Gene-Expression-Based Predictors for Breast Cancer

Development of a Machine Learning-Based Prognostic Model for Hormone Receptor-Positive Breast Cancer Using Nine-Gene Expression Signature

Background

Determining the prognosis of hormone receptor positive (HR+) breast cancer (BC), which accounts for 80% of all BCs, is critical in improving survival outcomes. Stratifying individuals at high risk of BC-related mortality and improving prognosis has been the focus of research for over a decade. However, these tools are not universal as they are limited to clinical factors. We hypothesized that a new framework for predicting prognosis in HR+ BC patients can develop using artificial intelligence.

Methods

A total of 2,338 HR+ human epidermal growth factor receptor 2 negative (HER2-) BC cases were analyzed from Molecular Taxonomy of Breast Cancer International Consortium (METABRIC), The Cancer Genome Atlas (TCGA), and Gene Expression Omnibus (GEO) cohorts. Groups were then divided into high- and low-risk categories utilizing a recurrence prediction model (RPM). An RPM was created by extracting nine prognosis-related genes from over 18,000 genes using a logistic progression model.

Results

Risk classification by RPM was significantly stratified in both the discovery cohort and validation cohort. In the time-dependent area under the curve analysis, there was some variation depending on the cohort, but accuracy was found to decline significantly after about 10 years. Cell cycle related gene sets, MYC, and PI3K-AKT-mTOR signaling were enriched in high-risk tumors by the Gene Set Enrichment Analysis. High-risk tumors were associated with high levels of immune cells from the lymphoid and myeloid lineage and immune cytolytic activity, as well as low levels of stem cells and stromal cells. High-risk tumors were also associated with poor therapeutic effects of chemotherapy and endocrine therapy.

Conclusions

This model was able to stratify prognosis in multiple cohorts. This is because the model reflects major BC therapeutic target pathways and tumor immune microenvironment and, further is supported by the therapeutic effect of chemotherapy and endocrine therapy.

The Interleukin-1 (IL-1) Superfamily Cytokines and Their Single Nucleotide Polymorphisms (SNPs)

Interleukins (ILs)-which are important members of cytokines-consist of a vast group of molecules, including a wide range of immune mediators that contribute to the immunological responses of many cells and tissues. ILs are immune-glycoproteins, which directly contribute to the growth, activation, adhesion, differentiation, migration, proliferation, and maturation of immune cells; and subsequently, they are involved in the pro and anti-inflammatory responses of the body, by their interaction with a wide range of receptors. Due to the importance of immune system in different organisms, the genes belonging to immune elements, such as ILs, have been studied vigorously. The results of recent investigations showed that the genes pertaining to the immune system undergo progressive evolution with a constant rate. The occurrence of any mutation or polymorphism in IL genes may result in substantial changes in their biology and function and may be associated with a wide range of diseases and disorders. Among these abnormalities, single nucleotide polymorphisms (SNPs) can represent as important disruptive factors. The present review aims at concisely summarizing the current knowledge available on the occurrence, properties, role, and biological consequences of SNPs within the IL-1 family members.

Artificial image objects for classification of breast cancer biomarkers with transcriptome sequencing data and convolutional neural network algorithms

BACKGROUND

Transcriptome sequencing has been broadly available in clinical studies. However, it remains a challenge to utilize these data effectively for clinical applications due to the high dimension of the data and the highly correlated expression between individual genes.

METHODS

We proposed a method to transform RNA sequencing data into artificial image objects (AIOs) and applied convolutional neural network (CNN) algorithms to classify these AIOs. With the AIO technique, we considered each gene as a pixel in an image and its expression level as pixel intensity. Using the GSE96058 (n = 2976), GSE81538 (n = 405), and GSE163882 (n = 222) datasets, we created AIOs for the subjects and designed CNN models to classify biomarker Ki67 and Nottingham histologic grade (NHG).

RESULTS

With fivefold cross-validation, we accomplished a classification accuracy and AUC of 0.821 ± 0.023 and 0.891 ± 0.021 for Ki67 status. For NHG, the weighted average of categorical accuracy was 0.820 ± 0.012, and the weighted average of AUC was 0.931 ± 0.006. With GSE96058 as training data and GSE81538 as testing data, the accuracy and AUC for Ki67 were 0.826 ± 0.037 and 0.883 ± 0.016, and that for NHG were 0.764 ± 0.052 and 0.882 ± 0.012, respectively. These results were 10% better than the results reported in the original studies. For Ki67, the calls generated from our models had a better power for prediction of survival as compared to the calls from trained pathologists in survival analyses.

CONCLUSIONS

We demonstrated that RNA sequencing data could be transformed into AIOs and be used to classify Ki67 status and NHG with CNN algorithms. The AIO method could handle high-dimensional data with highly correlated variables, and there was no need for variable selection. With the AIO technique, a data-driven, consistent, and automation-ready model could be developed to classify biomarkers with RNA sequencing data and provide more efficient care for cancer patients.

KMT2C is a potential biomarker of prognosis and chemotherapy sensitivity in breast cancer

PURPOSE

Epigenetic regulation plays critical roles in cancer progression, and high-frequency mutations or expression variations in epigenetic regulators have been frequently observed in tumorigenesis, serving as biomarkers and targets for cancer therapy. Here, we aimed to explore the function of epigenetic regulators in breast cancer.

METHODS

The mutational landscape of epigenetic regulators in breast cancer samples was investigated based on datasets from the Cancer Genome Atlas. The Kaplan-Meier method was used for survival analysis. RNA sequencing (RNA-seq) in MCF-7 cells transfected with control siRNA or KMT2C siRNA was performed. Quantitative reverse transcription-PCR and chromatin immunoprecipitation were used to validate the RNA-seq results.

RESULTS

Among the 450 epigenetic regulators, KMT2C was frequently mutated in breast cancer samples. The tumor mutational burden (TMB) was elevated in breast cancer samples with KMT2C mutations or low KMT2C mRNA levels compared to their counterparts with wild-type KMT2C or high KMT2C mRNA levels. Somatic mutation and low expression of KMT2C were independently correlated with the poor overall survival (OS) and disease-free survival (DFS) of the breast cancer samples, respectively. RNA-seq analysis combined with chromatin immunoprecipitation and qRT-PCR assays revealed that the depletion of KMT2C remarkably affected the expression of DNA damage repair-related genes. More importantly, the low expression of KMT2C was related to breast cancer cell sensitivity to chemotherapy and longer OS of breast cancer patients who underwent chemotherapy.

CONCLUSION

We conclude that KMT2C could serve as a potential biomarker of prognosis and chemotherapy sensitivity by affecting the DNA damage repair-related genes of breast cancer.

Different associations of tumor PIK3CA mutations and clinical outcomes according to aspirin use among women with metastatic hormone receptor positive breast cancer

Introduction

The relationships among PIK3CA mutations, medication use and tumor progression remains poorly understood. Aspirin use post-diagnosis may modify components of the PI3K pathway, including AKT and mTOR, and has been associated with lower risk of breast cancer recurrence and mortality. We assessed time to metastasis (TTM) and survival with respect to aspirin use and tumor PIK3CA mutations among women with metastatic breast cancer.

Methods

Patients with hormone receptor positive, HER2 negative (HR+/HER2-) metastatic breast cancer treated in 2009–2016 who received tumor genotyping were included. Aspirin use between primary and metastatic diagnosis was extracted from electronic medical records. TTM and survival were estimated using Cox proportional hazards regression.

Results

Among 267 women with metastatic breast cancer, women with PIK3CA mutated tumors had longer TTM than women with PIK3CA wildtype tumors (7.1 vs. 4.7 years, p = 0.008). There was a significant interaction between PIK3CA mutations and aspirin use on TTM (p = 0.006) and survival (p = 0.026). PIK3CA mutations were associated with longer TTM among aspirin non-users (HR = 0.60 95% CI:0.44–0.82 p = 0.001) but not among aspirin users (HR = 1.57 0.86–2.84 p = 0.139). Similarly, PIK3CA mutations were associated with reduced mortality among aspirin non-users (HR = 0.70 95% CI:0.48–1.02 p = 0.066) but not among aspirin users (HR = 1.75 95% CI:0.88–3.49 p = 0.110).

Conclusions

Among women who develop metastatic breast cancer, tumor PIK3CA mutations are associated with slower time to progression and mortality only among aspirin non-users. Larger studies are needed to confirm this finding and examine the relationship among aspirin use, tumor mutation profile, and the overall risk of breast cancer progression.

'Omics Approaches to Explore the Breast Cancer Landscape

Breast cancer incidence is increasing worldwide with more than 600,000 deaths reported in 2018 alone. In current practice treatment options for breast cancer patients consists of surgery, chemotherapy, radiotherapy or targeting of classical markers of breast cancer subtype: estrogen receptor (ER) and HER2. However, these treatments fail to prevent recurrence and metastasis. Improved understanding of breast cancer and metastasis biology will help uncover novel biomarkers and therapeutic opportunities to improve patient stratification and treatment. We will first provide an overview of current methods and models used to study breast cancer biology, focusing on 2D and 3D cell culture, including organoids, and on in vivo models such as the MMTV mouse model and patient-derived xenografts (PDX). Next, genomic, transcriptomic, and proteomic approaches and their integration will be considered in the context of breast cancer susceptibility, breast cancer drivers, and therapeutic response and resistance to treatment. Finally, we will discuss how 'Omics datasets in combination with traditional breast cancer models are useful for generating insights into breast cancer biology, for suggesting individual treatments in precision oncology, and for creating data repositories to undergo further meta-analysis. System biology has the potential to catalyze the next great leap forward in treatment options for breast cancer patients.

Development of Malignancy-Risk Gene Signature Assay for Predicting Breast Cancer Risk

BACKGROUND

Breast cancer (BC) risk assessment models are statistical estimates based on patient characteristics. We developed a gene expression assay to assess BC risk using benign breast biopsy tissue.

METHODS

A NanoString-based malignancy risk (MR) gene signature was validated for formalin-fixed paraffin-embedded (FFPE) tissue. It was applied to FFPE benign and BC specimens obtained from women who underwent breast biopsy, some of whom developed BC during follow-up to evaluate diagnostic capability of the MR signature. BC risk was calculated with MR score, Gail risk score, and both tests combined. Logistic regression and receiver operating characteristic curves were used to evaluate these 3 models.

RESULTS

NanoString MR demonstrated concordance between fresh frozen and FFPE malignant samples (r = 0.99). Within the validation set, 563 women with benign breast biopsies from 2007 to 2011 were identified and followed for at least 5 y; 50 women developed BC (affected) within 5 y from biopsy. Three groups were compared: benign tissue from unaffected and affected patients and malignant tissue from affected patients. Kruskal-Wallis test suggested difference between the groups (P = 0.09) with trend in higher predicted MR score for benign tissue from affected patients before development of BC. Neither the MR signature nor Gail risk score were statistically different between affected and unaffected patients; combining both tests demonstrated best predictive value (AUC = 0.71).

CONCLUSIONS

FFPE gene expression assays can be used to develop a predictive test for BC. Further investigation of the combined MR signature and Gail Model is required. Our assay was limited by scant cellularity of archived breast tissue.

A 23 gene-based molecular prognostic score precisely predicts overall survival of breast cancer patients

BACKGROUND

Although many prognosis-predicting molecular scores for breast cancer have been developed, they are applicable to only limited disease subtypes. We aimed to develop a novel prognostic score that is applicable to a wider range of breast cancer patients.

METHODS

We initially examined The Cancer Genome Atlas breast cancer cohort to identify potential prognosis-related genes. We then performed a meta-analysis of 36 international breast cancer cohorts to validate such genes. We trained artificial intelligence models (random forest and neural network) to predict prognosis precisely, and we finally validated our prediction with the log-rank test.

FINDINGS

We identified a comprehensive list of 184 prognosis-related genes, most of which have been not extensively studied to date. We then established a universal molecular prognostic score (mPS) that relies on the expression status of only 23 of these genes. The mPS system is almost universally applicable to breast cancer patients (log-rank P < 0.05) in a manner independent of platform (microarray or RNA sequencing).

INTERPRETATION

The mPS system is simple and cost-effective to apply and yet is able to reveal previously unrecognized heterogeneity among patient subpopulations in a platform-independent manner. The combination of mPS and clinical stage stratifies prognosis even more precisely and should prove of value for avoidance of overtreatment. In addition, the prognosis-related genes uncovered in this study are potential drug targets. FUND: This work was supported by KAKENHI grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan to H.S. (19K20403) and to K.I·N (18H05215).

SALMON: Survival Analysis Learning With Multi-Omics Neural Networks on Breast Cancer

Improved cancer prognosis is a central goal for precision health medicine. Though many models can predict differential survival from data, there is a strong need for sophisticated algorithms that can aggregate and filter relevant predictors from increasingly complex data inputs. In turn, these models should provide deeper insight into which types of data are most relevant to improve prognosis. Deep Learning-based neural networks offer a potential solution for both problems because they are highly flexible and account for data complexity in a non-linear fashion. In this study, we implement Deep Learning-based networks to determine how gene expression data predicts Cox regression survival in breast cancer. We accomplish this through an algorithm called SALMON (Survival Analysis Learning with Multi-Omics Neural Networks), which aggregates and simplifies gene expression data and cancer biomarkers to enable prognosis prediction. The results revealed improved performance when more omics data were used in model construction. Rather than use raw gene expression values as model inputs, we innovatively use eigengene modules from the result of gene co-expression network analysis. The corresponding high impact co-expression modules and other omics data are identified by feature selection technique, then examined by conducting enrichment analysis and exploiting biological functions, escalated the interpretation of input feature from gene level to co-expression modules level. Our study shows the feasibility of discovering breast cancer related co-expression modules, sketch a blueprint of future endeavors on Deep Learning-based survival analysis. SALMON source code is available at https://github.com/huangzhii/SALMON/.

Strong impact of MammaPrint and BluePrint on treatment decisions in luminal early breast cancer: results of the WSG-PRIMe study

PURPOSE

The WSG-PRIMe Study prospectively evaluated the impact of the 70-gene signature MammaPrint® (MP) and the 80-gene molecular subtyping assay BluePrint® on clinical therapy decisions in luminal early breast cancer.

METHODS

452 hormone receptor (HR)-positive and HER2-negative patients were recruited (N0, N1). Physicians provided initial therapy recommendations based on clinicopathological factors. After prospective risk classification by MammaPrint/BluePrint was revealed, post-test treatment recommendations and actual treatment were recorded. Decisional Conflict and anxiety were measured by questionnaires.

RESULTS

Post-test switch (in chemotherapy (CT) recommendation) occurred in 29.1% of cases. Overall, physician adherence to MP risk assessment was 92.3% for low-risk and 94.3% for high-risk MP scores. Adherence was remarkably high in "discordant" groups: 74.7% of physicians initially recommending CT switched to CT omission following low-risk MP scores; conversely, 88.9% of physicians initially recommending CT omission switched to CT recommendations following high-risk MP scores. Most patients (99.2%) recommended to forgo CT post-test and 21.3% of patients with post-test CT recommendations did not undergo CT; among MP low-risk patients with pre-test and post-test CT recommendations, 40% did not actually undergo CT. Luminal subtype assessment by BluePrint was discordant with IHC assessment in 34% of patients. Patients' State Anxiety scores improved significantly overall, particularly in MP low-risk patients. Trait Anxiety scores increased slightly in MP high risk and decreased slightly in MP low-risk patients.

CONCLUSIONS

MammaPrint and BluePrint test results strongly impacted physicians' therapy decisions in luminal EBC with up to three involved lymph nodes. The high adherence to genetically determined risk assessment represents a key prerequisite for achieving a personalized cost-effective approach to disease management of early breast cancer.

Unraveling the South African Breast Cancer Story: The Relationship of Patients, Delay to Diagnosis, and Tumor Biology With Stage at Presentation in an Urban Setting

BACKGROUND

Adverse outcomes from breast cancer disproportionately affect women in sub-Saharan Africa, with delay the most studied contribution to advanced stage at presentation. However, tumor molecular biology and its contribution to advanced stage are yet to be explored.

MATERIALS AND METHODS

Patients newly diagnosed with breast cancer in a South African tertiary breast center completed a questionnaire and file review concerning socioeconomics, delay to care, stage at presentation, and molecular characteristics. Logistic regression was done to determine the relative risk of advanced stage presentation.

RESULTS

Advanced stage was present in 70.1% (n = 162) of the 231 participants, with 55.8% stage III (n = 129) and 32% (n = 72) having a T4 tumor. The median age was 56 y with 21.6% (n = 47) aged <45 y. Most common subtype was luminal B (57.7%, n = 128) followed by luminal A (21.6%, n = 48), triple negative (13.9%, n = 31), and HER2 positive (6.7%, n = 15). Lobular cancer (incidence risk ratio [IRR], 1.29; 95% confidence interval [CI], 1.03-1.62), high grade and intermediate grade tumors (IRR, 1.90; 95% CI, 1.15-3.13 and IRR, 1.95; 95% CI, 1.18-3.22, respectively), high Ki67 proliferation index (IRR, 1.30; 95% CI, 1.02-1.66), and HER2 overexpression (IRR, 1.32; 95% CI, 1.12-1.55) were more likely to present with advanced disease, as were luminal B (HER2+) cancers (adjusted IRR [aIRR], 1.46; 95% CI, 1.10-1.95). Although on univariate analysis Black and young participants were both more likely to have advanced stage (IRR, 1.23; 95% CI, 1.01-1.49 and IRR, 1.25; 95% CI, 1.04-1.51, respectively), in multivariate analysis controlling for tumor biology and delay, these were no longer significant (aIRR, 1.12; 95% CI, 0.91-1.37 and aIRR, 1.17; 95% CI, 0.94-1.48, respectively).

CONCLUSIONS

Tumor biology has a compelling role in the etiology of advanced-stage disease irrespective of socioeconomic factors. Accurate pathologic assessment is important in planning breast cancer care in Africa.

Precision Medicine and Radiogenomics in Breast Cancer: New Approaches toward Diagnosis and Treatment

Precision medicine is medicine optimized to the genotypic and phenotypic characteristics of an individual and, when present, his or her disease. It has a host of targets, including genes and their transcripts, proteins, and metabolites. Studying precision medicine involves a systems biology approach that integrates mathematical modeling and biology genomics, transcriptomics, proteomics, and metabolomics. Moreover, precision medicine must consider not only the relatively static genetic codes of individuals, but also the dynamic and heterogeneous genetic codes of cancers. Thus, precision medicine relies not only on discovering identifiable targets for treatment and surveillance modification, but also on reliable, noninvasive methods of identifying changes in these targets over time. Imaging via radiomics and radiogenomics is poised for a central role. Radiomics, which extracts large volumes of quantitative data from digital images and amalgamates these together with clinical and patient data into searchable shared databases, potentiates radiogenomics, which is the combination of genetic and radiomic data. Radiogenomics may provide voxel-by-voxel genetic information for a complete, heterogeneous tumor or, in the setting of metastatic disease, set of tumors and thereby guide tailored therapy. Radiogenomics may also quantify lesion characteristics, to better differentiate between benign and malignant entities, and patient characteristics, to better stratify patients according to risk for disease, thereby allowing for more precise imaging and screening. This report provides an overview of precision medicine and discusses radiogenomics specifically in breast cancer. ^© RSNA, 2018.

Actions of sesquiterpene lactones isolated from Moquiniastrum polymorphum subsp. floccosum in MCF7 cell line and their potentiating action on doxorubicin

Background

In order to obtain better clinical results in anticancer therapies, polychemotherapy or combination therapies are used. For this, the combinations are required to increase the efficacy and reduce the adverse reactions of the associated chemotherapies. The aim of this study was to evaluate the cytotoxic, apoptotic and (anti)proliferative potential of two sesquiterpene lactones isolated from Moquiniastrum polymorphum, 11,13-diidrozaluzanin C (1) and gochnatiolide C (2), and their associations with chemotherapeutic agents irinotecan, tamoxifen, cisplatin, 5-fluouracyl and doxorubicin in the tumoral lineage of MCF-7 breast adenocarcinoma.

Methods

The analyses were performed by MTT cytotoxicity assays, drug combination index (CI), apoptosis morphological assay and cell proliferation assay. Treatments were evaluated with short exposure times (4 h), followed or not by recovery in drug-free medium for 24 h. For the cell viability assay the statistical analysis was performed using software INSTAT, and the ANOVA/Tukey test was applied. Combination Indices (CI) was made using CompuSyn software and demonstrated through isoboles. The assays that evaluated cell death and proliferation used statistical analysis SAS 9.4 (Statistical Analysis System), and the procedure adopted was PROC NPAR1WAY. The Wilcoxon test at 5% level was applied for comparing statistical differences.

Results

The results demonstrated that the compounds decrease cell viability and increase their action when associated with irinotecan, tamoxifen and doxorubicin (CI < 1 and CI = 1). In periods of 4 h-exposure, the compounds cause cell death by apoptosis and after 24 h, they increase the mean number of cells in programmed cell death, especially when treated with 2. In addition, the association with doxorubicin increases the apoptotic potential induced by tested compounds. Both isolates had effect on the reduction of the number of mitoses, especially when 2 at its highest concentration is associated with doxorubicin.

Conclusions

Finally, these compounds are presented as potential agents in chemotherapy combined with doxorubicin, since they trigger the mechanism of apoptosis, which, through the mechanism of action of sesquiterpene lactones, leads to a reduction in toxicity. In addition, the tested compounds have the ability to exert a synergistic action with doxorubicin, possibly by down-regulating the drug resistance mechanisms.

Testing chemotherapy efficacy in HER2 negative breast cancer using patient-derived spheroids

BACKGROUND

Targeted anti-HER2 therapy has greatly improved the prognosis for many breast cancer patients. However, treatment for HER2 negative disease is currently still selected from a multitude of untargeted chemotherapeutic treatment options. A predictive test was developed using patient-derived spheroids to identify the most effective therapy for patients with HER2 negative breast cancer of all stages, for clinically relevant subgroups, as well as individual patients.

METHODS

Tumor samples from 120 HER2 negative patients obtained through biopsy or surgical excision were tested in the breast cancer spheroid model using scaffold-free cell culture. Similarly, spheroids were also generated from established HER2 negative breast cancer cell lines T-47D, MCF7, HCC1143, and HCC1937 to compare treatment efficacy of heterogeneous cell populations from patient tumor tissue with homogeneous cell lines. Spheroids were treated in vitro with guideline-recommended compounds. Treatment mediated impact on cell survival was subsequently quantified using an ATP assay.

RESULTS

Differences were observed in the metabolic activity of the untreated spheroids, whereby cell lines consistently achieved higher values compared to tissue spheroids (p < 0.001). A higher number of cells per spheroid correlated with a higher basal metabolic activity in tissue-derived spheroids (p < 0.01), while the opposite was observed for cell line spheroids (p < 0.01). Recurrent tumors showed a higher mean vitality (p < 0.01) compared to primary tumors. Except for taxanes, treatment efficacy for most tested compounds differed significantly between breast cancer tissue spheroids and breast cancer cell lines. Overall a high variability in treatment response in vitro was seen in the tissue spheroids regardless of the tested substances. A greater response to anthracycline/docetaxel was observed for hormone receptor negative samples (p < 0.01). A higher response to 5-FU (p < 0.01) and anthracycline (p < 0.05) was seen in high grade tumors. Smaller tumor size and negative lymph node status were both associated with a higher treatment efficacy to anthracycline treatment combined with 5-FU (cT1/2 vs cT3/4, p = 0.035, cN+ vs cN-, p < 0.05).

CONCLUSIONS

The tissue spheroid model reflects current guideline treatment recommendations for HER2 negative breast cancer, whereas tested cell lines did not. This model represents a unique diagnostic method to select the most effective therapy out of several equivalent treatment options.