Cancer Grade Model: a multi-gene machine learning-based risk classification for improving prognosis in breast cancer

Investigating the effects of artificial intelligence on the personalization of breast cancer management: a systematic study

BACKGROUND

Providing appropriate specialized treatment to the right patient at the right time is considered necessary in cancer management. Targeted therapy tailored to the genetic changes of each breast cancer patient is a desirable feature of precision oncology, which can not only reduce disease progression but also potentially increase patient survival. The use of artificial intelligence alongside precision oncology can help physicians by identifying and selecting more effective treatment factors for patients.

METHOD

A systematic review was conducted using the PubMed, Embase, Scopus, and Web of Science databases in September 2023. We performed the search strategy with keywords, namely: Breast Cancer, Artificial intelligence, and precision Oncology along with their synonyms in the article titles. Descriptive, qualitative, review, and non-English studies were excluded. The quality assessment of the articles and evaluation of bias were determined based on the SJR journal and JBI indices, as well as the PRISMA2020 guideline.

RESULTS

Forty-six studies were selected that focused on personalized breast cancer management using artificial intelligence models. Seventeen studies using various deep learning methods achieved a satisfactory outcome in predicting treatment response and prognosis, contributing to personalized breast cancer management. Two studies utilizing neural networks and clustering provided acceptable indicators for predicting patient survival and categorizing breast tumors. One study employed transfer learning to predict treatment response. Twenty-six studies utilizing machine-learning methods demonstrated that these techniques can improve breast cancer classification, screening, diagnosis, and prognosis. The most frequent modeling techniques used were NB, SVM, RF, XGBoost, and Reinforcement Learning. The average area under the curve (AUC) for the models was 0.91. Moreover, the average values for accuracy, sensitivity, specificity, and precision were reported to be in the range of 90-96% for the models.

CONCLUSION

Artificial intelligence has proven to be effective in assisting physicians and researchers in managing breast cancer treatment by uncovering hidden patterns in complex omics and genetic data. Intelligent processing of omics data through protein and gene pattern classification and the utilization of deep neural patterns has the potential to significantly transform the field of complex disease management.

Investigation of recurrence prediction ability of EndoPredict® using microarray data from fresh frozen tissues in ER-positive, HER2-negative breast cancer and indication expansion of EndoPredict® from microarray data from fresh-frozen to FFPE tissues

BACKGROUND

EndoPredict® (EP) is a multigene assay to predict distant recurrence risk in luminal breast cancer. EP measures the expression of 12 genes in primary tumor by qRT-PCR from formalin-fixed paraffin-embedded (FFPE) tissues and calculates EP risk score that indicates the risk of distant recurrence. We evaluated the performance of EP in predicting distant recurrence risk using microarray data from fresh frozen (FF) tissues. We also examined the applicability of EP to microarray data from FFPE tissues.

METHODS

We analyzed the publicly available data of 431 node-negative and 270 node-positive patients with luminal breast cancer who received endocrine therapy alone. We evaluated the prognostic value of EP using microarray data from FF tissues. Next, we created an algorithm to calculate EP risk score using microarray data from FFPE tissues. We examined the correlation coefficient of EP risk score and concordance rate of EP risk high/low using microarray data from FFPE/FF tissue pairs in a validation set of 39 patients.

RESULTS

In 431 node-negative patients, the distant recurrence-free survival (DRFS) rate was significantly worse in those with high EP risk scores (P = 3.68 × 10-6, log-rank). The 5-year DRFS was 95.2% in those with low EP risk score. In the validation set, the correlation coefficient of EP risk score was 0.93 and the concordance rate of EP risk high/low was 91.7%.

CONCLUSIONS

EP using microarray data from FF tissues was useful in predicting distant recurrence risk in luminal breast cancer, and EP might be utilized in microarray data from FFPE tissues.

Exploring Prognostic Gene Factors in Breast Cancer via Machine Learning

Breast cancer remains the most prevalent cancer in women. To date, its underlying molecular mechanisms have not been fully uncovered. The determination of gene factors is important to improve our understanding on breast cancer, which can correlate the specific gene expression and tumor staging. However, the knowledge in this regard is still far from complete. Thus, this study aimed to explore these knowledge gaps by analyzing existing gene expression profile data from 3149 breast cancer samples, where each sample was represented by the expression of 19,644 genes and classified into Nottingham histological grade (NHG) classes (Grade 1, 2, and 3). To this end, a machine learning-based framework was designed. First, the profile data were analyzed by using seven feature ranking algorithms to evaluate the importance of features (genes). Seven feature lists were generated, each of which sorted features in accordance with feature importance evaluated from a special aspect. Then, the incremental feature selection method was applied to each list to determine essential features for classification and building efficient classifiers. Consequently, overlapping genes, such as AURKA, CBX2, and MYBL2, were deemed as potentially related to breast cancer malignancy and prognosis, indicating that such genes were identified to be important by multiple feature ranking algorithms. In addition, the study formulated classification rules to reflect special gene expression patterns for three NHG classes. Some genes and rules were analyzed and supported by recent literature, providing new references for studying breast cancer.

Endoplasmic reticulum stress in breast cancer: a predictive model for prognosis and therapy selection

Background

Breast cancer (BC) is a leading cause of mortality among women, underscoring the urgent need for improved therapeutic predictio. Developing a precise prognostic model is crucial. The role of Endoplasmic Reticulum Stress (ERS) in cancer suggests its potential as a critical factor in BC development and progression, highlighting the importance of precise prognostic models for tailored treatment strategies.

Methods

Through comprehensive analysis of ERS-related gene expression in BC, utilizing both single-cell and bulk sequencing data from varied BC subtypes, we identified eight key ERS-related genes. LASSO regression and machine learning techniques were employed to construct a prognostic model, validated across multiple datasets and compared with existing models for its predictive accuracy.

Results

The developed ERS-model categorizes BC patients into distinct risk groups with significant differences in clinical prognosis, confirmed by robust ROC, DCA, and KM analyses. The model forecasts survival rates with high precision, revealing distinct immune infiltration patterns and treatment responsiveness between risk groups. Notably, we discovered six druggable targets and validated Methotrexate and Gemcitabine as effective agents for high-risk BC treatment, based on their sensitivity profiles and potential for addressing the lack of active targets in BC.

Conclusion

Our study advances BC research by establishing a significant link between ERS and BC prognosis at both the molecular and cellular levels. By stratifying patients into risk-defined groups, we unveil disparities in immune cell infiltration and drug response, guiding personalized treatment. The identification of potential drug targets and therapeutic agents opens new avenues for targeted interventions, promising to enhance outcomes for high-risk BC patients and paving the way for personalized cancer therapy.

Associations of accumulated persistent organic pollutants in breast adipose tissue with the evolution of breast cancer after surgery

Chronic exposure to persistent organic pollutants (POPs) is suspected to contribute to the onset of breast cancer, but the impact on the evolution of patients after diagnosis is unclear. We aimed to analyze the contribution of long-term exposure to five POPs to overall mortality, cancer recurrence, metastasis, and development of second primary tumors over a global follow-up of 10 years after surgery in breast cancer patients in a cohort study. Between 2012 and 2014, a total of 112 newly diagnosed breast cancer patients were recruited from a public hospital in Granada, Southern Spain. Historical exposure to POPs was estimated by analyzing their concentrations in breast adipose tissue samples. Sociodemographic data were collected through face-to-face interviews, while data on evolution tumor were retrieved from clinical records. Statistical analyses were performed using Cox regression (overall survival, breast cancer recurrence or metastasis) and binary logistic regression models (joint outcome variable). We also tested for statistical interactions of POPs with age, residence, and prognostic markers. The third vs first tertile of hexachlorobenzene concentrations was associated with a lower risk of all-cause mortality (Hazard Ratio, HR = 0.26; 95 % Confidence Interval, CI = 0.07-0.92) and of the appearance of any of the four events (Odds Ratio = 0.37; 95 % CI = 0.14-1.03). Polychlorinated biphenyl 138 concentrations were significantly and inversely associated with risk of metastasis (HR = 0.65; 95 % CI = 0.44-0.97) and tumor recurrence (HR = 0.69; 95 % CI = 0.49-0.98). Additionally, p,p'-dichlorodiphenyldichloroethylene showed inverse associations with risk of metastasis in women with ER-positive tumors (HR = 0.49; 95 % CI = 0.25-0.93) and in those with a tumor size <2.0 cm (HR = 0.39; 95 % CI = 0.18-0.87). The observed paradoxical inverse associations of POP exposure with breast cancer evolution might be related to either a better prognosis of hormone-dependent tumors, which have an approachable pharmacological target, or an effect of sequestration of circulating POPs by adipose tissue.

Drug repurposing and prediction of multiple interaction types via graph embedding

BACKGROUND

Finding drugs that can interact with a specific target to induce a desired therapeutic outcome is key deliverable in drug discovery for targeted treatment. Therefore, both identifying new drug-target links, as well as delineating the type of drug interaction, are important in drug repurposing studies.

RESULTS

A computational drug repurposing approach was proposed to predict novel drug-target interactions (DTIs), as well as to predict the type of interaction induced. The methodology is based on mining a heterogeneous graph that integrates drug-drug and protein-protein similarity networks, together with verified drug-disease and protein-disease associations. In order to extract appropriate features, the three-layer heterogeneous graph was mapped to low dimensional vectors using node embedding principles. The DTI prediction problem was formulated as a multi-label, multi-class classification task, aiming to determine drug modes of action. DTIs were defined by concatenating pairs of drug and target vectors extracted from graph embedding, which were used as input to classification via gradient boosted trees, where a model is trained to predict the type of interaction. After validating the prediction ability of DT2Vec+, a comprehensive analysis of all unknown DTIs was conducted to predict the degree and type of interaction. Finally, the model was applied to propose potential approved drugs to target cancer-specific biomarkers.

CONCLUSION

DT2Vec+ showed promising results in predicting type of DTI, which was achieved via integrating and mapping triplet drug-target-disease association graphs into low-dimensional dense vectors. To our knowledge, this is the first approach that addresses prediction between drugs and targets across six interaction types.

Machine Learning Models for the Identification of Prognostic and Predictive Cancer Biomarkers: A Systematic Review

The identification of biomarkers plays a crucial role in personalized medicine, both in the clinical and research settings. However, the contrast between predictive and prognostic biomarkers can be challenging due to the overlap between the two. A prognostic biomarker predicts the future outcome of cancer, regardless of treatment, and a predictive biomarker predicts the effectiveness of a therapeutic intervention. Misclassifying a prognostic biomarker as predictive (or vice versa) can have serious financial and personal consequences for patients. To address this issue, various statistical and machine learning approaches have been developed. The aim of this study is to present an in-depth analysis of recent advancements, trends, challenges, and future prospects in biomarker identification. A systematic search was conducted using PubMed to identify relevant studies published between 2017 and 2023. The selected studies were analyzed to better understand the concept of biomarker identification, evaluate machine learning methods, assess the level of research activity, and highlight the application of these methods in cancer research and treatment. Furthermore, existing obstacles and concerns are discussed to identify prospective research areas. We believe that this review will serve as a valuable resource for researchers, providing insights into the methods and approaches used in biomarker discovery and identifying future research opportunities.

Rapid identification of breast cancer subtypes using micro-FTIR and machine learning methods

Breast cancer (BC) molecular subtypes diagnosis involves improving clinical uptake by Fourier transform infrared (FTIR) spectroscopic imaging, which is a non-destructive and powerful technique, enabling label free extraction of biochemical information towards prognostic stratification and evaluation of cell functionality. However, methods of measurements of samples demand a long time to achieve high quality images, making its clinical use impractical because of the data acquisition speed, poor signal to noise ratio, and deficiency of optimized computational framework procedures. To address those challenges, machine learning (ML) tools can facilitate obtaining an accurate classification of BC subtypes with high actionability and accuracy. Here, we propose a ML-algorithm-based method to distinguish computationally BC cell lines. The method is developed by coupling the K-neighbors classifier (KNN) with neighborhood components analysis (NCA), and hence, the NCA-KNN method enables to identify BC subtypes without increasing model size as well as adding additional computational parameters. By incorporating FTIR imaging data, we show that classification accuracy, specificity, and sensitivity improve, respectively, 97.5%, 96.3%, and 98.2%, even at very low co-added scans and short acquisition times. Moreover, a clear distinctive accuracy (up to 9 %) difference of our proposed method (NCA-KNN) was obtained in comparison with the second best supervised support vector machine model. Our results suggest a key diagnostic NCA-KNN method for BC subtypes classification that may translate to advancement of its consolidation in subtype-associated therapeutics.

Determination of a DNA repair-related gene signature with potential implications for prognosis and therapeutic response in pancreatic adenocarcinoma

Background

Pancreatic adenocarcinoma (PAAD) is one of the leading causes of cancer death worldwide. Alterations in DNA repair-related genes (DRGs) are observed in a variety of cancers and have been shown to affect the development and treatment of cancers. The aim of this study was to develop a DRG-related signature for predicting prognosis and therapeutic response in PAAD.

Methods

We constructed a DRG signature using least absolute shrinkage and selection operator (LASSO) Cox regression analysis in the TCGA training set. GEO datasets were used as the validation set. A predictive nomogram was constructed based on multivariate Cox regression. Calibration curve and decision curve analysis (DCA) were applied to validate the performance of the nomogram. The CIBERSORT and ssGSEA algorithms were utilized to explore the relationship between the prognostic signature and immune cell infiltration. The pRRophetic algorithm was used to estimate sensitivity to chemotherapeutic agents. The CellMiner database and PAAD cell lines were used to investigate the relationship between DRG expression and therapeutic response.

Results

We developed a DRG signature consisting of three DRGs (RECQL, POLQ, and RAD17) that can predict prognosis in PAAD patients. A prognostic nomogram combining the risk score and clinical factors was developed for prognostic prediction. The DCA curve and the calibration curve demonstrated that the nomogram has a higher net benefit than the risk score and TNM staging system. Immune infiltration analysis demonstrated that the risk score was positively correlated with the proportions of activated NK cells and monocytes. Drug sensitivity analysis indicated that the signature has potential predictive value for chemotherapy. Analyses utilizing the CellMiner database showed that RAD17 expression is correlated with oxaliplatin. The dynamic changes in three DRGs in response to oxaliplatin were examined by RT-qPCR, and the results show that RAD17 is upregulated in response to oxaliplatin in PAAD cell lines.

Conclusion

We constructed and validated a novel DRG signature for prediction of the prognosis and drug sensitivity of patients with PAAD. Our study provides a theoretical basis for further unraveling the molecular pathogenesis of PAAD and helps clinicians tailor systemic therapies within the framework of individualized treatment.

A fully-convolutional residual encoder-decoder neural network to localize breast cancer on histopathology images

Cancer detection in its early stages may allow patients to receive the proper treatment and save lives along with recovering the routine lifestyles. Breast cancer is of the top leading causes of mortality among women all around the globe. A source to find these cancerous nuclei is through analyzing histopathology images. These images, however, are very complex and large. Thus, locating the cancerous nuclei in them is very challenging. Hence, if an expert fails to diagnose their patients via these images, the situation may be exacerbated. Therefore, this study aims to introduce a method to mask as many cancer nuclei on histopathology images as possible with a high visual aesthetic to make them distinguishable by experts easily. A tailored residual fully convolutional encoder-decoder neural network based on end-to-end learning is proposed to issue the matter. The proposed method is evaluated quantitatively and qualitatively on ER + BCa H&E-stained dataset. The average detection accuracy achieved by the method is 98.61%, which is much better than that of competitors.

DNA Nanowire Guided-Catalyzed Hairpin Assembly Nanoprobe for In Situ Profiling of Circulating Extracellular Vesicle-Associated MicroRNAs

Extracellular vesicle-associated miRNAs (EV-miRNAs) are emerging as a new type of noninvasive biomarker for disease diagnosis. Their relatively low abundance, however, makes accurate detection challenging. Here, we designed a DNA nanowire guided-catalyzed hairpin assembly (NgCHA) nanoprobe for profiling EV-miRNAs. NgCHA showed high penetrability to EVs, which allowed rapid delivery of the probes into EVs. In the presence of targeted miRNAs within EVs, a fluorescent signal could be generated and amplified by confining the catalytic hairpin assembly system within the nanowires, thus greatly enhancing the analytical sensitivity. We showed that EV-miRNAs from various cell lines could be accurately quantified by NgCHA in situ. By using a four-EV-miRNA panel, this platform can identify patients with breast cancer at an early stage with 95.2% sensitivity and 86.7% specificity. Its applications for risk assessment as well as cancer type prediction were also successfully demonstrated. This platform is sensitive, low-cost, and simple compared with current methods. It may thus serve as a promising tool for the noninvasive diagnosis and monitoring of cancers and other diseases through EV-miRNA profiling.

Presentation of Novel Architecture for Diagnosis and Identifying Breast Cancer Location Based on Ultrasound Images Using Machine Learning

Breast cancer is one of the main causes of death among women worldwide. Early detection of this disease helps reduce the number of premature deaths. This research aims to design a method for identifying and diagnosing breast tumors based on ultrasound images. For this purpose, six techniques have been performed to detect and segment ultrasound images. Features of images are extracted using the fractal method. Moreover, k-nearest neighbor, support vector machine, decision tree, and Naïve Bayes classification techniques are used to classify images. Then, the convolutional neural network (CNN) architecture is designed to classify breast cancer based on ultrasound images directly. The presented model obtains the accuracy of the training set to 99.8%. Regarding the test results, this diagnosis validation is associated with 88.5% sensitivity. Based on the findings of this study, it can be concluded that the proposed high-potential CNN algorithm can be used to diagnose breast cancer from ultrasound images. The second presented CNN model can identify the original location of the tumor. The results show 92% of the images in the high-performance region with an AUC above 0.6. The proposed model can identify the tumor's location and volume by morphological operations as a post-processing algorithm. These findings can also be used to monitor patients and prevent the growth of the infected area.