Model Comparison for Breast Cancer Prognosis Based on Clinical Data

Modelling-based joint embedding of histology and genomics using canonical correlation analysis for breast cancer survival prediction

Traditional approaches to predicting breast cancer patients' survival outcomes were based on clinical subgroups, the PAM50 genes, or the histological tissue's evaluation. With the growth of multi-modality datasets capturing diverse information (such as genomics, histology, radiology and clinical data) about the same cancer, information can be integrated using advanced tools and have improved survival prediction. These methods implicitly exploit the key observation that different modalities originate from the same cancer source and jointly provide a complete picture of the cancer. In this work, we investigate the benefits of explicitly modelling multi-modality data as originating from the same cancer under a probabilistic framework. Specifically, we consider histology and genomics as two modalities originating from the same breast cancer under a probabilistic graphical model (PGM). We construct maximum likelihood estimates of the PGM parameters based on canonical correlation analysis (CCA) and then infer the underlying properties of the cancer patient, such as survival. Equivalently, we construct CCA-based joint embeddings of the two modalities and input them to a learnable predictor. Real-world properties of sparsity and graph-structures are captured in the penalized variants of CCA (pCCA) and are better suited for cancer applications. For generating richer multi-dimensional embeddings with pCCA, we introduce two novel embedding schemes that encourage orthogonality to generate more informative embeddings. The efficacy of our proposed prediction pipeline is first demonstrated via low prediction errors of the hidden variable and the generation of informative embeddings on simulated data. When applied to breast cancer histology and RNA-sequencing expression data from The Cancer Genome Atlas (TCGA), our model can provide survival predictions with average concordance-indices of up to 68.32% along with interpretability. We also illustrate how the pCCA embeddings can be used for survival analysis through Kaplan-Meier curves.

Big data and variceal rebleeding prediction in cirrhosis patients

Big data has convincing merits in developing risk stratification strategies for diseases. The 6 “V”s of big data, namely, volume, velocity, variety, veracity, value, and variability, have shown promise for real-world scenarios. Big data can be applied to analyze health data and advance research in preclinical biology, medicine, and especially disease initiation, development, and control. A study design comprises data selection, inclusion and exclusion criteria, standard confirmation and cohort establishment, follow-up strategy, and events of interest. The development and efficiency verification of a prognosis model consists of deciding the data source, taking previous models as references while selecting candidate predictors, assessing model performance, choosing appropriate statistical methods, and model optimization. The model should be able to inform disease development and outcomes, such as predicting variceal rebleeding in patients with cirrhosis. Our work has merits beyond those of other colleagues with respect to cirrhosis patient screening and data source regarding variceal bleeding.

Long-term effects of aromatase inhibitors on body mass index among postmenopausal breast cancer survivors in Africa: observational cohort study

PURPOSE

this study was conducted to assess the impact of AIs on body mass index and high sensitivity as prognostic predictors to be incorporated into point of care technology (POCT) testing in postmenopausal breast cancer women after a 24 month follow up in Africa. An observational cohort study was conducted; including 126 female BC patients with stages ranging from 0-III initially subjected to AIs and subsequently followed up for 24 months. Multiple imputation model was conducted to predict missing data.

RESULTS

Random effects model was used to monitor the changes over the time. The study revealed stronger statistically association between BMI and homocysteine (p = 0.021, 95%CI: 0.0083 to 0.1029). Weight and total body fat were strongly associated after 24 months follow up. Hs-CRP was associated with BMI (p = 0.0001), and hs-CRP was associated with other biomedical markers such as calcium (p = 0.021, 95% CI: 0.01 to 0.10), phosphate (p = 0.039, 95%CI: 0.01 to 0.10), and ferritin (p = 0.002, 95%CI: 0.02 to 0.08) and calcium. The patients subjected to AIs are likely to develop cardiovascular adverse events. POCT of care strategy which include clinical, biomedical and genetic predictor's measurement is required to improve BC survivorship.

Development of an Artificial Intelligence-Based Breast Cancer Detection Model by Combining Mammograms and Medical Health Records

BACKGROUND

Artificial intelligence (AI)-based computational models that analyze breast cancer have been developed for decades. The present study was implemented to investigate the accuracy and efficiency of combined mammography images and clinical records for breast cancer detection using machine learning and deep learning classifiers.

METHODS

This study was verified using 731 images from 357 women who underwent at least one mammogram and had clinical records for at least six months before mammography. The model was trained on mammograms and clinical variables to discriminate benign and malignant lesions. Multiple pre-trained deep CNN models to detect cancer in mammograms, including X-ception, VGG16, ResNet-v2, ResNet50, and CNN3 were employed. Machine learning models were constructed using k-nearest neighbor (KNN), support vector machine (SVM), random forest (RF), Artificial Neural Network (ANN), and gradient boosting machine (GBM) in the clinical dataset.

RESULTS

The detection performance obtained an accuracy of 84.5% with a specificity of 78.1% at a sensitivity of 89.7% and an AUC of 0.88. When trained on mammography image data alone, the result achieved a slightly lower score than the combined model (accuracy, 72.5% vs. 84.5%, respectively).

CONCLUSIONS

A breast cancer-detection model combining machine learning and deep learning models was performed in this study with a satisfactory result, and this model has potential clinical applications.

Addressing the Clinical Feasibility of Adopting Circulating miRNA for Breast Cancer Detection, Monitoring and Management with Artificial Intelligence and Machine Learning Platforms

Detecting breast cancer (BC) at the initial stages of progression has always been regarded as a lifesaving intervention. With modern technology, extensive studies have unraveled the complexity of BC, but the current standard practice of early breast cancer screening and clinical management of cancer progression is still heavily dependent on tissue biopsies, which are invasive and limited in capturing definitive cancer signatures for more comprehensive applications to improve outcomes in BC care and treatments. In recent years, reviews and studies have shown that liquid biopsies in the form of blood, containing free circulating and exosomal microRNAs (miRNAs), have become increasingly evident as a potential minimally invasive alternative to tissue biopsy or as a complement to biomarkers in assessing and classifying BC. As such, in this review, the potential of miRNAs as the key BC signatures in liquid biopsy are addressed, including the role of artificial intelligence (AI) and machine learning platforms (ML), in capitalizing on the big data of miRNA for a more comprehensive assessment of the cancer, leading to practical clinical utility in BC management.

Characterization of breast tumors using machine learning based upon multiparametric magnetic resonance imaging features

Magnetic resonance imaging (MRI) is playing an important role in the classification of breast tumors. MRI can be used to obtain multiparametric (mp) information, such as structural, hemodynamic, and physiological information. Quantitative analysis of mp-MRI data has shown potential in improving the accuracy of breast tumor classification. In general, a large set of quantitative and texture features can be generated depending upon the type of methodology used. A suitable combination of selected quantitative and texture features can further improve the accuracy of tumor classification. Machine learning (ML) classifiers based upon features derived from MRI data have shown potential in tumor classification. There is a need for further research studies on selecting an appropriate combination of features and evaluating the performance of different ML classifiers for accurate classification of breast tumors. The objective of the current study was to develop and optimize an ML framework based upon mp-MRI features for the characterization of breast tumors (malignant vs. benign and low- vs. high-grade). This study included the breast mp-MRI data of 60 female patients with histopathology results. A total of 128 features were extracted from the mp-MRI tumor data followed by features selection. Five ML classifiers were evaluated for tumor classification using 10-fold crossvalidation with 10 repetitions. The support vector machine (SVM) classifier based on optimum features selected using a wrapper method with an adaptive boosting (AdaBoost) technique provided the highest sensitivity (0.96 ± 0.03), specificity (0.92 ± 0.09), and accuracy (94% ± 2.91%) in the classification of malignant versus benign tumors. This method also provided the highest sensitivity (0.94 ± 0.07), specificity (0.80 ± 0.05), and accuracy (90% ± 5.48%) in the classification of low- versus high-grade tumors. These findings suggest that the SVM classifier outperformed other ML methods in the binary classification of breast tumors.

Modeling breast cancer survival and metastasis rates from moderate-sized clinical data

Predicting time-dependent survival probability of a breast cancer patient using information such as primary tumor size, grade, node spread status, and patient age at the time of surgery can be of immense help in managing life expectations and strategizing postoperative treatment. However, for moderate-sized clinical datasets the application of standard Kaplan-Meier theory to determine survival probability as a function of multiple cofactors can become challenging when continuous variables like tumor diameter and survival time are segmented into a large number of narrow intervals, a problem commonly termed the curse of dimensionality. We circumvent this problem by modeling the patient-to-patient distribution of primary tumor diameter with a realistic, right-skewed function, and then matching the diameter-marginalized survival with the mean Kaplan-Meier survival for the data. We apply this procedure on a recent clinical data from 1875 breast cancer patients and develop parameters that can be readily used to estimate post-surgery survival for an arbitrary time length. Finally, we show that the observed fraction of node-positive patients can be quantitatively explained within a simple tumor growth and metastasis framework. Employing two different tumor growth models from the literature (i.e., Gompertz and logistic growth models), we utilize the observed fraction-node-positive data to determine metastasis rates from the surface of a primary tumor and its patient-to-patient distribution.

Direct Comparison of Total Clearance Prediction: Computational Machine Learning Model versus Bottom-Up Approach Using In Vitro Assay

The in vitro-in vivo extrapolation (IVIVE) approach for predicting total plasma clearance (CL_tot) has been widely used to rank order compounds early in discovery. More recently, a computational machine learning approach utilizing physicochemical descriptors and fingerprints calculated from chemical structure information has emerged, enabling virtual predictions even earlier in discovery. Previously, this approach focused more on in vitro intrinsic clearance (CL_int) prediction. Herein, we directly compare these two approaches for predicting CL_tot in rats. A structurally diverse set of 1114 compounds with known in vivo CL_tot, in vitro CL_int, and plasma protein binding was used as the basis for this evaluation. The machine learning models were assessed by validation approaches using the time- and cluster-split training and test sets, and five-fold cross validation. Assessed by five-fold validation, the random forest regression (RF) and radial basis function (RBF) models demonstrated better prediction performance in eight attempted machine learning models. The CL_tot values predicted by the RF and RBF models were within two-fold of the observed values for 67.7 and 71.9% of cluster-split test set compounds, respectively, while the predictivity was worse in the time-split dataset. The predictivity of both models tended to be improved by incorporating in vitro parameters, unbound fraction in plasma (f_u,p), and CL_int. CL_tot prediction utilizing in vitro CL_int and the well-stirred model, correcting for the fraction unbound in blood, was substantially worse compared to machine learning approaches for the same cluster-split test set. The reason that CL_tot is underestimated by IVIVE is not fully explained by considering the calculated microsomal unbound fraction (cf_u,mic), extended clearance classification system (ECCS), and omitting high clearance compounds in excess of hepatic blood flow. The analysis suggests that in silico machine learning models may have the power to reduce reliance on or replace in vitro and in vivo studies for chemical structure optimization in early drug discovery.

Use of Machine Learning Techniques for Case-Detection of Varicella Zoster Using Routinely Collected Textual Ambulatory Records: Pilot Observational Study

Background

The detection of infectious diseases through the analysis of free text on electronic health reports (EHRs) can provide prompt and accurate background information for the implementation of preventative measures, such as advertising and monitoring the effectiveness of vaccination campaigns.

Objective

The purpose of this paper is to compare machine learning techniques in their application to EHR analysis for disease detection.

Methods

The Pedianet database was used as a data source for a real-world scenario on the identification of cases of varicella. The models’ training and test sets were based on two different Italian regions’ (Veneto and Sicilia) data sets of 7631 patients and 1,230,355 records, and 2347 patients and 569,926 records, respectively, for whom a gold standard of varicella diagnosis was available. Elastic-net regularized generalized linear model (GLMNet), maximum entropy (MAXENT), and LogitBoost (boosting) algorithms were implemented in a supervised environment and 5-fold cross-validated. The document-term matrix generated by the training set involves a dictionary of 1,871,532 tokens. The analysis was conducted on a subset of 29,096 tokens, corresponding to a matrix with no more than a 99% sparsity ratio.

Results

The highest predictive values were achieved through boosting (positive predicative value [PPV] 63.1, 95% CI 42.7-83.5 and negative predicative value [NPV] 98.8, 95% CI 98.3-99.3). GLMNet delivered superior predictive capability compared to MAXENT (PPV 24.5% and NPV 98.3% vs PPV 11.0% and NPV 98.0%). MAXENT and GLMNet predictions weakly agree with each other (agreement coefficient 1 [AC1]=0.60, 95% CI 0.58-0.62), as well as with LogitBoost (MAXENT: AC1=0.64, 95% CI 0.63-0.66 and GLMNet: AC1=0.53, 95% CI 0.51-0.55).

Conclusions

Boosting has demonstrated promising performance in large-scale EHR-based infectious disease identification.

Identification of Prognostic Signatures for Predicting the Overall Survival of Uveal Melanoma Patients

Uveal melanoma (UM) is an aggressive cancer which has a high percentage of metastasis and with a poor prognosis. Identifying the potential prognostic markers of uveal melanoma may provide information for early detection of metastasis and treatment. In this work, we analyzed 80 uveal melanoma samples from The Cancer Genome Atlas (TCGA). We developed an 18-gene signature which can significantly predict the prognosis of UM patients. Firstly, we performed a univariate Cox regression analysis to identify significantly prognostic genes in uveal melanoma (P<0.01). Then the glmnet Cox analysis was used to generate a powerful prognostic gene model. Further, we established a risk score formula for every patient based on the 18-gene prognostic model with multivariate Cox regression. We stratified patients into high- and low-risk subtypes with median risk score and found that patients in high-risk group had worse prognosis than patients in low-risk group. Multivariate Cox regression analysis demonstrated that 18-gene model risk score was independent of clinical prognostic factors. We identified four genes whose mutations were closely to UM patients' prognosis or risk score. We also explored the relationship between copy number variation and risk score and found that high risk group showed more chromosome aberrations than low risk group. Gene Set Enrichment Analysis (GSEA) analysis showed that the different biological pathways and functions between low and high risk group. In summary, our findings constructed an 18-gene signature for estimating overall survival (OS) of UM. Patients were categorized into two subtypes based on the risk score and we found that high risk group showed more chromosome aberrations than low risk group.

Developing a periodontal disease antibody array for the prediction of severe periodontal disease using machine learning classifiers

BACKGROUND

The aim of this study was to simultaneously and quantitatively assess the expression levels of 20 periodontal disease-related proteins in gingival crevicular fluid (GCF) from normal controls (NOR) and severe periodontitis (SP) patients with an antibody array.

METHODS

Antibodies against 20 periodontal disease-related proteins were spotted onto a glass slide to create a periodontal disease antibody array (PDD). The array was then incubated with GCF samples collected from 25 NOR and 25 SP patients. Differentially expressed proteins between NOR and SP patients were then used to build receiver operator characteristic (ROC) curves and compare five classification models, including support vector machine, random forest, k nearest neighbor, linear discriminant analysis, and Classification and Regression Trees.

RESULTS

Seven proteins (C-reactive protein, interleukin [IL]-1α, interleukin-1β, interleukin-8, matrix metalloproteinase-13, osteoprotegerin, and osteoactivin) were significantly upregulated in SP patients compared with NOR, while receptor activator of nuclear factor-kappa was significantly downregulated. The highest diagnostic accuracy using a ROC curve was observed for IL-1β with an area under the curve of 0.984. Five of the proteins (IL-1β, IL-8, MMP-13, osteoprotegerin, and osteoactivin) were identified as important features for classification. Linear discriminant analysis had the highest classification accuracy across the five classification models that were tested.

CONCLUSION

This study highlights the potential of antibody arrays to diagnose periodontal disease.

Association Analysis of Deep Genomic Features Extracted by Denoising Autoencoders in Breast Cancer

Artificial intelligence-based unsupervised deep learning (DL) is widely used to mine multimodal big data. However, there are few applications of this technology to cancer genomics. We aim to develop DL models to extract deep features from the breast cancer gene expression data and copy number alteration (CNA) data separately and jointly. We hypothesize that the deep features are associated with patients’ clinical characteristics and outcomes. Two unsupervised denoising autoencoders (DAs) were developed to extract deep features from TCGA (The Cancer Genome Atlas) breast cancer gene expression and CNA data separately and jointly. A heat map was used to view and cluster patients into subgroups based on these DL features. Fisher’s exact test and Pearson’ Chi-square test were applied to test the associations of patients’ groups and clinical information. Survival differences between the groups were evaluated by Kaplan–Meier (KM) curves. Associations between each of the features and patient’s overall survival were assessed using Cox’s proportional hazards (COX-PH) model and a risk score for each feature set from the different omics data sets was generated from the survival regression coefficients. The risk scores for each feature set were binarized into high- and low-risk patient groups to evaluate survival differences using KM curves. Furthermore, the risk scores were traced back to their gene level DAs weights so that the three gene lists for each of the genomic data points were generated to perform gene set enrichment analysis. Patients were clustered into two groups based on concatenated features from the gene expression and CNA data and these two groups showed different overall survival rates (p-value = 0.049) and different ER (Estrogen receptor) statuses (p-value = 0.002, OR (odds ratio) = 0.626). All the risk scores from the gene expression and CNA data and their concatenated one were significantly associated with breast cancer survival. The patients with the high-risk group were significantly associated with patients’ worse outcomes (p-values ≤ 0.0023). The concatenated risk score was enriched by the AMP-activated protein kinase (AMPK) signaling pathway, the regulation of DNA-templated transcription, the regulation of nucleic acid-templated transcription, the regulation of apoptotic process, the positive regulation of gene expression, the positive regulation of cell proliferation, heart morphogenesis, the regulation of cellular macromolecule biosynthetic process, with FDR (false discovery rate) less than 0.05. We confirmed DAs can effectively extract meaningful genomic features from genomic data and concatenating multiple data sources can improve the significance of the features associated with breast cancer patients’ clinical characteristics and outcomes.

Health-Related Quality of Life in a Predictive Model for Mortality in Older Breast Cancer Survivors

OBJECTIVES

To develop a predictive model and risk score for 10-year mortality using health-related quality of life (HRQOL) in a cohort of older women with early-stage breast cancer.

DESIGN

Prospective cohort.

SETTING

Community.

PARTICIPANTS

U.S. women aged 65 and older diagnosed with Stage I to IIIA primary breast cancer (N=660).

MEASUREMENTS

We used medical variables (age, comorbidity), HRQOL measures (10-item Physical Function Index and 5-item Mental Health Index from the Medical Outcomes Study (MOS) 36-item Short-Form Survey; 8-item Modified MOS Social Support Survey), and breast cancer variables (stage, surgery, chemotherapy, endocrine therapy) to develop a 10-year mortality risk score using penalized logistic regression models. We assessed model discriminative performance using the area under the receiver operating characteristic curve (AUC), calibration performance using the Hosmer-Lemeshow test, and overall model performance using Nagelkerke R² (NR).

RESULTS

Compared to a model including only age, comorbidity, and cancer stage and treatment variables, adding HRQOL variables improved discrimination (AUC 0.742 from 0.715) and overall performance (NR 0.221 from 0.190) with good calibration (p=0.96 from HL test).

CONCLUSION

In a cohort of older women with early-stage breast cancer, HRQOL measures predict 10-year mortality independently of traditional breast cancer prognostic variables. These findings suggest that interventions aimed at improving physical function, mental health, and social support might improve both HRQOL and survival.

Breast cancer data analysis for survivability studies and prediction

BACKGROUND

Breast cancer is the most common cancer affecting females worldwide. Breast cancer survivability prediction is challenging and a complex research task. Existing approaches engage statistical methods or supervised machine learning to assess/predict the survival prospects of patients.

OBJECTIVE

The main objectives of this paper is to develop a robust data analytical model which can assist in (i) a better understanding of breast cancer survivability in presence of missing data, (ii) providing better insights into factors associated with patient survivability, and (iii) establishing cohorts of patients that share similar properties.

METHODS

Unsupervised data mining methods viz. the self-organising map (SOM) and density-based spatial clustering of applications with noise (DBSCAN) is used to create patient cohort clusters. These clusters, with associated patterns, were used to train multilayer perceptron (MLP) model for improved patient survivability analysis. A large dataset available from SEER program is used in this study to identify patterns associated with the survivability of breast cancer patients. Information gain was computed for the purpose of variable selection. All of these methods are data-driven and require little (if any) input from users or experts.

RESULTS

SOM consolidated patients into cohorts of patients with similar properties. From this, DBSCAN identified and extracted nine cohorts (clusters). It is found that patients in each of the nine clusters have different survivability time. The separation of patients into clusters improved the overall survival prediction accuracy based on MLP and revealed intricate conditions that affect the accuracy of a prediction.

CONCLUSIONS

A new, entirely data driven approach based on unsupervised learning methods improves understanding and helps identify patterns associated with the survivability of patient. The results of the analysis can be used to segment the historical patient data into clusters or subsets, which share common variable values and survivability. The survivability prediction accuracy of a MLP is improved by using identified patient cohorts as opposed to using raw historical data. Analysis of variable values in each cohort provide better insights into survivability of a particular subgroup of breast cancer patients.