A deep learning survival model for triple-negative breast cancer patients.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. e12594-e12594
Author(s):  
Yiyue Xu ◽  
Bing Zou ◽  
Bingjie Fan ◽  
Wanlong Li ◽  
Shijiang Wang ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Deep Learning ◽  
Random Forest ◽  
Cox Regression ◽  
Learning Model ◽  
Survival Model ◽  
Cox Regression Analysis ◽  
Validation Set ◽  
Deep Learning Model

e12594 Background: Triple-negative breast cancer (TNBC) is the subtype of breast cancer with the worst prognosis. There is no reliable model for survival prediction of TNBC patients. The traditional Cox regression analysis with poor prediction power cannot satisfy the clinical needs. The purpose was to establish a deep learning model and develop a new prognostic system for TNBC patients. Methods: This study collected data of TNBC patients from the Surveillance, Epidemiology, and End Results (SEER) program between 2010 and 2016. 70% were used to develop the deep learning model, 15% were used as the validation set, and 15% as the independent testing set. Then the concordance-index (c-index) and Brier score (IBS) were calculated and compared with the Cox regression analysis and random forest. Finally, according to the classification of the deep survival model, an individualized prognosis system was established. Results: A total of 37,818 patients were enrolled in this study. In the validation set, the c-index of the deep learning was 0.799, which was better than the traditional Cox regression model (0.774) and random forest (0.763). The independent testing set further proved the robustness of the deep survival model (c-index 0.788). The new prognosis system based on the deep survival model reached an area under the curve (AUC) of 0.805, which was better than the Tumor, Node, Metastases (TNM) staging system (0.771). Conclusions: Deep learning model had better prediction power than the Cox regression analysis and the random forest. The established prognosis system can better predict prognosis and aid individual risk stratification for TNBC patients patients.

scholarly journals Immune-related lncRNAs as predictors of survival in breast cancer: a prognostic signature

2020 ◽  
Vol 18 (1) ◽  
Author(s):  
Wei Ma ◽  
Fangkun Zhao ◽  
Xinmiao Yu ◽  
Shu Guan ◽  
Huandan Suo ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Cox Regression ◽  
Risk Group ◽  
Curve Analysis ◽  
Time Dependent ◽  
Training Set ◽  
Roc Curve Analysis ◽  
Cox Regression Analysis ◽  
Validation Set

Abstract Background Breast cancer is a highly heterogeneous disease, this poses challenges for classification and management. Long non-coding RNAs play acrucial role in the breast cancersdevelopment and progression, especially in tumor-related immune processes which have become the most rapidly investigated area. Therefore, we aimed at developing an immune-related lncRNA signature to improve the prognosis prediction of breast cancer. Methods We obtained breast cancer patient samples and corresponding clinical data from The Cancer Genome Atlas (TCGA) database. Immune-related lncRNAs were screened by co-expression analysis of immune-related genes which were downloaded from the Immunology Database and Analysis Portal (ImmPort). Clinical patient samples were randomly separated into training and testing sets. In the training set, univariate Cox regression analysis and LASSO regression were utilized to build a prognostic immune-related lncRNA signature. The signature was validated in the training set, testing set, and whole cohorts by the Kaplan–Meier log-rank test, time-dependent ROC curve analysis, principal component analysis, univariate andmultivariate Cox regression analyses. Results A total of 937 immune- related lncRNAs were identified, 15 candidate immune-related lncRNAs were significantly associated with overall survival (OS). Eight of these lncRNAs (OTUD6B-AS1, AL122010.1, AC136475.2, AL161646.1, AC245297.3, LINC00578, LINC01871, AP000442.2) were selected for establishment of the risk prediction model. The OS of patients in the low-risk group was higher than that of patients in the high-risk group (p = 1.215e − 06 in the training set; p = 0.0069 in the validation set; p = 1.233e − 07 in whole cohort). The time-dependent ROC curve analysis revealed that the AUCs for OS in the first, eighth, and tenth year were 0.812, 0.81, and 0.857, respectively, in the training set, 0.615, 0.68, 0.655 in the validation set, and 0.725, 0.742, 0.741 in the total cohort. Multivariate Cox regression analysis indicated the model was a reliable and independent indicator for the prognosis of breast cancer in the training set (HR = 1.432; 95% CI 1.204–1.702, p < 0.001), validation set (HR = 1.162; 95% CI 1.004–1.345, p = 0.044), and whole set (HR = 1.240; 95% CI 1.128–1.362, p < 0.001). GSEA analysis revealed a strong connection between the signature and immune-related biological processes and pathways. Conclusions We constructed and verified a robust signature of 8 immune-related lncRNAs for the prediction of breast cancer patient survival.

scholarly journals Immune-Related LncRNAs as Predictors of Survival in Breast Cancer: A Prognostic Signature

2020 ◽  
Author(s):  
wei ma ◽  
fangkun zhao ◽  
xinmiao yu ◽  
shu guan ◽  
huandan suo ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Cox Regression ◽  
Risk Group ◽  
Curve Analysis ◽  
Time Dependent ◽  
Training Set ◽  
Roc Curve Analysis ◽  
Cox Regression Analysis ◽  
Validation Set

Abstract Background: Breast cancer is a highly heterogeneous disease, this poses challenges for classification and management. Long non-coding RNAs play acrucial role in the breast cancers development and progression, especially in tumor-related immune processes which have become the most rapidly investigated area. Methods: We obtained breast cancer patient samples and corresponding clinical data from The Cancer Genome Atlas (TCGA) database. Immune-related lncRNAs were screened by co-expression analysis of immune-related genes which were downloaded from the Immunology Database and Analysis Portal (ImmPort). Clinical patient samples were randomly separatedinto training and testing sets. In the training set, univariate Cox regression analysis and LASSO regression were utilized to build a prognostic immune-related lncRNA signature. The signature was validated in the training set, testing set, and whole cohorts by the Kaplan–Meier log-rank test, time-dependent ROC curve analysis, principal component analysis, univariate and multivariate Cox regression analyses. Results: A total of 937 immune- related lncRNAs were identified, 15 candidate immune-related lncRNAs were significantly associated with overall survival (OS). Eight of these lncRNAs (OTUD6B-AS1, AL122010.1, AC136475.2, AL161646.1, AC245297.3, LINC00578, LINC01871, AP000442.2) were selected for establishment of the risk prediction model. The OS of patients in the low-risk group was higher than that of patients in the high-risk group( p= 1.215e−06 in the training set; p =0.0069 in the validation set; p =1.233e−07 in whole cohort). The time-dependent ROC curve analysis revealed that the AUCs for OS in the first, eighth, and tenth year were 0.812, 0.81, and 0.857, respectively, in the training set, 0.615, 0.68, 0.655 in the validation set, and 0.725, 0.742, 0.741 in the total cohort. Multivariate Cox regression analysis indicated the model was a reliable and independent indicator for the prognosis of breast cancer in the training set (HR= 1.432; 95% CI 1.204−1.702, p <0.001), validation set (HR= 1.162; 95% CI 1.004−1.345, p = 0.044), and whole set (HR=1.240; 95% CI 1.128−1.362, p <0.001). GSEA analysis revealed a strong connection between the signature and immune-related biological processes and pathways. Conclusions: We constructed and verified a robust signature of 8 immune-related lncRNAs for the prediction of breast cancer patient survival.

scholarly journals Immune-related lncRNAs as predictors of survival in breast cancer: a prognostic signature

2020 ◽  
Author(s):  
wei ma ◽  
fangkun zhao ◽  
xinmiao yu ◽  
shu guan ◽  
huandan suo ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Cox Regression ◽  
Risk Group ◽  
Curve Analysis ◽  
Time Dependent ◽  
Training Set ◽  
Roc Curve Analysis ◽  
Cox Regression Analysis ◽  
Validation Set

Abstract Background: Breast cancer is a highly heterogeneous disease, this poses challenges for classification and management. Long non-coding RNAs play acrucial role in the breast cancersdevelopment and progression, especially in tumor-related immune processes which have become the most rapidly investigated area. Therefore, we aimed at developing an immune-related lncRNA signature to improve the prognosis prediction of breast cancer.Methods: We obtained breast cancer patient samples and corresponding clinical data from The Cancer Genome Atlas (TCGA) database. Immune-related lncRNAs were screened by co-expression analysis of immune-related genes which were downloaded from the Immunology Database and Analysis Portal (ImmPort). Clinical patient samples were randomly separated into training and testing sets. In the training set, univariate Cox regression analysis and LASSO regression were utilized to build a prognostic immune-related lncRNA signature. The signature was validated in the training set, testing set, and whole cohorts by the Kaplan–Meier log-rank test, time-dependent ROC curve analysis, principal component analysis, univariate andmultivariate Cox regression analyses.Results:A total of 937 immune- related lncRNAs were identified, 15 candidate immune-related lncRNAs were significantly associated with overall survival (OS). Eight of these lncRNAs (OTUD6B-AS1, AL122010.1, AC136475.2, AL161646.1, AC245297.3, LINC00578, LINC01871, AP000442.2) were selected for establishment of the risk prediction model. The OS of patients in the low-risk group was higher than that of patients in the high-risk group(p=1.215e−06 in the training set; p=0.0069 in the validation set; p=1.233e−07 in whole cohort). The time-dependent ROC curve analysis revealed that the AUCs for OS in the first, eighth, and tenth year were 0.812, 0.81, and 0.857, respectively, in the training set,0.615, 0.68, 0.655 in the validation set, and 0.725, 0.742, 0.741 in the total cohort. Multivariate Cox regression analysis indicated the model was a reliable and independent indicator for the prognosis of breast cancer in the training set (HR= 1.432; 95% CI 1.204−1.702, p<0.001), validation set (HR= 1.162; 95% CI 1.004−1.345, p = 0.044), and whole set (HR=1.240; 95% CI 1.128−1.362, p<0.001). GSEA analysis revealed a strong connection between the signature and immune-related biological processes and pathways.Conclusions:We constructed and verified a robust signature of 8 immune-related lncRNAs for the prediction of breast cancer patient survival.

Predicting biochemical recurrence after prostatectomy: Can machine learning beat CAPRA score? Results of a multicentric retrospective analysis on 4,700 patients.

2020 ◽  
Vol 38 (6_suppl) ◽  
pp. 343-343 ◽  
Author(s):  
Paul Sargos ◽  
Nicolas Leduc ◽  
Nicolas Giraud ◽  
Giorgio Gandaglia ◽  
Mathieu Roumiguie ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Logistic Regression ◽  
Regression Analysis ◽  
Deep Learning ◽  
Random Forest ◽  
Biochemical Recurrence ◽  
Cox Regression ◽  
P Value ◽  
Cox Regression Analysis

343 Background: Recent advances in machine learning algorithms and deep learning solutions paved the way for improved accuracy in survival analysis. We aim to investigate the accuracy of conventional machine learning and deep learning methods in the prediction of 3-year biochemical recurrence (BCR) as compared to CAPRA score prediction. Methods: A total of 5043 men who underwent RP between 2000 and 2015 for clinically localized PCa iwere analyzed retrospectively. Three-year BCR was predicted using the following models: CAPRA score, Cox regression analysis, logistic regression, k-nearest neighbor, random forest and densely connected feed-forward neural network classifier. The discrimination of the models was quantified using the C-index or the area under the receiver operating characteristics curve. Results: Patients with CAPRA score 2 and 3 accounted for 64% of the population. C-index measuring performance for the prediction of the three-year BCR for CAPRA score was 0.63. C-index values for k-neighbor classifier, logistic regression, Cox regression analysis, random forest classifier and densely optimized neural network were respectively 0.55, 0.63, 0.64, 0.64 and 0.70 (pairwise, adjusted p-value < 0.01). After inclusion of available post-surgical variables, C-index value reached respectively 0.58, 0.77, 0.74, 0.75 and 0.84 (pairwise, adjusted p-value < 0.05). Conclusions: Our results show that CAPRA score performed poorly in intermediate-risk patients undergoing RP. Densely connected neural networks with simple architecture further increased predictive power with low computational cost. In order to predict 3-years BCR, adding post-surgical features to the model greatly enhanced its performance.

scholarly journals Identification and development of an independent immune-related genes prognostic model for breast cancer

BMC Cancer ◽  
2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Lin Chen ◽  
Yuxiang Dong ◽  
Yitong Pan ◽  
Yuhan Zhang ◽  
Ping Liu ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Cox Regression ◽  
Gene Set Enrichment Analysis ◽  
Risk Scores ◽  
Validation Dataset ◽  
Immune Gene ◽  
Cox Regression Analysis ◽  
Immune Genes ◽  
Immune Related Genes

Abstract Background Breast cancer is one of the main malignant tumors that threaten the lives of women, which has received more and more clinical attention worldwide. There are increasing evidences showing that the immune micro-environment of breast cancer (BC) seriously affects the clinical outcome. This study aims to explore the role of tumor immune genes in the prognosis of BC patients and construct an immune-related genes prognostic index. Methods The list of 2498 immune genes was obtained from ImmPort database. In addition, gene expression data and clinical characteristics data of BC patients were also obtained from the TCGA database. The prognostic correlation of the differential genes was analyzed through Survival package. Cox regression analysis was performed to analyze the prognostic effect of immune genes. According to the regression coefficients of prognostic immune genes in regression analysis, an immune risk scores model was established. Gene set enrichment analysis (GSEA) was performed to probe the biological correlation of immune gene scores. P < 0.05 was considered to be statistically significant. Results In total, 556 immune genes were differentially expressed between normal tissues and BC tissues (p < 0. 05). According to the univariate cox regression analysis, a total of 66 immune genes were statistically significant for survival risk, of which 30 were associated with overall survival (P < 0.05). Finally, a 15 immune genes risk scores model was established. All patients were divided into high- and low-groups. KM survival analysis revealed that high immune risk scores represented worse survival (p < 0.001). ROC curve indicated that the immune genes risk scores model had a good reliability in predicting prognosis (5-year OS, AUC = 0.752). The established risk model showed splendid AUC value in the validation dataset (3-year over survival (OS) AUC = 0.685, 5-year OS AUC = 0.717, P = 0.00048). Moreover, the immune risk signature was proved to be an independent prognostic factor for BC patients. Finally, it was found that 15 immune genes and risk scores had significant clinical correlations, and were involved in a variety of carcinogenic pathways. Conclusion In conclusion, our study provides a new perspective for the expression of immune genes in BC. The constructed model has potential value for the prognostic prediction of BC patients and may provide some references for the clinical precision immunotherapy of patients.

scholarly journals The Investigation of Associations between TP53 rs1042522, BBC3 rs2032809, CCND1 rs9344, EGFR rs2227983 Polymorphisms and Breast Cancer Phenotype and Prognosis

Diagnostics ◽  
2021 ◽  
Vol 11 (8) ◽  
pp. 1419
Author(s):  
Justina Bekampytė ◽  
Agnė Bartnykaitė ◽  
Aistė Savukaitytė ◽  
Rasa Ugenskienė ◽  
Erika Korobeinikova ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Cox Regression ◽  
Cancer Prognosis ◽  
Nucleotide Polymorphisms ◽  
Single Nucleotide ◽  
Cox Regression Analysis ◽  
Pcr Rflp ◽  
Oncological Diseases ◽  
Cancer Phenotype

Breast cancer is one of the most common oncological diseases among women worldwide. Cell cycle and apoptosis—related genes TP53, BBC3, CCND1 and EGFR play an important role in the pathogenesis of breast cancer. However, the roles of single nucleotide polymorphisms (SNPs) in these genes have not been fully defined. Therefore, this study aimed to analyze the association between TP53 rs1042522, BBC3 rs2032809, CCND1 rs9344 and EGFR rs2227983 polymorphisms and breast cancer phenotype and prognosis. For the purpose of the analysis, 171 Lithuanian women were enrolled. Genomic DNA was extracted from peripheral blood; PCR-RFLP was used for SNPs analysis. The results showed that BBC3 rs2032809 was associated with age at the time of diagnosis, disease progression, metastasis and death. CCND1 rs9344 was associated with tumor size, however an association resulted in loss of significance after Bonferroni correction. In survival analysis, significant associations were observed between BBC3 rs2032809 and OS, PFS and MFS. EGFR rs2227983 also showed some associations with OS and PFS (univariate Cox regression analysis). However, the results were in loss of significance (multivariate Cox regression analysis). In conclusion, BBC3 rs2032809 polymorphism was associated with breast cancer phenotype and prognosis. Therefore, it could be applied as potential markers for breast cancer prognosis.

Factors Predictive of Distant Metastases in Patients With Breast Cancer Who Have a Pathologic Complete Response After Neoadjuvant Chemotherapy

2005 ◽  
Vol 23 (28) ◽  
pp. 7098-7104 ◽  
Author(s):  
Ana M. Gonzalez-Angulo ◽  
Sean E. McGuire ◽  
Thomas A. Buchholz ◽  
Susan L. Tucker ◽  
Henry M. Kuerer ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Regression Analysis ◽  
Neoadjuvant Chemotherapy ◽  
Distant Metastasis ◽  
Cox Regression ◽  
Clinical Stage ◽  
Pathologic Complete Response ◽  
Distant Metastases ◽  
Complete Response ◽  
Cox Regression Analysis

Purpose To identify clinicopathological factors predictive of distant metastasis in patients who had a pathologic complete response (pCR) after neoadjuvant chemotherapy (NC). Methods Retrospective review of 226 patients at our institution identified as having a pCR was performed. Clinical stage at diagnosis was I (2%), II (36%), IIIA (27%), IIIB (23%), and IIIC (12%). Eleven percent of all patients were inflammatory breast cancers (IBC). Ninety-five percent received anthracycline-based chemotherapy; 42% also received taxane-based therapy. The relationship of distant metastasis with clinicopathologic factors was evaluated, and Cox regression analysis was performed to identify independent predictors of development of distant metastasis. Results Median follow-up was 63 months. There were 31 distant metastases. Ten-year distant metastasis-free rate was 82%. Multivariate Cox regression analysis using combined stage revealed that clinical stages IIIB, IIIC, and IBC (hazard ratio [HR], 4.24; 95% CI, 1.96 to 9.18; P < .0001), identification of ≤ 10 lymph nodes (HR, 2.94; 95% CI, 1.40 to 6.15; P = .004), and premenopausal status (HR, 3.08; 95% CI, 1.25 to 7.59; P = .015) predicted for distant metastasis. Freedom from distant metastasis at 10 years was 97% for no factors, 88% for one factor, 77% for two factors, and 31% for three factors (P < .0001). Conclusion A small percentage of breast cancer patients with pCR experience recurrence. We identified factors that independently predicted for distant metastasis development. Our data suggest that premenopausal patients with advanced local disease and suboptimal axillary node evaluation may be candidates for clinical trials to determine whether more aggressive or investigational adjuvant therapy will be of benefit.

scholarly journals Deep learning model for fully automated breast cancer detection system from thermograms

PLoS ONE ◽  
2022 ◽  
Vol 17 (1) ◽  
pp. e0262349
Author(s):  
Esraa A. Mohamed ◽  
Essam A. Rashed ◽  
Tarek Gaber ◽  
Omar Karam
Keyword(s):  
Breast Cancer ◽  
Deep Learning ◽  
Cancer Detection ◽  
Detection System ◽  
Diagnostic Technique ◽  
Learning Model ◽  
The Body ◽  
Breast Cancer Detection ◽  
Benchmark Database ◽  
Deep Learning Model

Breast cancer is one of the most common diseases among women worldwide. It is considered one of the leading causes of death among women. Therefore, early detection is necessary to save lives. Thermography imaging is an effective diagnostic technique which is used for breast cancer detection with the help of infrared technology. In this paper, we propose a fully automatic breast cancer detection system. First, U-Net network is used to automatically extract and isolate the breast area from the rest of the body which behaves as noise during the breast cancer detection model. Second, we propose a two-class deep learning model, which is trained from scratch for the classification of normal and abnormal breast tissues from thermal images. Also, it is used to extract more characteristics from the dataset that is helpful in training the network and improve the efficiency of the classification process. The proposed system is evaluated using real data (A benchmark, database (DMR-IR)) and achieved accuracy = 99.33%, sensitivity = 100% and specificity = 98.67%. The proposed system is expected to be a helpful tool for physicians in clinical use.

Sign in / Sign up

Export Citation Format

Share Document