Correlation-centred variable selection of a gene expression signature to predict breast cancer metastasis.
Ontology highlight
ABSTRACT:
SUBMITTER: Hikichi S
PROVIDER: S-EPMC7220948 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Metastatic cancer remains the leading cause of death for patients with breast cancer. To understand the mechanisms underlying the development of distant metastases to specific sites is therefore important and of potential clinical value. From 157 primary breast tumours of the patients with known metastatic disease, gene expression profiling data were generated and correlated to metastatic behaviour including site-specific metastasis, metastasis pattern and survival outcomes. We analysed gene expression signatures specifically associated with the development of bone metastases. As a validation cohort, we used a published dataset of 376 breast carcinomas for which gene expression data and site-specific metastasis information were available. 80.5 % of luminal-type tumours developed bone metastasis as opposed to 41.7 % of basal and 55.6 % of HER2-like tumours. A novel 15-gene signature identified 82.4 % of the tumours with bone metastasis, 85.2 % of the tumours which had bone metastasis as first site of metastasis and 100 % of the ones with bone metastasis only (p 9.99e-09), in the training set. In the independent dataset, 81.2 % of the positive tested tumours had known metastatic disease to the bone (p 4.28e-10). This 15-gene signature showed much better correlation with the development of bone metastases than previously identified signatures and was predictive in both ER-positive as well as in ER-negative tumours. Multivariate analyses revealed that together with the molecular subtype, our 15-gene expression signature was significantly correlated to bone metastasis status (p <0.001, 95 % CI 3.86-48.02 in the training set; p 0.001, 95 % CI 1.54-5.00 in the independent set). The 15 genes, APOPEC3B, ATL2, BBS1, C6orf61, C6orf167, MMS22L, KCNS1, MFAP3L, NIP7, NUP155, PALM2, PH-4, PGD5, SFT2D2 and STEAP3, encoded mainly membrane-bound molecules with molecular function of protein binding. The expression levels of the up-regulated genes (NAT1, BBS1 and PH-4) were also found to be correlated to epithelial to mesenchymal transition status of the tumour. We have identified a novel 15-gene expression signature associated with the development of bone metastases in breast cancer patients. This bone metastasis signature is the first to be identified using a supervised classification approach in a large series of patients and will help forward research in this area towards clinical applications.
| S-EPMC4819548 | biostudies-literature
Project description:Partial correlation based variable selection method was proposed for normal linear regression models by Bühlmann, Kalisch and Maathuis (2010) as a comparable alternative method to regularization methods for variable selection. This paper addresses two important issues related to partial correlation based variable selection method: (a) whether this method is sensitive to normality assumption, and (b) whether this method is valid when the dimension of predictor increases in an exponential rate of the sample size. To address issue (a), we systematically study this method for elliptical linear regression models. Our finding indicates that the original proposal may lead to inferior performance when the marginal kurtosis of predictor is not close to that of normal distribution. Our simulation results further confirm this finding. To ensure the superior performance of partial correlation based variable selection procedure, we propose a thresholded partial correlation (TPC) approach to select significant variables in linear regression models. We establish the selection consistency of the TPC in the presence of ultrahigh dimensional predictors. Since the TPC procedure includes the original proposal as a special case, our theoretical results address the issue (b) directly. As a by-product, the sure screening property of the first step of TPC was obtained. The numerical examples also illustrate that the TPC is competitively comparable to the commonly-used regularization methods for variable selection.
| S-EPMC5484095 | biostudies-literature
Project description:Purpose:Pathological complete response (pCR) to neoadjuvant chemotherapy (NACT) is associated with favourable outcomes of patients with triple-negative breast cancer (TNBC). However, a proportion of TNBC patients with the residual disease do not relapse and achieve long-term survival. The aim of this study was to identify biomarkers that predict clinical outcomes in these patients. Patients and Methods:A retrospective series of 10 TNBC patients who displayed non-pCR to NACT were included in the discovery cohort. Total RNA from pre-NACT core biopsies and paired surgical specimens were subjected to the Affymetrix Human Transcriptome Array. Gene set enrichment analysis (GSEA) was used to identify signal pathways and gene signatures associated with metastasis. The Cox proportional hazard model and Kaplan-Meier survival curves were employed to assess the prognostic value of the identified signature in two independent TNBC datasets included in Gene Expression Omnibus (GEO). Results:The epithelial-mesenchymal transition (EMT) pathway was markedly more enriched in pre- (NES = 1.92; p.adjust = 0.019) and post-NACT samples (NES = 2.02; p.adjust = 0.010) from patients who developed metastasis after NACT. A subset of 6 EMT genes including LUM, SFRP4, COL6A3, MMP2, CXCL12, and HTRA1 were expressed constantly at higher levels in samples from patients who progressed to metastatic disease. The potential of the 6-EMT gene signature to predict TNBC metastasis after NACT was validated with a GEO dataset (HR=0.36, p=0.0008, 95% CI: 0.200-0.658). Moreover, the signature appeared of predictive value in another GEO dataset of TNBC patients who received surgery followed by adjuvant chemotherapy (HR = 0.46, 95% CI: 0.225-0.937). Conclusion:Expression analysis of the 6-EMT gene signature at diagnosis may be of predictive value for metastasis in TNCB patients who did not achieve pCR to NACT and for patients treated with surgery in combination with adjuvant therapy.
| S-EPMC7342558 | biostudies-literature
Project description:BackgroundVisceral organ metastasis is associated with poor survival outcomes in terms of metastasis free- and overall survival in breast carcinomas. Identification of a gene expression profile in tumours that selects a subpopulation of patients that is more likely to develop visceral organ metastases will help elucidate mechanisms for the development of distant metastases and could be of clinical value. With this study we aimed to determine genomic predictors that would help to distinguish breast cancer patients with more likelihood to develop visceral metastasis.MethodsGene expression profiling data of 157 primary tumours from breast cancer patients who developed distant metastases were analyzed and differentially expressed genes between the group of tumours with visceral metastasis and the those without visceral metastases were identified. Published data were used to validate our findings. Multivariate logistic regression tests were applied to further investigate the association between the gene-expression-signature and clinical variables. Survival analyses were performed by the Kaplan-Meier method.ResultsFourteen differentially expressed genes (WDR6, CDYL, ATP6V0A4, CHAD, IDUA, MYL5, PREP, RTN4IP1, BTG2, TPRG1, ABHD14A, KIF18A, S100PBP and BEND3) were identified between the group of tumours with and without visceral metastatic disease. Five of these genes (CDYL, ATP6V0A4, PREP, RTN4IP1 and KIF18A) were up-regulated and the other genes were down-regulated. This gene expression signature was validated in the training and in the independent data set (p 2.13e- 08 and p 9.68e- 06, respectively). Multivariate analyses revealed that the 14-gene-expression-signature was associated with visceral metastatic disease (p 0.001, 95% CI 1.43-4.27), independent of other clinicopathologic features. This signature has been also found to be associated with survival status of the patients (p < .001).ConclusionWe have identified an unique gene expression signature which is specific to visceral metastasis. This 14-gene-expression-signature may play a role in identifying the subgroup of patients with potential to develop visceral metastasis.
| S-EPMC6454625 | biostudies-literature
Project description:Background: The existing metabolic gene signatures for predicting breast cancer outcomes only focus on gene expression data without considering clinical characteristics. Therefore, this study aimed to establish a predictive risk model combining metabolic enzyme genes and clinicopathological characteristics to predict the overall survival in patients with breast cancer. Methods: Transcriptomics and corresponding clinical data for patients with breast cancer were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. Differentially expressed metabolic genes between tumors and normal tissues were identified in the TCGA dataset (training dataset). A prognostic model was then built using univariate and multifactorial Cox proportional hazards regression analyses in the training dataset. The capability of the predictive model was then assessed using the receiver operating characteristic in both datasets. Pathway enrichment analysis and immune cell infiltration were performed using Kyoto Encyclopedia of Genes and Genomes (KEGG)/Gene Ontology (GO) enrichment and CIBERSORT algorithm, respectively. Results: In breast cancer and normal tissues, 212 metabolic enzyme genes were differentially expressed. The predictive model included four factors: age, stage, and expression of SLC35A2 and PLA2G10. Patients with breast cancer were classified into high- and low-risk groups based on the model; the high-risk group had a significantly poorer overall survival rate than the low-risk group. Furthermore, the two risk groups showed different activation of pathways and alterations in the properties of tumor microenvironment-infiltrating immune cells. Conclusion: We developed a powerful model to predict prognosis in patients with breast cancer by combining the gene expression of metabolic enzymes with clinicopathological characteristics.
| S-EPMC9277072 | biostudies-literature
Project description:PurposeThe identification of a molecular signature predicting the relapse of tamoxifen-treated primary breast cancers should help the therapeutic management of estrogen receptor-positive cancers.Experimental designA series of 132 primary tumors from patients who received adjuvant tamoxifen were analyzed for expression profiles at the whole-genome level by 70-mer oligonucleotide microarrays. A supervised analysis was done to identify an expression signature.ResultsWe defined a 36-gene signature that correctly classified 78% of patients with relapse and 80% of relapse-free patients (79% accuracy). Using 23 independent tumors, we confirmed the accuracy of the signature (78%) whose relevance was further shown by using published microarray data from 60 tamoxifen-treated patients (63% accuracy). Univariate analysis using the validation set of 83 tumors showed that the 36-gene classifier is more efficient in predicting disease-free survival than the traditional histopathologic prognostic factors and is as effective as the Nottingham Prognostic Index or the "Adjuvant!" software. Multivariate analysis showed that the molecular signature is the only independent prognostic factor. A comparison with several already published signatures demonstrated that the 36-gene signature is among the best to classify tumors from both training and validation sets. Kaplan-Meier analyses emphasized its prognostic power both on the whole cohort of patients and on a subgroup with an intermediate risk of recurrence as defined by the St. Gallen criteria.ConclusionThis study identifies a molecular signature specifying a subgroup of patients who do not gain benefits from tamoxifen treatment. These patients may therefore be eligible for alternative endocrine therapies and/or chemotherapy.
| S-EPMC2912334 | biostudies-literature
Project description:Despite the current standard of care, breast cancer remains one of the leading causes of mortality in women worldwide, thus emphasizing the need for better predictive and therapeutic targets. ABI1 is associated with poor survival and an aggressive breast cancer phenotype, although its role in tumorigenesis, metastasis, and the disease outcome remains to be elucidated. Here, we define the ABI1-based seven-gene prognostic signature that predicts survival of metastatic breast cancer patients; ABI1 is an essential component of the signature. Genetic disruption of Abi1 in primary breast cancer tumors of PyMT mice led to significant reduction of the number and size of lung metastases in a gene dose-dependent manner. The disruption of Abi1 resulted in deregulation of the WAVE complex at the mRNA and protein levels in mouse tumors. In conclusion, ABI1 is a prognostic metastatic biomarker in breast cancer. We demonstrate, for the first time, that lung metastasis is associated with an Abi1 gene dose and specific gene expression aberrations in primary breast cancer tumors. These results indicate that targeting ABI1 may provide a therapeutic advantage in breast cancer patients.
| S-EPMC9297774 | biostudies-literature
Project description:BackgroundThe prognosis of lymph node-negative triple-negative breast cancer (TNBC) is still worse than that of other subtypes despite adjuvant chemotherapy. Reliable prognostic biomarkers are required to identify lymph node-negative TNBC patients at a high risk of distant metastasis and optimize individual treatment.MethodsWe analyzed the RNA sequencing data of primary tumor tissue and the clinicopathological data of 202 lymph node-negative TNBC patients. The cohort was randomly divided into training and validation sets. Least absolute shrinkage and selection operator Cox regression and multivariate Cox regression were used to construct the prognostic model.ResultsA clinical prognostic model, seven-gene signature, and combined model were constructed using the training set and validated using the validation set. The seven-gene signature was established based on the genomic variables associated with distant metastasis after shrinkage correction. The difference in the risk of distant metastasis between the low- and high-risk groups was statistically significant using the seven-gene signature (training set: P < 0.001; validation set: P = 0.039). The combined model showed significance in the training set (P < 0.001) and trended toward significance in the validation set (P = 0.071). The seven-gene signature showed improved prognostic accuracy relative to the clinical signature in the training data (AUC value of 4-year ROC, 0.879 vs. 0.699, P = 0.046). Moreover, the composite clinical and gene signature also showed improved prognostic accuracy relative to the clinical signature (AUC value of 4-year ROC: 0.888 vs. 0.699, P = 0.029; AUC value of 5-year ROC: 0.882 vs. 0.693, P = 0.038). A nomogram model was constructed with the seven-gene signature, patient age, and tumor size.ConclusionsThe proposed signature may improve the risk stratification of lymph node-negative TNBC patients. High-risk lymph node-negative TNBC patients may benefit from treatment escalation.
| S-EPMC8481824 | biostudies-literature
Project description:The partially linear model (PLM) is a useful semiparametric extension of the linear model that has been well studied in the statistical literature. This paper proposes a variable selection procedure for the PLM with ultrahigh dimensional predictors. The proposed method is different from the existing penalized least squares procedure in that it relies on partial correlation between the partial residuals of the response and the predictors. We systematically study the theoretical properties of the proposed procedure and prove its model consistency property. We further establish the root-n convergence of the estimator of the regression coefficients and the asymptotic normality of the estimate of the baseline function. We conduct Monte Carlo simulations to examine the finite-sample performance of the proposed procedure and illustrate the proposed method with a real data example.
| S-EPMC6555488 | biostudies-literature
Project description:BackgroundInflammatory breast cancer (IBC) is an aggressive disease. To date, no molecular feature reliably predicts either the response to chemotherapy (CT) or the survival. Using DNA microarrays, we searched for multigene predictors.Patients and methodsThe World IBC Consortium generated whole-genome expression profiles of 137 IBC and 252 non-IBC (nIBC) samples. We searched for transcriptional profiles associated with pathological complete response (pCR) to neoadjuvant anthracycline-based CT and distant metastasis-free survival (DMFS) in respective subsets of 87 and 106 informative IBC samples. Correlations were investigated with predictive and prognostic gene expression signatures published in nIBC (nIBC-GES). Supervised analyses tested genes and activation signatures of 19 biological pathways and 234 transcription factors.ResultsThree of five tested prognostic nIBC-GES and the two tested predictive nIBC-GES discriminated between IBC with and without pCR, as well as two interferon activation signatures. We identified a 107-gene signature enriched for immunity-related genes that distinguished between responders and nonresponders in IBC. Its robustness was demonstrated by external validation in three independent sets including two IBC sets and one nIBC set, with independent significant predictive value in IBC and nIBC validation sets in multivariate analysis. We found no robust signature associated with DMFS in patients with IBC, and neither of the tested prognostic GES, nor the molecular subtypes were informative, whereas they were in our nIBC series (220 stage I-III informative samples).ConclusionDespite the relatively small sample size, we show that response to neoadjuvant CT in IBC is, as in nIBC, associated with immunity-related processes, suggesting that similar mechanisms responsible for pCR exist. Analysis of a larger IBC series is warranted regarding the correlation of gene expression profiles and DMFS.
| S-EPMC3905779 | biostudies-literature