Breast Cancer Data Research Articles

Delta Radiomics Based on Longitudinal Dual-modal Ultrasound Can Early Predict Response to Neoadjuvant Chemotherapy in Breast Cancer Patients

To develop a monitoring model using radiomics analysis based on longitudinal B-mode ultrasound (BUS) and shear wave elastography (SWE) to early predict pathological response to neoadjuvant chemotherapy (NAC) in breast cancer patients. In this prospective study, 112 breast cancer patients who received NAC between September 2016 and March 2022 were included. The BUS and SWE data of breast cancer were obtained prior to treatment as well as after two and four cycles of NAC. Radiomics features were extracted followed by measuring the changes in radiomics features compared to baseline after the second and fourth cycles of NAC(△R[C2], △R [C4]), respectively. The delta radiomics signatures were established using a support vector machine classifier. The area under receiver operating characteristic curve (AUC) values of △RBUS (C2) and △RBUS (C4) for predicting the response to NAC were 0.83 and 0.84, while those of △RSWE (C2) and △RSWE (C4) were 0.88 and 0.90, respectively. △RSWE exhibited significantly superior performance to △RBUS for predicting NAC response (Delong test, p<0.01). No significant differences were observed in the performances between △R (C2) and △R (C4) based on BUS or SWE data. The longitudinal dual-modal ultrasound radiomics (LDUR) model had an excellent discrimination, good calibration and clinical usefulness, with the AUC, sensitivity and specificity of 0.97, 95.52% and 91.11%, respectively. The LDUR model achieved excellent performance in predicting the pathological response to chemotherapy during the early stages of NAC for breast cancer.

Abstract C002: Integrating multiomics data to identify African ancestry-informative markers acting as eQTLs in breast cancer

Abstract Ancestry-informative markers (AIMs) are SNPs that are found at relatively higher frequencies in a specific population and can be used to estimate genetic ancestry. Genetic ancestry is becoming more commonly used in biomedical research investigating the genetic basis of diseases. For many cancers, the literature suggests that African ancestry is associated with poor response to treatment, poorer prognosis, and poorer survival. More specifically, African American women with breast cancer are more likely to have more aggressive subtypes of the disease and diagnosed with a later stage of the disease at an earlier age compared to their European American counterparts. Some of this disparity remains after accounting for socioeconomic factors, suggesting a genetic basis for the health disparities observed in many cancers. To evaluate the genetic basis of cancer health disparities in populations of African ancestry, we identified a set of 46,787 African ancestry informative markers (AIMs) using 1000 Genomes Project data from African and European populations. These AIMs were functionally annotated and investigated for potential associations with cancer in the literature, using tools such as VEP, SIFT, and Polyphen. Using multiomics data (RNAseq, genotype, methylation, copy number variation) from The Cancer Genome Atlas and an integrative eQTL modeling approach, we are investigating which of these AIMs act as expression quantitative trait loci in breast cancer. Of the 47K SNPs identified, 0.6% percent was found to be in the coding regions of the genome and AIMS located on chromosomes 8, 10, 11, 15, 18, and 19 were found to be associated with different cancers in the literature including breast and colorectal cancers. Approximately 5K of the AIMs were found in the TCGA breast cancer data and preliminary results identified cis-eqtls in 4 different genes including SERINC2 and ZSCAN23, and trans-eqtls in about 8 genes including CLCN3 and DEDD. These eQTL AIMs are being evaluated for potential association with ancestry-specific differential survival and response to treatment in breast cancer. Citation Format: Samantha Boudeau. Integrating multiomics data to identify African ancestry-informative markers acting as eQTLs in breast cancer [abstract]. In: Proceedings of the 16th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2023 Sep 29-Oct 2;Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2023;32(12 Suppl):Abstract nr C002.

Analisis Perbandingan Pengukuran Jarak Algoritma K-Nearest Neighbor Dengan Menggunakan Data Breast Cancer Dan Data Heart Disease

&#x0D; &#x0D; &#x0D; &#x0D; Breast Cancer is a cancerous condition that appears in the breast area. This type of cancer is often experienced by women with a characteristic feature of Breast Cancer, namely the appearance of unusual lumps in the breast area. Heart or Heart Disease is a type of Non-Communicable Disease (PTM): which results in a fairly high mortality rate. Heart Disease is caused by several risk factors including smoking, an unhealthy lifestyle, high cholesterol, hypertension, and diabetes.&#x0D; Based on these facts, an appropriate algorithm is needed to classify Breast Caner and Heart Disease as an effort to prevent an increase in mortality rates due to Breast Cancer and Heart Disease. And the algorithm that will be used is the K-Nearest Neighbor algorithm with 3 distance measurement methods, namely Euclidean distance, Manhattan distance, and Minkowsky distance .&#x0D; &#x0D; &#x0D; &#x0D; &#x0D; From the stages that have been carried out, the final results of the Euclidean distance method obtained an Accuracy value of 80.88% Breast Cancer data at K = 11, and 78.69% heart Disease data at K = 11. The Manhattan distance method obtained an Accuracy value of 89.71% of Breast Cancer data on K=11, and 78.69% of Heart Disease data on K=20.The Minkowsky distance method obtained an Accuracy value of 98.53% of Breast Cancer data on K=11, and 79.41% of Heart Disease data on K=11. This shows that the Minkowsky distance method works more optimally than the Euclidean distance and Manhattan distance methods.

Particle Swarm Optimization for Penalize cox models in long-term prediction of breast cancer data

The particle swarm optimization algorithm (PSO) was applied to penalize the Cox model for predicting long-term outcomes of breast cancer patients. This study makes use of data on 198 patients' breast cancer survival, including their age, estrogen receptor status, tumor size, and grade, as well as the expression levels of 76 genes. The objective was to identify a subset of features that could accurately predict patient survival while controlling for overfitting and model complexity. PSO was used to search for optimal model parameters. The algorithm was designed to minimize a penalized partial likelihood function, which balances the tradeoff between an accurate model and model complexity. The values of the objective function were compared with other feature selection techniques, including the Least Absolute Shrinkage and Selection Operator (LASSO) and Elastic regression, and were found to perform better in relation to predictive accuracy and feature selection. The study results showed that PSO-based Cox models with cross-validation to regularization parameters had higher prediction accuracy than models trained with other feature selection methods. The PSO algorithm identified a subset of features that were consistently selected across multiple iterations, indicating their importance in predicting patient survival. Overall, the study demonstrates the potential of PSO-based feature selection in improving the accuracy and interpretability of Cox regression models for predicting long-term outcomes in breast cancer patients.

LncRNA SEMA3B-AS1 suppresses the tumor-initiating characteristics of triple negative breast cancer via engaging in MLL4-mediated H3K4 trimethylation.

Long noncoding RNAs (lncRNAs) are crucial regulators of tumor-initiating cells (TICs) and hold particular importance in triple negative breast cancer (TNBC). Yet, the precise mechanisms by which TIC-associated lncRNAs influence TNBC remain unclear. Our research utilized The Cancer Genome Atlas Breast Cancer (BC)data set to identify prognostic lncRNAs. We then conducted extensive assays to explore their impact on the tumor-initiating phenotype of TNBC cells and the underlying mechanisms. Notably, we found that low expression of lncRNA SEMA3B-AS1 correlated with unfavorable survival in BC patients. SEMA3B-AS1 was also downregulated in TNBC and linked to advanced tumor stage. Functional experiments confirmed its role as a TIC-suppressing lncRNA, curtailing mammosphere formation, ALDH + TIC cell proportion, and impairing clonogenicity, migration, and invasion. Mechanistic insights unveiled SEMA3B-AS1's nuclear localization and interaction with MLL4 (mixed-lineage leukemia 4), triggering H3K4 methylation-associated transcript activation and thus elevating the expression of SEMA3B, a recognized tumor suppressor gene. Our findings emphasize SEMA3B-AS1's significance as a TNBC-suppressing lncRNA that modulates TIC behavior. This study advances our comprehension of lncRNA's role in TNBC progression, advocating for their potential as therapeutic targets in this aggressive BC subtype.

Cardiac protective techniques in left breast radiotherapy: rapid selection criteria for routine clinical decision making

ObjectiveIn left breast radiotherapy (RT) desired heart doses may be achieved without heart-sparing RT techniques in some patients. We aimed to examine the existence of predictive factors and cutoff points to determine which patients are the main candidates for heart-sparing RT techniques.Material and methodDosimetric data for left breast cancer was examined. RT plans were made at conventional doses to the breast and peripheral lymph nodes. Statistical analyses were performed using SPSS 22.0 (SPSS Inc., IBM Corp., Armonk, NY).Result114 cases were evaluated by ROC (Receiver operating characteristic) analysis in the breast-conserving surgery (BCS) and mastectomy groups. While only left lung volume (AUC: 0.74, 95% CI 0.61–0.87, p = 0.002) was significant in BCS cases, in cases with mastectomy, left lung volume (AUC: 0.81, 95% CI 0.69–0.94, p = 0.002) and lung/heart volume ratio (AUC: 0.83, 95% CI 0.70–0.96, p = 0.001) had a significant relationship with the relevance of heart doses. The cutoff point of 1.92 was selected for the lung/heart volume ratio for the mastectomized patients. Moreover, the cutoff point 1154 cc and 1208 cc was determined for the left lung volume for the BCS and mastectomized patients, respectively.ConclusionVarious cutoff points in left breast RT can be used to predict whether RT plans will meet QUANTEC (Quantitative Analysis of Normal Tissue Effects in the Clinic) heart dose limits. Evaluating only these few cutoff points before planning makes it possible to eliminate 70% of patients with BCS and 40% of patients with mastectomy from respiratory-controlled methods, which require time and effort. Patients with lung volume and lung/heart volume ratio smaller than the cutoff values can be considered primary candidates for heart-sparing techniques.

ScLSTM: single-cell type detection by siamese recurrent network and hierarchical clustering

MotivationCategorizing cells into distinct types can shed light on biological tissue functions and interactions, and uncover specific mechanisms under pathological conditions. Since gene expression throughout a population of cells is averaged out by conventional sequencing techniques, it is challenging to distinguish between different cell types. The accumulation of single-cell RNA sequencing (scRNA-seq) data provides the foundation for a more precise classification of cell types. It is crucial building a high-accuracy clustering approach to categorize cell types since the imbalance of cell types and differences in the distribution of scRNA-seq data affect single-cell clustering and visualization outcomes.ResultTo achieve single-cell type detection, we propose a meta-learning-based single-cell clustering model called ScLSTM. Specifically, ScLSTM transforms the single-cell type detection problem into a hierarchical classification problem based on feature extraction by the siamese long-short term memory (LSTM) network. The similarity matrix derived from the improved sigmoid kernel is mapped to the siamese LSTM feature space to analyze the differences between cells. ScLSTM demonstrated superior classification performance on 8 scRNA-seq data sets of different platforms, species, and tissues. Further quantitative analysis and visualization of the human breast cancer data set validated the superiority and capability of ScLSTM in recognizing cell types.

The impact of risk stratification by polygenic risk and age on breast cancer screening in women aged 40–49 years: a modelling study

BackgroundPolygenic Risk Scores (PRSs) have been proposed as a mechanism for risk-stratification of screening, increasing efficiency and enabling extension of existing programmes to improve survival in our aging population. We sought to model the impact of three hypothetical programmes of annual breast cancer screening in women aged 40–49 years: screening the PRS-defined high-risk quintile, screening the oldest quintile, and screening the full population. MethodsIn this UK-based modelling study, we used the published estimate of the area under the curve (AUC) of a currently available breast cancer PRS (0·64) to calculate the proportion of cancers captured by the PRS-defined high-risk quintile. We used population size estimates from the Office for National Statistics alongside age-stratified incidence rates of breast cancer, and age or stage-specific survival data from the National Cancer Registry, to build our model. We used stage-specific route-to-diagnosis data to reassign stage-specific survival for screen-detected cancers. Ethics approval was not required. FindingsThe PRS-defined high-risk quintile, oldest quintile, and full population capture 37% (n=2811), 29% (n=2198), and 100% (n=7533) of breast cancers occurring in women aged 40–49 each year. Annual screening of each group using digital mammography (sensitivity 70%, specificity 92%) would identify 1968, 1538, and 5273 breast cancers per year, respectively. This corresponds to an improvement in survival of 1·4% (102 deaths averted), 1·1% (80 deaths averted) and 3·6% (274 deaths averted) compared with baseline (no screening). Full population screening would require 4 369 703 mammograms and 354 246 confirmatory tests (breast biopsies) every year, while screening the oldest quintile would require 937 850 mammograms and 76 390 biopsies. Screening the PRS-defined high-risk quintile would require 873 941 mammograms and 71 658 biopsies, in addition to a PRS for all women in the age group (4 369 703). InterpretationUnder favourable assumptions, stratifying screening by PRS rather than age results in modest gains in survival but increases overdiagnoses, logistical complexity, and economic costs. Our study is limited by our modelling parameters (anticipated to maximise survival estimates), including complete uptake of PRS profiling and cancer screening, no interval cancers, and application of screening tools superior to those currently available in the UK. Only with randomised controlled trials, can the uptake, clinical impact, costs, and harms of PRS-stratified screening be definitively assessed. FundingThe Wellcome Trust.

Backward Elimination for Feature Selection on Breast Cancer Classification Using Logistic Regression and Support Vector Machine Algorithms

Breast cancer is a prevalent form of cancer that afflicts women across all nations globally. One of the ways that can be done as a prevention to reduce elevated fatality due to breast cancer is with a detection system that can determine whether a cancer is benign or malignant. Logistic Regression and Support Vector Machine (SVM) classification algorithms are often used to detect this disease, but the use of these two algorithms often doesn’t give optimal results when applied to datasets that have many features, so additional algorithm is needed to improve classification performance by using Backward Elimination feature selection. The comparison of Logistic Regression and SVM algorithms was carried out by applying feature selection to breast cancer data to see the best model. The breast cancer dataset has 30 features and two classes, Benign and Malignant. Backward Elimination has reduced features from 30 features to 13 features, thereby increasing the performance of both classification models. The best classification was obtained by using the Backward Elimination feature selection and linear kernel SVM with an increase in accuracy value from 96.14% to 97.02%, precision from 98.06% to 99.49%, recall from 90.48% to 92.38%, and the AUC from 0.95 to 0.96.

Identification of cuproptosis-related gene SLC31A1 and upstream LncRNA-miRNA regulatory axis in breast cancer

Mounting evidence indicate that cuproptosis, a novel form of programmed cell death, contributes to cancer development and progression. However, a comprehensive analysis regarding the expressions, functions, and regulatory network of cuproptosis-related genes is still lacking. In the present work, cuproptosis-related genes, upstream miRNAs and lncRNAs, and clinical data of breast cancer from TCGA database were analyzed by R language including Cox regression analysis, correlation calculation, ROC curve construction, and survival evaluation, and were further verified by public-available databases. Chemosensitivity and immune infiltration were also evaluated by online tools. SLC31A1 was significantly increased in breast cancer samples than those in normal tissues. SLC31A1 was negatively related to a favorable outcome in breast cancer, and the AUC value increased with the prolongation of follow-up time. LINC01614 and miR-204-5p were potential upstream regulators of SLC31A1. Moreover, SLC31A1 was significantly positively correlated with different immune cells infiltration, immune cell biomarkers, and immune checkpoints in breast cancer. SLC31A1 was a potential cuproptosis-related gene in breast cancer, which was significantly upregulated and was able to predict diagnosis, prognosis, chemosensitivity, and immune infiltration. LINC01640/miR-204-5p/SLC31A1 might be a significant and promising axis during cuproptosis in breast cancer.

The association between thyroid and breast cancers: a bidirectional mendelian randomization study.

Thyroid and breast cancers are the two most frequent types of endocrine-related tumors among women worldwide, and their incidence is still on the rise. Observational studies have shown a relationship between thyroid and breast cancers. Nevertheless, many confounders predispose the results to interference. Accordingly, we performed a two-sample Mendelian randomization (MR) study to investigate the causal association between thyroid and breast cancers. We acquired breast cancer data from the UK Biobank (13,879 breast cancer cases and 198,523 controls) and the Breast Cancer Association Consortium (BCAC; 122,977 breast cancer cases and 105,974 controls), and thyroid cancer data from FinnGen Biobank (989 thyroid cancer and 217,803 controls). Then, the multiplicative random effects inverse variance weighting (IVW), weight median (WM), and MR Egger methods were executed for MR analysis. Overall, IVW showed a causal effect of breast cancer on thyroid cancer using the BCAC dataset (odds ratio [OR] = 1.17; 95% confidence interval [CI] = 1.036-1.322; P = 0.011), and this relationship was also supported by the UK Biobank dataset (OR = 23.899; 95% CI = 2.331-245.003; P = 0.007), which showed that breast cancer patients were more likely to be diagnosed with thyroid cancer. On the whole, the reverse MR analysis did not show a causal effect of breast cancer on thyroid cancer. However, IVW showed a causal effect of thyroid cancer on estrogen receptor -negative breast cancer using the BCAC dataset (OR = 1.019; 95% CI = 1.001-1.038; P = 0.043), which suggested that people with thyroid cancer were more likely to develop breast cancer. Breast cancer represents a possible risk factor for thyroid cancer and thyroid cancer also represents a possible risk factor for ER-negative breast cancer. Future studies using powerful genetic tools to determine the causal relationship between breast and thyroid cancers are required.

The prevalence and clinical significance of HER2 expression in prostate adenocarcinoma

AimsAbnormalities in HER2 are well-established oncogenic drivers and are approved therapeutic targets in various malignancies. Prior studies evaluating HER2 expression in prostate cancer (PCa) have yielded variable results. Most of these studies used immunohistochemistry scoring systems based on breast cancer data. The goal of this study was to determine the prevalence and clinical significance of HER2 expression using a scoring system that better reflects the HER2 staining pattern observed in PCa. MethodsWe randomly selected similar numbers of localized low risk (AJCC stage I), locally advanced (AJCC stages II & III), and metastatic (AJCC stage IV) PCa patients treated at the DC VA Medical Center between 2000 and 2022. Among them, we included patients who had sufficient PCa tissue samples and clinical information required for this analysis. Two experienced pathologists independently scored HER2 expression (Ventana Pathway anti-HER2) according to a modified gastric cancer HER2 scoring system. ResultsOut of the 231 patients included, 85 % self-identified as Black. 58.9 % of patients expressed HER2 (1+: 35.5 %; 2+: 18.2 %; 3+: 5.2 %). Validity of the results was confirmed using the HercepTest antibody. Higher HER2 expression was associated with a higher Gleason Score/Grade Group and advanced disease. ConclusionsOur findings support the adverse prognostic impact on HER2 in PCa. We propose the use of a modified scoring system to evaluate HER2 expression in PCa. The observed high prevalence of HER2 expression supports the consideration of novel HER2-targeted therapies addressing even low levels of HER2 expression in future PCa trials.

Integrative bioinformatics analysis of miRNA and mRNA expression profiles identified some potential biomarkers for breast cancer

BackgroundBreast cancer is a common cause of cancer death among women with a complex and heterogeneous picture in histological, molecular and clinical features. The aim of this study was to identify hub gene and their target microRNAs in related pathways for breast cancer.MethodsWe selected screening methods for differentially expressed mRNAs and miRNAs using expression profile data of breast cancer from the cancer genome atlas. Using some databases for annotation, the functional and pathway enrichment for differential expression genes was performed. We selected genes and miRNAs with differential expression pattern. Then we determined target genes for differential expression miRNAs (DEMIs) and intersection between them was selected as differentially expressed miRNA–target genes for breast cancer. In the next step, we constructed miRNA–mRNA regulatory network and protein–protein interaction (PPI) network for more information.ResultsTop 10 DEMIs were identified from miRNA profile. Then, we selected 354 genes as target gene for 10 DEMIs. The miRNA–mRNA and PPI network were constructed, and 10 hub genes and 5 miRNAs identified that some of them are new for breast cancer. Also, miRNA–target genes with differential expressions in this study were all mainly involved in signaling pathways and developmental process.ConclusionThis study identified some candidate biomarkers for breast cancer that they have a potential role in pathways related to breast. These findings can be used for research, early diagnosis and therapeutic goals.

Analysing Breast Cancer using Convolution Neural Network

Technological development of Soft Computing and Artificial Intelligence contributed a lot towards disease diagnosis in medical science. For providing solutions to the biological inspired problems in medical domain like Breast Cancer (BC) soft computing methods can give the flexible information. As per 2020 about 2 million women were detected with Breast Cancer (BC) creating most common malignancy among women worldwide. This rises both incidence and mortality has occurred during the past three decades due to evolving risk factors, improved cancer registries, and earlier diagnosis. There is a large number of risk factors for BC, some of which can be changed and others cannot. Eighty percent of people diagnosed with BC nowadays are over the age of fifty. Molecular subtype and developmental stage are both important in determining the likelihood of survival. When it comes to clinical presentation, behaviour and shape, invasive BC span a broad spectrum of tumours. In this paper, Convolution Neural Network (CNN) used to recognize the BC tumor because it is another sort of neural network that can discover key information in both image and time series data. By applying CNN on the 2023 RSNA (Radiological Society for North America) Screening Mammography Breast Cancer data we analysed how best CNN algorithm is for identifying breast cancer with accuracy and also tried to analyse at what age Breast Cancer is mostly occurred in women.

Discovering SNP-disease relationships in genome-wide SNP data using an improved harmony search based on SNP locus and genetic inheritance patterns.

Advances in high-throughput sequencing technologies have made it possible to access millions of measurements from thousands of people. Single nucleotide polymorphisms (SNPs), the most common type of mutation in the human genome, have been shown to play a significant role in the development of complex and multifactorial diseases. However, studying the synergistic interactions between different SNPs in explaining multifactorial diseases is challenging due to the high dimensionality of the data and methodological complexities. Existing solutions often use a multi-objective approach based on metaheuristic optimization algorithms such as harmony search. However, previous studies have shown that using a multi-objective approach is not sufficient to address complex disease models with no or low marginal effect. In this research, we introduce a locus-driven harmony search (LDHS), an improved harmony search algorithm that focuses on using SNP locus information and genetic inheritance patterns to initialize harmony memories. The proposed method integrates biological knowledge to improve harmony memory initialization by adding SNP combinations that are likely candidates for interaction and disease causation. Using a SNP grouping process, LDHS generates harmonies that include SNPs with a higher potential for interaction, resulting in greater power in detecting disease-causing SNP combinations. The performance of the proposed algorithm was evaluated on 200 synthesized datasets for disease models with and without marginal effect. The results show significant improvement in the power of the algorithm to find disease-related SNP sets while decreasing computational cost compared to state-of-the-art algorithms. The proposed algorithm also demonstrated notable performance on real breast cancer data, showing that integrating prior knowledge can significantly improve the process of detecting disease-related SNPs in both real and synthesized data.

IoSearch: An approach for identifying interacting multiomics biomarkers using a novel algorithm with application on breast cancer data sets.

Identification of biomarkers by integrating multiple omics together is important because complex diseases occur due to an intricate interplay of various genetic materials. Traditional single-omics association tests neither explore this crucial interomics dependence nor identify moderately weak signals due to the multiple-testing burden. Conversely, multiomics data integration imparts complementary information but suffers from an increased multiple-testing burden, data diversity inherent with different omics features, high-dimensionality, and so forth. Most of the available methods address subtype classification using dimension-reduction techniques to circumvent the sample size issue but interacting multiomics biomarker identification methods are unavailable. We propose a two-step model that first investigates phenotype-omics association using logistic regression. Then, selects disease-associated omics using sparse principal components which explores the interrelationship of multiple variables from two omics in a multivariate multiple regression framework. On the basis of this model, we developed a multiomics biomarker identification algorithm, interacting omics search (ioSearch), that jointly tests the effect of multiple omics with disease and between-omics associations by using pathway information that subsequently reduces the multiple-testing burden. Further, inference in terms of p values potentially makes it an easily interpretable biomarker identification tool. Extensive simulation demonstrates ioSearch as statistically powerful with a controlled Type-I error rate. Its application to publicly available breast cancer data sets identified relevant omics features in important pathways.

THU486 Identification Of Metabolic And Molecular Mechanisms Driving Estrogen Receptor Positive Breast Cancer Disparities Using A Machine-Learning Pipeline

Abstract Disclosure: A. Santaliz Casiano: None. D. Mehta: None. H. Patel: None. G. Rauscher: None. J. Kim: None. J.M. Frasor: None. K. Hoskins: None. Background: African American (AA) women in the United States have a 40% higher breast cancer mortality rate compared with Non-Hispanic White (NHW) women. The survival disparity is particularly striking among ER+ breast cancer cases. The purpose of this study is to examine whether there are racial differences in metabolic pathways typically activated in patients with ER+ positive breast cancer. Methods: We collected pretreatment plasma from AA and NHW ER+ breast cancer cases and cancer-free controls to conduct an untargeted metabolomics analysis using gas chromatography mass spectrometry (GC-MS) to identify metabolites that may be altered in the different racial groups. Statistical methods combined with multiple feature selection and prediction models were employed to identify race-specific altered metabolic signatures. This was followed by the identification of altered metabolic pathways with a focus in AA patients with breast cancer. The clinical relevance of the identified pathways was further examined in PanCancer Atlas breast cancer data set from TCGA. Results: We identified differential metabolic signatures between NHW and AA patients. In AA patients, we observed changes in metabolites associated with amino acid metabolism, while fatty acid metabolism was significantly enriched in NHW patients. By mapping these metabolites to potential epigenetic regulatory mechanisms, this study identified significant association with regulators of metabolism such as methionine adenosyltransferase 1A (MAT1A), DNA Methyltransferases and Histone methyltransferases for AA individuals, and Fatty acid Synthase (FASN) and Monoacylglycerol lipase (MGL) for NHW individuals. Specific gene NELFE with histone methyltransferase activity was associated with poor survival exclusively for AA individuals. Conclusion: We employed a comprehensive and novel approach that integrates multiple machine learning methods and statistical methods, coupled with human functional pathway analyses. The metabolic profile of plasma samples identified may help elucidate underlying molecular drivers of disproportionately aggressive ER+ tumor biology in AA women, and may ultimately lead to identification of novel therapeutic targets. To our knowledge, this is a novel finding that describes a link between metabolic alterations and epigenetic regulation in AA breast cancer, and underscores the need for detailed investigations into the biological underpinnings of breast cancer health disparities. Presentation: Thursday, June 15, 2023

Detection of influential observations in high-dimensional survival data

Survival analysis is a statistical technique mainly used to analyze time-to-event data. Identification of influential observation attains greater importance since it leads to discovering new prognostic factors. Influential observation in survival typically points to individuals whose survival time is extremely short or long in comparison to others. Particularly, when the data possess more covariates than the observations, all classical approaches fail to perform. Hence, dimensionality reduction is necessary for choosing appropriate variables and it has been done by popular techniques such as LASSO and elastic net algorithm. This paper consider high-dimensional breast cancer data, and its dimensionality is reduced using variable selection methods. Subsequently, the rank product test and martingale residuals are used to identify an influential observation. Furthermore, a resampling technique is used to validate the consistency and robustness of the methods. The novelty of this paper lies in comparing the prediction accuracy of datasets with and without outliers using Random Survival Forest (RSF) for different training fractions. Comparatively, the RSF result demonstrates that the LASSO approach outperform others in the absence of outliers. Therefore, we suggest reducing dimensionality using the LASSO variable selection technique first, followed by removing likely outliers to improve the performance of classification algorithms.

A novel prognostic signature and potential therapeutic drugs based on tumor immune microenvironment characterization in breast cancer

Tumor microenvironment (TME) is closely correlated to the occurrence and progression of breast cancer, however its potentiality in assisting diagnosis and therapeutic decision remains unclear. Therefore, the major aim of this study is to explore the prognostic value of TME related gene in breast cancer. Expression matrices and clinical data of breast cancer obtained from public databases were divided into TME relevant clusters according to immune characterization. A 12-gene molecular classifier was generated through the utilization of differentially expressed genes identified between distinct Tumor Microenvironment (TME) clusters, coupled with correlative regression analysis. The performance of this TME-driven prognostic signature (TPS) were examined across both the training and validation cohorts. Furthermore, our study revealed that breast cancer cases classified as high-risk based on the TPS exhibited the phenotype with elevated immune cell infiltration, higher tumor mutational burden, and a notably worse overall prognostic outcome. To conclude, the novel TME-based TPS was able to serve as a superior prognosis indicator for breast cancer, alone or jointly with other clinical factors. Also, breast cancer patients belong to different risk subgroups of TPS were found potentially suitable for distinguished therapeutic agents, which might improve personalized treatment for breast cancer in the future.

158P In silico analysis of HER2 enriched subtype and a HER2 index based on transcriptomic data of breast cancer compared to gastric and uterine serous carcinomas

158P In silico analysis of HER2 enriched subtype and a HER2 index based on transcriptomic data of breast cancer compared to gastric and uterine serous carcinomas