretable linear model with l1-regularization that incorporates prior biological knowledge to the prediction of breast cancer outcomes. Two different approaches to integrate biological knowledge in BLASSO, Gene-specific and Gene-disease, are proposed to test their predictive performance and biomarker stability on a public RNA-Seq gene expression dataset for breast cancer. The relevance of the genetic signature for the model is inspected by a functional analysis.ResultsBLASSO has been compared with a baseline LASSO model. Using 10-fold cross-validation with 100 repetitions for models’ assessment, average AUC values of 0.7 and 0.69 were obtained for the Gene-specific and the Gene-disease approaches, respectively. These efficacy rates outperform the average AUC of 0.65 obtained with the LASSO. With respect to the stability of the genetic signatures found, BLASSO outperformed the baseline model in terms of the robustness index (RI). The Gene-specific approach gave RI of 0.15±0.03, compared to RI of 0.09±0.03 given by LASSO, thus being 66% times more robust. The functional analysis performed to the genetic signature obtained with the Gene-disease approach showed a significant presence of genes related with cancer, as well as one gene (IFNK) and one pseudogene (PCNAP1) which a priori had not been described to be related with cancer.ConclusionsBLASSO has been shown as a good choice both in terms of predictive efficacy and biomarker stability, when compared to other similar approaches. Further functional analyses of the genetic signatures obtained with BLASSO has not only revealed genes with important roles in cancer, but also genes that should play an unknown or collateral role in the studied disease.

Hadamard Kernel SVM with applications for breast cancer outcome predictions

BackgroundBreast cancer is one of the leading causes of deaths for women. It is of great necessity to develop effective methods for breast cancer detection and diagnosis. Recent studies have focused on gene-based signatures for outcome predictions. Kernel SVM for its discriminative power in dealing with small sample pattern recognition problems has attracted a lot attention. But how to select or construct an appropriate kernel for a specified problem still needs further investigation.ResultsHere we propose a novel kernel (Hadamard Kernel) in conjunction with Support Vector Machines (SVMs) to address the problem of breast cancer outcome prediction using gene expression data. Hadamard Kernel outperform the classical kernels and correlation kernel in terms of Area under the ROC Curve (AUC) values where a number of real-world data sets are adopted to test the performance of different methods.ConclusionsHadamard Kernel SVM is effective for breast cancer predictions, either in terms of prognosis or diagnosis. It may benefit patients by guiding therapeutic options. Apart from that, it would be a valuable addition to the current SVM kernel families. We hope it will contribute to the wider biology and related communities.

Improved prediction of breast cancer outcome by identifying heterogeneous biomarkers.

Identification of genes that can be used to predict prognosis in patients with cancer is important in that it can lead to improved therapy, and can also promote our understanding of tumor progression on the molecular level. One of the common but fundamental problems that render identification of prognostic genes and prediction of cancer outcomes difficult is the heterogeneity of patient samples. To reduce the effect of sample heterogeneity, we clustered data samples using K-means algorithm and applied modified PageRank to functional interaction (FI) networks weighted using gene expression values of samples in each cluster. Hub genes among resulting prioritized genes were selected as biomarkers to predict the prognosis of samples. This process outperformed traditional feature selection methods as well as several network-based prognostic gene selection methods when applied to Random Forest. We were able to find many cluster-specific prognostic genes for each dataset. Functional study showed that distinct biological processes were enriched in each cluster, which seems to reflect different aspect of tumor progression or oncogenesis among distinct patient groups. Taken together, these results provide support for the hypothesis that our approach can effectively identify heterogeneous prognostic genes, and these are complementary to each other, improving prediction accuracy. https://github.com/mathcom/CPR. jgahn@inu.ac.kr. Supplementary data are available at Bioinformatics online.

Quantification of hypoxia-related gene expression as a potential approach for clinical outcome prediction in breast cancer

Breast cancers are solid tumors frequently characterized by regions with low oxygen concentrations. Cellular adaptations to hypoxia are mainly determined by “hypoxia inducible factors” that mediate transcriptional modifications involved in drug resistance and tumor progression leading to metastasis and relapse occurrence. In this study, we investigated the prognostic value of hypoxia-related gene expression in breast cancer. A systematic review was conducted to select a set of 45 genes involved in hypoxia signaling pathways and breast tumor progression. Gene expression was quantified by RT-qPCR in a retrospective series of 32 patients with invasive ductal carcinoma. Data were analyzed in relation to classical clinicopathological criteria and relapse occurrence. Coordinated overexpression of selected genes was observed in high-grade and HER2+ tumors. Hierarchical cluster analysis of gene expression significantly segregated relapsed patients (p = 0.008, Chi2 test). All genes (except one) were up-regulated and six markers were significantly expressed in tumors from recurrent patients. The expression of this 6-gene set was used to develop a basic algorithm for identifying recurrent patients according to a risk score of relapse. Analysis of Kaplan-Meier relapse-free survival curves allowed the definition of a threshold score of 2 (p = 0.021, Mantel-Haenszel test). The risk of recurrence was increased by 40% in patients with a high score. In addition to classical prognostic factors, we showed that hypoxic markers have potential prognostic value for outcome and late recurrence prediction, leading to improved treatment decision-making for patients with early-stage invasive breast cancer. It will be necessary to validate the clinical relevance of this prognostic approach through independent studies including larger prospective patient cohorts.

Use of Digital Image Analysis for Outcome Prediction in Breast Cancer

Introduction/ Background Breast cancer is the most common cancer in theUK. Although 10 year survival has increased over last decades, significant improvement is still needed. Clinical management decisions are largely dependent on assessment of histological features. The traditional approach to histopathological assessment has been the expert manual reporting of cases as viewed with a light microscope and has remained virtually unchanged since 1928, with minor modifications that have led to the current routinely applied semi-quantitative tumour grading system. However, the abundance of information within the tumour microenvironment is not reflected in the traditionally evaluated histological features, and there remain morphological features with prognostic potential that have previously been beyond investigation by traditional manual microscopic means. Tumour prognosis is closely related to metastasis, a complex process involving tumour cell migration through the stromal microenvironment before entering the lymphovascular compartment. Tumour/stromal interaction is crucial in the process and represents a potential candidate for therapeutic intervention. This interaction is partly affected by the pattern of tumour migration, revealed in the tumour architecture, and partly by the stromal response. Aims Our working hypothesis for the proposed study framework was that, with the application of digital image analysis technology, previously unquantifiable tumour architectural and microenvironmental features can be rigorously assessed in detail and tested as potential prognostic parameters. Quantified features included tumour extracellular particles at the tumour-stroma interface, tumour infiltrating lymphocytes, tumour nest perimeter, number, size and shape. The selected prognostic parameter was axillary lymph node metastasis. Methods Our initial study included diagnostic core biopsies from 19 HER2 positive breast cancers, with approximately equal number of ER strongly positive or weakly positive/ negative cases. Her2 immunohistochemistry allowed rigorous segregation of epithelial elements. Immunostained sections were digitised using a Hamamatsu scanner and x10 magnification consecutive segments from .ndpi files were captured as .jpeg files and analysed using Fiji (Image J), a public domain image processing program. The entirety of each core was examined in all cases. Several native Fiji Functions and Fiji plugins, including Trainable Weka Segmentation, Colour Segmentation and Colour Deconvolution were employed in different combinations for different types of analysis. The analysis is currently being expanded to a large set of digitised breast cancer tissue microarray (TMA) slides which have been stained with cytokeratin to highlight tumour cells. The set includes breast cancer cases from all molecular subtypes and is linked with detailed histological and outcome data. Results Increased number of extracellular particles at the tumour-stroma interface and decreased number of tumour-infiltrating lymphocytes were significantly associated with axillary lymph node metastasis (p=0.0062 and p=0.0154 respectively). Combination of the two parameters increased further the strength of the association (p=0.0011). Increased total tumour nest perimeter, tumour nest number and tumour nest shape irregularity were also significantly associated with axillary lymph node metastasis (p=0.0288, p=0.0085 and 0.0203 respectively). Data from the analysis of TMAs are currently analyzed and will be presented.

FERAL: network-based classifier with application to breast cancer outcome prediction.

Motivation: Breast cancer outcome prediction based on gene expression profiles is an important strategy for personalize patient care. To improve performance and consistency of discovered markers of the initial molecular classifiers, network-based outcome prediction methods (NOPs) have been proposed. In spite of the initial claims, recent studies revealed that neither performance nor consistency can be improved using these methods. NOPs typically rely on the construction of meta-genes by averaging the expression of several genes connected in a network that encodes protein interactions or pathway information. In this article, we expose several fundamental issues in NOPs that impede on the prediction power, consistency of discovered markers and obscures biological interpretation.Results: To overcome these issues, we propose FERAL, a network-based classifier that hinges upon the Sparse Group Lasso which performs simultaneous selection of marker genes and training of the prediction model. An important feature of FERAL, and a significant departure from existing NOPs, is that it uses multiple operators to summarize genes into meta-genes. This gives the classifier the opportunity to select the most relevant meta-gene for each gene set. Extensive evaluation revealed that the discovered markers are markedly more stable across independent datasets. Moreover, interpretation of the marker genes detected by FERAL reveals valuable mechanistic insight into the etiology of breast cancer.Availability and implementation: All code is available for download at: http://homepage.tudelft.nl/53a60/resources/FERAL/FERAL.zip.Contact: j.deridder@tudelft.nlSupplementary information: Supplementary data are available at Bioinformatics online.

Individualization of tamoxifen therapy: much more than just CYP2D6 genotyping.

Clinical response to tamoxifen varies widely among women treated with this drug for hormone receptor-positive breast cancer. The principal active metabolite - endoxifen - is generated through hepatic metabolism of tamoxifen, with key roles for cytochrome P450 (CYP) CYP2D6 and CYP3A. By influencing endoxifen formation, genetic variants of CYP2D6 may affect response to tamoxifen. After a decade of research, examining the effects of CYP2D6 genetic variants on tamoxifen efficacy, there is still no agreement on the clinical utility of CYP2D6 genotype as biomarker for the prediction of breast cancer outcome, because studies revealed conflicting results. However, tamoxifen metabolism is complex and involves several other drug-metabolizing enzymes. Genetic variants of other CYP enzymes, including CYP3A4 and CYP2C9/19, as well as co-medication interfering with the metabolic activity of CYP2D6 and CYP3A4 have been shown to affect endoxifen concentrations and may also contribute to the variability in response to tamoxifen. Phenotyping strategies can predict endoxifen exposure more accurately than CYP2D6 genotype, but do not take into account all factors influencing endoxifen exposure. Therapeutic drug monitoring (TDM) is likely to be the optimal strategy for individualization of tamoxifen treatment. According to a growing amount of literature, endoxifen concentration seems to be a predictor of clinical outcome. The relationship between endoxifen levels and breast cancer outcomes has to be replicated and confirmed and the value of TDM should be evaluated in prospective clinical trials. Caution is advised regarding the concomitant use of medications which could interact with tamoxifen, including inhibitors and inducers of CYP enzymes.

Optimized outcome prediction in breast cancer by combining the 70-gene signature with clinical risk prediction algorithms.

Clinical guidelines for breast cancer treatment differ in their selection of patients at a high risk of recurrence who are eligible to receive adjuvant systemic treatment (AST). The 70-gene signature is a molecular tool to better guide AST decisions. The aim of this study was to evaluate whether adding the 70-gene signature to clinical risk prediction algorithms can optimize outcome prediction and consequently treatment decisions in early stage, node-negative breast cancer patients. A 70-gene signature was available for 427 patients participating in the RASTER study (cT1-3N0M0). Median follow-up was 61.6 months. Based on 5-year distant-recurrence free interval (DRFI) probabilities survival areas under the curve (AUC) were calculated and compared for risk estimations based on the six clinical risk prediction algorithms: Adjuvant! Online (AOL), Nottingham Prognostic Index (NPI), St. Gallen (2003), the Dutch National guidelines (CBO 2004 and NABON 2012), and PREDICT plus. Also, survival AUC were calculated after adding the 70-gene signature to these clinical risk estimations. Systemically untreated patients with a high clinical risk estimation but a low risk 70-gene signature had an excellent 5-year DRFI varying between 97.1 and 100 %, depending on the clinical risk prediction algorithms used in the comparison. The best risk estimation was obtained in this cohort by adding the 70-gene signature to CBO 2012 (AUC: 0.644) and PREDICT (AUC: 0.662). Clinical risk estimations by all clinical algorithms improved by adding the 70-gene signature. Patients with a low risk 70-gene signature have an excellent survival, independent of their clinical risk estimation. Adding the 70-gene signature to clinical risk prediction algorithms improves risk estimations and therefore might improve the identification of early stage node-negative breast cancer patients for whom AST has limited value. In this cohort, the PREDICT plus tool in combination with the 70-gene signature provided the best risk prediction.Electronic supplementary materialThe online version of this article (doi:10.1007/s10549-014-2954-2) contains supplementary material, which is available to authorized users.

Molecular Characteristics and Metastasis Predictor Genes of Triple-Negative Breast Cancer: A Clinical Study of Triple-Negative Breast Carcinomas

BackgroundTriple-negative breast cancer is a subtype of breast cancer with aggressive tumor behavior and distinct disease etiology. Due to the lack of an effective targeted medicine, treatment options for triple-negative breast cancer are few and recurrence rates are high. Although various multi-gene prognostic markers have been proposed for the prediction of breast cancer outcome, most of them were proven clinically useful only for estrogen receptor-positive breast cancers. Reliable identification of triple-negative patients with a favorable prognosis is not yet possible.Methodology/Principal FindingsClinicopathological information and microarray data from 157 invasive breast carcinomas were collected at National Taiwan University Hospital from 1995 to 2008. Gene expression data of 51 triple-negative and 106 luminal breast cancers were generated by oligonucleotide microarrays. Hierarchical clustering analysis revealed that the majority (94%) of triple-negative breast cancers were tightly clustered together carrying strong basal-like characteristics. A 45-gene prognostic signature giving 98% predictive accuracy in distant recurrence of our triple-negative patients was determined using the receiver operating characteristic analysis and leave-one-out cross validation. External validation of the prognostic signature in an independent microarray dataset of 59 early-stage triple-negative patients also obtained statistical significance (hazard ratio 2.29, 95% confidence interval (CI) 1.04–5.06, Cox P = 0.04), outperforming five other published breast cancer prognostic signatures. The 45-gene signature identified in this study revealed that TGF-β signaling of immune/inflammatory regulation may play an important role in distant metastatic invasion of triple-negative breast cancer.Conclusions/SignificanceGene expression data and recurrence information of triple-negative breast cancer were collected and analyzed in this study. A novel set of 45-gene signature was found to be statistically predictive in disease recurrence of triple-negative breast cancer. The 45-gene signature, if further validated, may be a clinically useful tool in risk assessment of distant recurrence for early-stage triple-negative patients.

Molecular characteristics and metastasis predictor genes of triple-negative breast cancer.

1043 Background: Triple-negative breast cancer(TNBC) is a subtype of breast cancer with aggressive tumor behavior and distinct disease etiology. Due to the lack of an effective targeted medicine, treatment options for triple-negative breast cancer are few and recurrence rates are high. Although various multi-gene prognostic markers have been proposed for the prediction of breast cancer outcome, most of them were proven clinically useful only for estrogen receptor-positive breast cancers. Reliable identification of triple-negative patients with a favorable prognosis is not yet possible. Methods: Clinicopathological information and microarray data from 157 invasive breast carcinomas were collected at National Taiwan University Hospital from 1995 to 2008. Gene expression data of 51 triple-negative and 106 luminal breast cancers were generated with oligonucleotide microarrays. A prognostic 45-gene signature for triple-negative breast cancer was identified using Student’s t test and receiver operating characteristic analysis. Results: Hierarchical clustering analysis revealed that the majority (94%) of triple-negative breast cancers were tightly clustered together carrying strong basal-like characteristics. A novel 45-gene signature giving 98% predictive accuracy in distant metastasis recurrence was identified in our triple-negative patient cohort. External validation of the prognostic signature in an independent microarray dataset of 59 early-stage triple-negative patients also obtained statistical significance (hazard ratio 2.29, 95% CI 1.04-5.06, Cox P = 0.04), outperforming five other published breast cancer prognostic signatures. The prognostic signature was statistically predictive with the node-negative triple-negative patients in the validation cohort. Conclusions: The 45-gene prognostic signature identified in this study revealed that TGF-β signaling in immune/inflammatory regulation may be critically involved in distant metastatic invasion of TNBC. The 45-gene signature, if further validated, may be a clinically useful tool in risk assessment of metastasis recurrence for early-stage triple-negative patients.

A Critical Evaluation of Network and Pathway-Based Classifiers for Outcome Prediction in Breast Cancer

Recently, several classifiers that combine primary tumor data, like gene expression data, and secondary data sources, such as protein-protein interaction networks, have been proposed for predicting outcome in breast cancer. In these approaches, new composite features are typically constructed by aggregating the expression levels of several genes. The secondary data sources are employed to guide this aggregation. Although many studies claim that these approaches improve classification performance over single genes classifiers, the gain in performance is difficult to assess. This stems mainly from the fact that different breast cancer data sets and validation procedures are employed to assess the performance. Here we address these issues by employing a large cohort of six breast cancer data sets as benchmark set and by performing an unbiased evaluation of the classification accuracies of the different approaches. Contrary to previous claims, we find that composite feature classifiers do not outperform simple single genes classifiers. We investigate the effect of (1) the number of selected features; (2) the specific gene set from which features are selected; (3) the size of the training set and (4) the heterogeneity of the data set on the performance of composite feature and single genes classifiers. Strikingly, we find that randomization of secondary data sources, which destroys all biological information in these sources, does not result in a deterioration in performance of composite feature classifiers. Finally, we show that when a proper correction for gene set size is performed, the stability of single genes sets is similar to the stability of composite feature sets. Based on these results there is currently no reason to prefer prognostic classifiers based on composite features over single genes classifiers for predicting outcome in breast cancer.