Identification of a four-gene metabolic signature predicting overall survival for hepatocellular carcinoma

While hundreds of consistently altered metabolic genes had been identified in hepatocellular carcinoma (HCC), the prognostic role of them remains to be further elucidated. Messenger RNA expression profiles and clinicopathological data were downloaded from The Cancer Genome Atlas—Liver Hepatocellular Carcinoma and GSE14520 data set from the Gene Expression Omnibus database. Univariate Cox regression analysis and lasso Cox regression model established a novel four-gene metabolic signature (including acetyl-CoA acetyltransferase 1, glutamic-oxaloacetic transaminase 2, phosphatidylserine synthase 2, and uridine-cytidine kinase 2) for HCC prognosis prediction. Patients in the high-risk group shown significantly poorer survival than patients in the low-risk group. The signature was significantly correlated with other negative prognostic factors such as higher α-fetoprotein. The signature was found to be an independent prognostic factor for HCC survival. Nomogram including the signature shown some clinical net benefit for overall survival prediction. Furthermore, gene set enrichment analyses revealed several significantly enriched pathways, which might help explain the underlying mechanisms. Our study identified a novel robust four-gene metabolic signature for HCC prognosis prediction. The signature might reflect the dysregulated metabolic microenvironment and provided potential biomarkers for metabolic therapy and treatment response prediction in HCC.     

Development and validation of a CTNNB1-associated metabolic prognostic model for hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is a heterogeneous malignancy closely related to metabolic reprogramming. We investigated how CTNNB1 mutation regulates the HCC metabolic phenotype and thus affects the prognosis of HCC. We obtained the mRNA expression profiles and clinicopathological data from The Cancer Genome Atlas (TCGA), the International Cancer Genomics Consortium (ICGC) and the Gene Expression Omnibus database ( GSE14520 and GSE116174 ). We conducted gene set enrichment analysis on HCC patients with and without mutant CTNNB1 through TCGA dataset. The Kaplan-Meier analysis and univariate Cox regression analysis assisted in screening metabolic genes related to prognosis, and the prognosis model was constructed using the Lasso and multivariate Cox regression analysis. The prognostic model showed good prediction performance in both the training cohort (TCGA) and the validation cohorts (ICGC, GSE14520 , GSE116174 ), and the high-risk group presented obviously poorer overall survival compared with low-risk group. Cox regression analysis indicated that the risk score can be used as an independent predictor for the overall survival of HCC. The immune infiltration in different risk groups was also evaluated in this study to explore underlying mechanisms. This study is also the first to describe an metabolic prognostic model associated with CTNNB1 mutations and could be implemented for determining the prognoses of individual patients in clinical practice.     

A prognostic five long–noncoding RNA signature for patients with rectal cancer

This study aimed to identify prognostic long noncoding RNAs (lncRNAs) signature for predicting the prognosis of patients with rectal cancer. LncRNA-sequencing data and clinicopathological data of patients with rectal cancer were retrieved from The Cancer Genome Atlas database. Univariate and multivariate Cox proportional hazards regression analysis, the least absolute shrinkage, and selection operator analysis and the Kaplan-Meier curve method were employed to identify prognostic lncRNAs and construct multi-lncRNA signature. Finally, five lncRNAs (AC079789.1, AC106900.2, AL121987.1, AP004609.1, and LINC02163) were identified to construct a five-lncRNA signature. According to the five-lncRNA signature, patients with rectal cancer were divided into a high-risk group and low-risk group. Patients with rectal cancer had significantly poorer overall survival in the high-risk group than in the low-risk group. We used a time-dependent receiver operating characteristic curve to assess the power of the five-lncRNA signature by calculating the area under the curve (AUC). The AUCs for predicting 3-year survival and 5-year survival were 0.742 and 0.935, respectively, which indicated a good performance of the five-lncRNA signature. The five-lncRNA signature was independently associated with the prognosis of patients with rectal cancer through using univariate and multivariate Cox regression analysis. The biological function of the five lncRNAs was enriched in some cancer-related biological processes and pathways by performing functional enrichment analysis of their correlated protein-coding genes. In conclusion, we developed a five-lncRNA signature as a potential indicator for rectal cancer.     

Integrated analysis of gene expression and DNA methylation datasets identified key genes and a 6-gene prognostic signature for primary lung adenocarcinoma

Lung adenocarcinoma (LUAD) is the main subtype of non-small cell lung cancer with a poor survival prognosis. In our study, gene expression, DNA methylation, and clinicopathological data of primary LUAD were utilized to identify potential prognostic markers for LUAD, which were recruited from The Cancer Genome Atlas (TCGA) database. Univariate regression analysis showed that there were 21 methylation-associated DEGs related to overall survival (OS), including 9 down- and 12 up-regulated genes. The 12 up-regulated genes with hypomethylation may be risky genes, whereas the other 9 down-regulated genes with hypermethylation might be protective genes. By using the Step-wise multivariate Cox analysis, a methylation-associated 6-gene (consisting of CCL20, F2, GNPNAT1, NT5E, B3GALT2, and VSIG2) prognostic signature was constructed and the risk score based on this gene signature classified patients into high- or low-risk groups. Patients of the high-risk group had shorter OS than those of the low-risk group in both the training and validation cohort. Multivariate Cox analysis and the stratified analysis revealed that the risk score was an independent prognostic factor for LUAD patients. The methylation-associated gene signature may serve as a prognostic factor for LUAD patients and the represent hypermethylated or hypomethylated genes might be potential targets for LUAD therapy.     

Comprehensive characterization of pathological stage-related genes of papillary thyroid cancer along with survival prediction

It is crucial to understand the differences across papillary thyroid cancer (PTC) stages, so as to provide a basis for individualized treatments. Here, comprehensive function characterization of PTC stage-related genes was performed and a new prognostic signature was developed for advanced patients. Two gene modules were confirmed to be closely associated with PTC stages and further six hub genes were identified that yield excellent diagnostic efficiency between tumour and normal tissues. Genetic alteration analysis indicates that they are much conservative since mutations in the DNA of them rarely occur, but changes of DNA methylation on these six genes show that 12 DNA methylation sites are significantly associated with their corresponding genes' expression. Validation data set testing also suggests that these six stage-related hub genes would be probably potential biomarkers for marking four stages. Subsequently, a 21-mRNA-based prognostic risk model was constructed for PTC stage III/IV patients and it could effectively predict the survival of patients with strong prognostic ability. Functional analysis shows that differential expression genes between high- and low-risk patients would promote the progress of PTC to some extent. Moreover, tumour microenvironment (TME) of high-risk patients may be more conducive to tumour growth by ESTIMATE analysis.     

Construction of a prognostic model for non-small-cell lung cancer based on ferroptosis-related genes

We wished to construct a prognostic model based on ferroptosis-related genes and to simultaneously evaluate the performance of the prognostic model and analyze differences between high-risk and low-risk groups at all levels. The gene-expression profiles and relevant clinical data of patients with non-small-cell lung cancer (NSCLC) were downloaded from public databases. Differentially expressed genes (DEGs) were obtained by analyzing differences between cancer tissues and paracancerous tissues, and common genes between DEGs and ferroptosis-related genes were identified as candidate ferroptosis-related genes. Next, a risk-score model was constructed using univariate Cox analysis and least absolute shrinkage and selection operator (Lasso) analysis. According to the median risk score, samples were divided into high-risk and low-risk groups, and a series of bioinformatics analyses were conducted to verify the predictive ability of the model. Single-sample gene set enrichment analysis (ssGSEA) was used to investigate differences in immune status between high-risk and low-risk groups, and differences in gene mutations between the two groups were investigated. A risk-score model was constructed based on 21 ferroptosis-related genes. A Kaplan–Meier curve and receiver operating characteristic curve showed that the model had good prediction ability. Univariate and multivariate Cox analyses revealed that ferroptosis-related genes associated with the prognosis may be used as independent prognostic factors for the overall survival time of NSCLC patients. The pathways enriched with DEGs in low-risk and high-risk groups were analyzed, and the enriched pathways were correlated significantly with immunosuppressive status.     

基于基因组不稳定性相关lncRNA的宫颈癌患者预后模型

结合TCGA数据库中宫颈癌的lncRNA表达谱和体细胞突变谱,构建基于突变假设的计算框架,鉴定出36个与宫颈癌基因组不稳定性相关的lncRNA;对其共表达的基因功能进行分析,发现与36个lncRNA共表达的基因在2-氧代戊二酸代谢过程和2-氧羧酸代谢通路中富集。构建了基于基因组不稳定性衍生的两个lncRNA的基因特征(GILncSig),将Train组患者分为高风险组和低风险组,两组患者生存率显著不同,这一结果在Test组患者中得到进一步验证。通过独立预后分析,结果显示GILncSig可独立于其他临床性状,作为宫颈癌患者的整体生存相关独立预后因子。总之,本研究为进一步探讨lncRNA在基因组不稳定性中的作用提供了关键的方法和资源,为识别基因组不稳定性相关的肿瘤标志物提供了新的预测方法。     

Novel gene signatures for prognosis prediction in ovarian cancer

Ovarian cancer (OV) is one of the leading causes of cancer deaths in women worldwide. Late diagnosis and heterogeneous treatment result to poor survival outcomes for patients with OV. Therefore, we aimed to develop novel biomarkers for prognosis prediction from the potential molecular mechanism of tumorigenesis. Eight eligible data sets related to OV in GEO database were integrated to identify differential expression genes (DEGs) between tumour tissues and normal. Enrichment analyses discovered DEGs were most significantly enriched in G2/M checkpoint signalling pathway. Subsequently, we constructed a multi‐gene signature based on the LASSO Cox regression model in the TCGA database and time‐dependent ROC curves showed good predictive accuracy for 1‐, 3‐ and 5‐year overall survival. Utility in various types of OV was validated through subgroup survival analysis. Risk scores formulated by the multi‐gene signature stratified patients into high‐risk and low‐risk, and the former inclined worse overall survival than the latter. By incorporating this signature with age and pathological tumour stage, a visual predictive nomogram was established, which was useful for clinicians to predict survival outcome of patients. Furthermore, SNRPD1 and EFNA5 were selected from the multi‐gene signature as simplified prognostic indicators. Higher EFNA5 expression or lower SNRPD1 indicated poorer outcome. The correlation between signature gene expression and clinical characteristics was observed through WGCNA. Drug‐gene interaction was used to identify 16 potentially targeted drugs for OV treatment. In conclusion, we established novel gene signatures as independent prognostic factors to stratify the risk of OV patients and facilitate the implementation of personalized therapies.     

Identification of a five-gene signature with prognostic value in colorectal cancer

Colorectal cancer (CRC) ranks as one of the most commonly diagnosed malignancies worldwide. Although mortality rates have been decreasing, the prognosis of CRC patients is still highly dependent on the individual. Therefore, identifying and understanding novel biomarkers for CRC prognosis remains crucial. The gene expression profiles of five-gene expression omnibus (GEO) data sets of CRC were first downloaded. A total of 352 consistent differentially expressed genes (DEGs) were identified for CRC and paired with normal tissues. Functional analysis including gene ontology and Kyoto encyclopedia of genes and genomes pathway enrichment revealed that these DEGs were related to metabolic pathways, tight junctions, and the cell cycle. Ten hub DEGs were identified based on the search tool for the retrieval of interacting genes database and protein–protein interaction networks. By using univariate Cox proportional hazard regression analysis, we found 11 survival-related genes among these DEGs. We finally established a five-gene signature (kinesin family member 15, N-acetyltransferase 2, glutathione peroxidase 3, secretogranin II, and chloride channel accessory 1) with prognostic value in CRC by step multivariate Cox regression analysis. Based on this risk scoring system, patients in the high-risk group had significantly poorer survival results compared with those in the low-risk group (log-rank test, p < 0.0001). Finally, we validated our gene signature scoring system in two independent GEO cohorts (GSE17536 and GSE33113). We found all five of the signature genes to be DEGs in The Cancer Genome Atlas database. In conclusion, our findings suggest that our five DEG-based signature can provide a novel biomarker with useful applications in CRC prognosis.     

Construction of a prognostic signature in Ewing's sarcoma: Based on metabolism-related genes

ObjectiveBy combining the expression profiles of metabolism-related genes (MRGS) with clinical information, the expression quantities of MRGS and the influence on development and prognosis were systematically analyzed, so as to provide a theoretical basis for the clinical study on the prognosis of Ewing's sarcoma.MethodsMRGs expression profiles of 64 patients with Ewing's sarcoma were obtained from GEO dataset. Univariate Cox regression analysis was used to identify metabolization-related differentially expressed genes (DEGs) related with prognosis in Ewing's sarcoma patients. Then, multivariate Cox analysis was used to calculate novel prognostic markers based on metabolism-related DEGs. Besides, We validate the model using ICGC datasets. Finally, the new prognostic index was verified on the basis of the prognostic models.ResultsMultivariate Cox regression analysis identified 74 metabolization-related DEGs, 25 of which were associated with Ewing's sarcoma patients' overall survival. Subsequently, we used 25 DEGs to construct metabolism-related prognostic signature for patients with Ewing's sarcoma. Based on the 18 DEGs regression coefficient, we propose the formula of each patient's risk score, and then divided the patients into high-risk group and low-risk group. The results indicated that the survival rate and survival time were higher in the low-risk group and lower in the high-risk group. Multivariate Cox analysis showed that risk score index was an independent prognostic factor for Ewing's sarcoma.ConclusionThe experimental results suggest that the 18 metabolism-related DEGs marker may be effective in predicting the prognosis of Ewing's sarcoma to some extent, helping to individualize treatment of patients at different risks.     

Construction of a prognostic signature in Ewing's sarcoma: Based on metabolism-related genes

ObjectiveBy combining the expression profiles of metabolism-related genes (MRGS) with clinical information, the expression quantities of MRGS and the influence on development and prognosis were systematically analyzed, so as to provide a theoretical basis for the clinical study on the prognosis of Ewing's sarcoma.MethodsMRGs expression profiles of 64 patients with Ewing's sarcoma were obtained from GEO dataset. Univariate Cox regression analysis was used to identify metabolization-related differentially expressed genes (DEGs) related with prognosis in Ewing's sarcoma patients. Then, multivariate Cox analysis was used to calculate novel prognostic markers based on metabolism-related DEGs. Besides, We validate the model using ICGC datasets. Finally, the new prognostic index was verified on the basis of the prognostic models.ResultsMultivariate Cox regression analysis identified 74 metabolization-related DEGs, 25 of which were associated with Ewing's sarcoma patients' overall survival. Subsequently, we used 25 DEGs to construct metabolism-related prognostic signature for patients with Ewing's sarcoma. Based on the 18 DEGs regression coefficient, we propose the formula of each patient's risk score, and then divided the patients into high-risk group and low-risk group. The results indicated that the survival rate and survival time were higher in the low-risk group and lower in the high-risk group. Multivariate Cox analysis showed that risk score index was an independent prognostic factor for Ewing's sarcoma.ConclusionThe experimental results suggest that the 18 metabolism-related DEGs marker may be effective in predicting the prognosis of Ewing's sarcoma to some extent, helping to individualize treatment of patients at different risks.     

Identification of 4-lncRNA prognostic signature in head and neck squamous cell carcinoma

Deregulated long noncoding RNAs (lncRNA) have been critically implicated in tumorigenesis and serve as novel diagnostic and prognostic biomarkers. Here we sought to develop a prognostic lncRNA signature in patients with head and neck squamous cell carcinoma (HNSCC). Original RNA-seq data of 499 HNSCC samples were retrieved from The Cancer Genome Atlas database, which was randomly divided into training and testing set. Univariate Cox regression survival analysis, robust likelihood-based survival model and random sampling iterations were applied to identify prognostic lncRNA candidates in the training cohort. A prognostic risk score was developed based on the Cox coefficient of four individual lncRNA imputed as follows: (0.14546 × expression level of RP11-366H4.1) + (0.27106 × expression level of LINC01123) + (0.54316 × expression level of RP11-110I1.14) + (−0.48794 × expression level of CTD-2506J14.1). Kaplan-Meier analysis revealed that patients with high-risk score had significantly reduced overall survival as compared with those with low-risk score when patients in training, testing, and validation cohorts were stratified into high- or low-risk subgroups. Multivariate survival analysis further revealed that this 4-lncRNA signature was a novel and important prognostic factor independent of multiple clinicopathological parameters. Importantly, ROC analyses indicated that predictive accuracy and sensitivity of this 4-lncRNA signature outperformed those previously well-established prognostic factors. Noticeably, prognostic score based on quantification of these 4-lncRNA via qRT-PCR in another independent HNSCC cohort robustly stratified patients into subgroups with high or low survival. Taken together, we developed a robust 4-lncRNA prognostic signature for HNSCC that might provide a novel powerful prognostic biomarker for precision oncology.     

Heat shock protein family A member 8 is a prognostic marker for bladder cancer: Evidences based on experiments and machine learning

Heat shock protein member 8 (HSPA8) is one of the most abundant chaperones in eukaryotic cells, but its biological roles in bladder cancer (BC) are largely unclear. First, we observed that HSPA8 was abundant in both cell lines and tissues of BC, and the HSPA8-high group had poorer T stages and overall survival (OS) than the HSPA8-low group in the TCGA patients. Next, when we knocked down HSPA8 in BC cells, the growth and migration abilities were significantly decreased, the apoptosis rates were significantly increased, and the Ki67 fluorescence intensity was decreased in BC cells. Moreover, caspase 3 was significantly decreased with overexpression of HSPA8 in BC cells. After that, a machine learning prognostic model was created based on the expression of HSPA8 by applying LASSO Cox regression in TCGA and GEO patients. The model indicated that the low-risk (LR) group with BC had better tumour stages, lymphovascular invasion, and OS than the high-risk (HR) group. Additionally, the risk score was demonstrated to be an independent risk factor for the prognosis of BC by univariate and multivariate Cox analyses. Moreover, the HR group showed a greater rate of TP53 mutations and was mostly enriched in the ECM-receptor interaction pathway than the LR group. Importantly, lower CD8⁺ T-cell and NK cell infiltration, higher immune exclusion scores, higher expression of PD-L1 and CTLA4 and poorer immune checkpoint therapy effects were found in the HR group. These findings demonstrated how crucial HSPA8 plays a role in determining the prognosis of bladder cancer.     

Gene-expression profiles predict survival of patients with lung adenocarcinoma

Histopathology is insufficient to predict disease progression and clinical outcome in lung adenocarcinoma. Here we show that gene-expression profiles based on microarray analysis can be used to predict patient survival in early-stage lung adenocarcinomas. Genes most related to survival were identified with univariate Cox analysis. Using either two equivalent but independent training and testing sets, or 'leave-one-out' cross-validation analysis with all tumors, a risk index based on the top 50 genes identified low-risk and high-risk stage I lung adenocarcinomas, which differed significantly with respect to survival. This risk index was then validated using an independent sample of lung adenocarcinomas that predicted high- and low-risk groups. This index included genes not previously associated with survival. The identification of a set of genes that predict survival in early-stage lung adenocarcinoma allows delineation of a high-risk group that may benefit from adjuvant therapy.     

Identification of a three-miRNA signature as a novel potential prognostic biomarker in patients with clear cell renal cell carcinoma

Current studies suggest that some microRNAs (miRNAs) are associated with prognosis in clear cell renal cell carcinoma (ccRCC). In this paper, we aimed to identify a miRNAs signature to improve prognostic prediction for ccRCC patients. Using ccRCC RNA-Seq data of The Cancer Genome Atlas (TCGA) database, we identified 177 differentially expressed miRNAs between ccRCC and paracancerous tissue. Then all the ccRCC tumor samples were divided into training set and validation set randomly. Three-miRNA signature including miR130b, miR-18a, and miR-223 were constructed by the least absolute shrinkage and selection operator (LASSO) Cox regression model in training set. According to optimal cut-off value of three-miRNA signature risk score, all the patients could be classified into high-risk group and low-risk group significantly. Survival of patients was significantly different between two groups (hazard ratio, 5.58, 95% confidence interval, 3.17-9.80; P < 0.0001), and three-miRNA signature performed favorably prognostic and predictive accuracy. The results were further validated in the validation set and total set. Multivariate Cox regression analyses and subgroup analyses showed that three-miRNA signature was an independent prognostic factor. Two nomograms that integrated three-miRNA signature and three clinicopathological risk factors were constructed to predict overall survival and disease-free survival after surgery for ccRCC patients. Functional enrichment analysis showed the possible roles of three-miRNA signature in some cancer-associated biological processes and pathways. In conclusion, we developed a novel three-miRNA signature that performed reliable prognostic for patient survival with ccRCC, it might facilitate ccRCC patients counseling and individualize management.     

Prognostic significance of elevated endothelin-1 levels in patients with colorectal cancer

BACKGROUND: Prognostic factors from clinical, laboratory and pathological data of patients with colorectal cancer are essential to identify high-risk groups to whom beneficial adjuvant therapy could be given. Endothelin-1, a growth factor, has been associated with the development and spread of solid tumours. This prospective study was performed to determine whether preoperative plasma big ET-1 levels might be useful as a prognostic indicator in patients with colorectal carcinoma. METHOD: Sixty-five consecutive patients with colorectal cancer confirmed by biopsy were included prospectively into this study over a 12-month period. Plasma samples from a peripheral vein were obtained prior to surgery. Univariate analysis of survival using age (< or > 70 years), sex, Dukes' stage (A&B versus C), tumour size (< or > 50 mm), vascular invasion and plasma big ET-1 levels was performed and significant factors were then analysed with the Cox regression model. RESULTS: Three variables, age, Dukes' tumour stage and plasma big ET-1 levels, were found to have prognostic significance (p<0.05). Factors associated with a poorer prognosis were age >70 years (p=0.02), Dukes' C tumours (p=0.04) and plasma big ET-1 levels >4.2 pg/mL (p=0.02). The Cox regression model identified the same three variables as having independent prognostic value for overall survival. CONCLUSION: Preoperative plasma big ET-1 levels may be useful in predicting overall survival in patients with colorectal cancer. Plasma big ET-1 levels may be useful in the selection of high-risk lymph node-negative patients with colorectal cancer for adjuvant therapy.     

Identification of a risk prediction model for clinical prognosis in HER2 positive breast cancer patients

BackgroundNew biomarkers are needed to identify different clinical outcomes for HER2+ breast cancer (BC).MethodsDifferential genes of HER2+ BC were screened based on TCGA database. We used WGCNA to identify the genes related to the survival. Genetic Algorithm was used to structure risk prediction model. The prognostic model was validated in GSE data.ResultsWe constructed a risk prediction model of 6 genes to identify prognosis of HER2+ BC, including CLEC9A, PLD4, PIM1, PTK2B, AKNAD1 and C15orf27. Kaplan-Meier curve showed that the model effectively distinguished the survival of HER2+ BC patients. The multivariate Cox regression suggested that the risk model was an independent predictor for HER2+ BC. Analysis related to immune showed that significant differences in immune infiltration between high- and low-risk groups classified by the prognostic model.ConclusionsOur study identified a risk prediction model of 6 genes that could distinguish the prognosis of HER2+ BC.     

A novel epithelial-mesenchymal transition molecular signature predicts the oncological outcomes in colorectal cancer

Epithelial-mesenchymal transition (EMT), a biological process involving the transformation of epithelial cells into mesenchymal cells, promotes tumour initiation and metastasis. The aim of this study was to construct an EMT molecular signature for predicting colorectal cancer (CRC) prognosis and evaluate the efficacy of the model. The risk scoring system, constructed by log-rank test and multivariate Cox regression analysis according to EMT-related gene expression in CRC patients from TCGA database, demonstrated the highest correlation with prognosis compared with other parameters in CRC patients. The risk scores were significantly correlated with more lymph node metastasis, distal metastasis and advanced clinical stage of CRC. The model was further successfully validated in two independent external cohorts from GEO database. Furthermore, we developed a nomogram to integrate the EMT signature with the pathological stage of CRC, which was found to perform well in predicting the overall survival. Additionally, this risk scoring model was found to be associated with immune cell infiltration, implying a potential role of EMT involved in immunity regulation in tumour microenvironment. Taken together, our novel EMT molecular model may be useful in identifying high-risk patients who need an intensive follow-up and more aggressive therapy, finally contributing to more precise individualized therapeutic strategies.     

Tamoxifen therapy benefit predictive signature combining with prognostic signature in surgical-only ER-positive breast cancer

Tamoxifen treatment is important assistant for estrogen-receptor-positive breast cancer (BRCA) after resection. This study aimed to identify signatures for predicting the prognosis of patients with BRCA after tamoxifen treatment. Data of gene-specific DNA methylation (DM), as well as the corresponding clinical data for the patients with BRCA, were obtained from The Cancer Genome Atlas and followed by systematic bioinformatics analyses. After mapping these DM CPG sites onto genes, we finally obtained 352 relapse-free survival (RFS) associated DM genes, with which 61,776 gene pairs were combined, including 1,614 gene pairs related to RFS. An 11 gene-pair signature was identified to cluster the 189 patients with BRCA into the surgical low-risk group (136 patients) and high-risk group (53 patients). Then, we further identified a tamoxifen-predictive signature that could classify surgical high-risk patients with significant differences on RFS. Combining surgical-only prognostic signature and tamoxifen-predictive signature, patients were clustered into surgical-only low-risk group, tamoxifen nonbenefit group, and tamoxifen benefit group. In conclusion, we identified that the gene pair PDHA2–APRT could serve as a potential prognostic biomarker for patients with BRCA after tamoxifen treatment.