A Bayesian derived network of breast pathology co-occurrence

Identifying risk factors of developing type 2 diabetes from an adult population with initial prediabetes using a Bayesian network

Background

It is known that people with prediabetes increase their risk of developing type 2 diabetes (T2D), which constitutes a global public health concern, and it is associated with other diseases such as cardiovascular disease.

Methods

This study aimed to determine those factors with high influence in the development of T2D once prediabetes has been diagnosed, through a Bayesian network (BN), which can help to prevent T2D. Furthermore, the set of features with the strongest influences on T2D can be determined through the Markov blanket. A BN model for T2D was built from a dataset composed of 12 relevant features of the T2D domain, determining the dependencies and conditional independencies from empirical data in a multivariate context. The structure and parameters were learned with the bnlearn package in R language introducing prior knowledge. The Markov blanket was considered to find those features (variables) which increase the risk of T2D.

Results

The BN model established the different relationships among features (variables). Through inference, a high estimated probability value of T2D was obtained when the body mass index (BMI) was instantiated to obesity value, the glycosylated hemoglobin (HbA1c) to more than 6 value, the fatty liver index (FLI) to more than 60 value, physical activity (PA) to no state, and age to 48-62 state. The features increasing T2D in specific states (warning factors) were ranked.

Conclusion

The feasibility of BNs in epidemiological studies is shown, in particular, when data from T2D risk factors are considered. BNs allow us to order the features which influence the most the development of T2D. The proposed BN model might be used as a general tool for prevention, that is, to improve the prognosis.

Application of Bayesian network modeling to pathology informatics

BACKGROUND

In the era of extensive data collection, there is a growing need for a large scale data analysis with tools that can handle many variables in one modeling framework. In this article, we present our recent applications of Bayesian network modeling to pathology informatics.

METHODS

Bayesian networks (BNs) are probabilistic graphical models that represent domain knowledge and allow investigators to process this knowledge following sound rules of probability theory. BNs can be built based on expert opinion as well as learned from accumulating data sets. BN modeling is now recognized as a suitable approach for knowledge representation and reasoning under uncertainty. Over the last two decades BN have been successfully applied to many studies on medical prognosis and diagnosis.

RESULTS

Based on data and expert knowledge, we have constructed several BN models to assess patient risk for subsequent specific histopathologic diagnoses and their related prognosis in gynecological cytopathology and breast pathology. These models include the Pittsburgh Cervical Cancer Screening Model assessing risk for histopathologic diagnoses of cervical precancer and cervical cancer, modeling of the significance of benign-appearing endometrial cells in Pap tests, diagnostic modeling to determine whether adenocarcinoma in tissue specimens is of endometrial or endocervical origin, and models to assess risk for recurrence of invasive breast carcinoma and ductal carcinoma in situ.

CONCLUSIONS

Bayesian network models can be used as powerful and flexible risk assessment tools on large clinical datasets and can quantitatively identify variables that are of greatest significance in predicting specific histopathologic diagnoses and their related prognosis. Resulting BN models are able to provide individualized quantitative risk assessments and prognostication for specific abnormal findings commonly reported in gynecological cytopathology and breast pathology.

Development of a Fully Cross-Validated Bayesian Network Approach for Local Control Prediction in Lung Cancer

The purpose of this study is to demonstrate that a Bayesian network (BN) approach can explore hierarchical biophysical relationships that influence tumor response and predict tumor local control (LC) in non-small-cell lung cancer (NSCLC) patients before and during radiotherapy from a large-scale dataset. Our BN building approach has two steps. First, relevant biophysical predictors influencing LC before and during the treatment are selected through an extended Markov blanket (eMB) method. From this eMB process, the most robust BN structure for LC prediction was found via a wrapper-based approach. Sixty-eight patients with complete feature information were used to identify a full BN model for LC prediction before and during the treatment. Fifty more recent patients with some missing information were reserved for independent testing of the developed pre- and during-therapy BNs. A nested cross-validation (N-CV) was developed to evaluate the performance of the two-step BN approach. An ensemble BN model is generated from the N-CV sampling process to assess its similarity with the corresponding full BN model, and thus evaluate the sensitivity of our BN approach. Our results show that the proposed BN development approach is a stable and robust approach to identify hierarchical relationships among biophysical features for LC prediction. Furthermore, BN predictions can be improved by incorporating during treatment information.

Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

INTRODUCTION

Identifying high-risk lung cancer individuals at an early disease stage is the most effective way of improving survival. The landmark National Lung Screening Trial (NLST) demonstrated the utility of low-dose computed tomography (LDCT) imaging to reduce mortality (relative to X-ray screening). As a result of the NLST and other studies, imaging-based lung cancer screening programs are now being implemented. However, LDCT interpretation results in a high number of false positives. A set of dynamic Bayesian networks (DBN) were designed and evaluated to provide insight into how longitudinal data can be used to help inform lung cancer screening decisions.

METHODS

The LDCT arm of the NLST dataset was used to build and explore five DBNs for high-risk individuals. Three of these DBNs were built using a backward construction process, and two using structure learning methods. All models employ demographics, smoking status, cancer history, family lung cancer history, exposure risk factors, comorbidities related to lung cancer, and LDCT screening outcome information. Given the uncertainty arising from lung cancer screening, a cancer state-space model based on lung cancer staging was utilized to characterize the cancer status of an individual over time. The models were evaluated on balanced training and test sets of cancer and non-cancer cases to deal with data imbalance and overfitting.

RESULTS

Results were comparable to expert decisions. The average area under the curve (AUC) of the receiver operating characteristic (ROC) for the three intervention points of the NLST trial was higher than 0.75 for all models. Evaluation of the models on the complete LDCT arm of the NLST dataset (N=25,486) demonstrated satisfactory generalization. Consensus of predictions over similar cases is reported in concordance statistics between the models' and the physicians' predictions. The models' predictive ability with respect to missing data was also evaluated with the sample of cases that missed the second screening exam of the trial (N=417). The DBNs outperformed comparison models such as logistic regression and naïve Bayes.

CONCLUSION

The lung cancer screening DBNs demonstrated high discrimination and predictive power with the majority of cancer and non-cancer cases.

Bayesian network modeling: A case study of an epidemiologic system analysis of cardiovascular risk

An extensive, in-depth study of cardiovascular risk factors (CVRF) seems to be of crucial importance in the research of cardiovascular disease (CVD) in order to prevent (or reduce) the chance of developing or dying from CVD. The main focus of data analysis is on the use of models able to discover and understand the relationships between different CVRF. In this paper a report on applying Bayesian network (BN) modeling to discover the relationships among thirteen relevant epidemiological features of heart age domain in order to analyze cardiovascular lost years (CVLY), cardiovascular risk score (CVRS), and metabolic syndrome (MetS) is presented. Furthermore, the induced BN was used to make inference taking into account three reasoning patterns: causal reasoning, evidential reasoning, and intercausal reasoning. Application of BN tools has led to discovery of several direct and indirect relationships between different CVRF. The BN analysis showed several interesting results, among them: CVLY was highly influenced by smoking being the group of men the one with highest risk in CVLY; MetS was highly influence by physical activity (PA) being again the group of men the one with highest risk in MetS, and smoking did not show any influence. BNs produce an intuitive, transparent, graphical representation of the relationships between different CVRF. The ability of BNs to predict new scenarios when hypothetical information is introduced makes BN modeling an Artificial Intelligence (AI) tool of special interest in epidemiological studies. As CVD is multifactorial the use of BNs seems to be an adequate modeling tool.

Positive Association of Fibroadenomatoid Change with HER2-Negative Invasive Breast Cancer: A Co-Occurrence Study

Background

Risk assessment of a benign breast disease/lesion (BBD) for invasive breast cancer (IBC) is typically done through a longitudinal study. For an infrequently-reported BBD, the shortage of occurrence data alone is a limiting factor to conducting such a study. Here we present an approach based on co-occurrence analysis, to help address this issue. We focus on fibroadenomatoid change (FAC), an under-studied BBD, as our preliminary analysis has suggested its previously unknown significant co-occurrence with IBC.

Methods

A cohort of 1667 female patients enrolled in the Clinical Breast Care Project was identified. A single experienced breast pathologist reviewed all pathology slides for each case and recorded all observed lesions, including FAC. Fibroadenoma (FA) was studied for comparison since FAC had been speculated to be an immature FA. FA and Fibrocystic Changes (FCC) were used for method validation since they have been comprehensively studied. Six common IBC and BBD risk/protective factors were also studied. Co-occurrence analyses were performed using logistic regression models.

Results

Common risk/protective factors were associated with FA, FCC, and IBC in ways consistent with the literature in general, and they were associated with FAC, FA, and FCC in distinct patterns. Age was associated with FAC in a bell-shape curve so that middle-aged women were more likely to have FAC. We report for the first time that FAC is positively associated with IBC with odds ratio (OR) depending on BMI (OR = 6.78, 95%CI = 3.43-13.42 at BMI<25 kg/m²; OR = 2.13, 95%CI = 1.20-3.80 at BMI>25 kg/m²). This association is only significant with HER2-negative IBC subtypes.

Conclusions

We conclude that FAC is a candidate risk factor for HER2-negative IBCs, and it is a distinct disease from FA. Co-occurrence analysis can be used for initial assessment of the risk for IBC from a BBD, which is vital to the study of infrequently-reported BBDs.

Bayesian networks for clinical decision support in lung cancer care

Survival prediction and treatment selection in lung cancer care are characterised by high levels of uncertainty. Bayesian Networks (BNs), which naturally reason with uncertain domain knowledge, can be applied to aid lung cancer experts by providing personalised survival estimates and treatment selection recommendations. Based on the English Lung Cancer Database (LUCADA), we evaluate the feasibility of BNs for these two tasks, while comparing the performances of various causal discovery approaches to uncover the most feasible network structure from expert knowledge and data. We show first that the BN structure elicited from clinicians achieves a disappointing area under the ROC curve of 0.75 (± 0.03), whereas a structure learned by the CAMML hybrid causal discovery algorithm, which adheres with the temporal restrictions, achieves 0.81 (± 0.03). Second, our causal intervention results reveal that BN treatment recommendations, based on prescribing the treatment plan that maximises survival, can only predict the recorded treatment plan 29% of the time. However, this percentage rises to 76% when partial matches are included.

Company Profile: Strategic Medicine, Inc. and Strategic Medicine, BV

Strategic Medicine, BV (The Hague, The Netherlands) and Strategic Medicine, Inc. (PA, USA) deliver products and services to therapeutic and diagnostic companies, healthcare providers and payers and the investment community based on unique methodologies for modeling disease processes, from predisease through diagnosis, disease and patient stratification and outcome. Strategic Medicine, Inc. focuses on the development of disease models that design and incorporate a personalized health record to model the progression of a patient from genetic risk, through interaction with lifestyle and environment, to early disease detection, disease and patient stratification and treatment decision support for enhanced outcomes. This model involves the integration of disparate data and databases, evaluation of data quality and completeness, data simulation (when necessary), systems modeling and quantitative risk/opportunity evaluation. Strategic Medicine, Inc. has been involved in applying these approaches in oncology, cardiology, women's health and pediatric conditions and rare diseases, and has focused on the development of both disease- and data-agnostic infrastructures and analytics.

Clinical decision support and individualized prediction of survival in colon cancer: bayesian belief network model

BACKGROUND

We used a large population-based data set to create a clinical decision support system (CDSS) for real-time estimation of overall survival (OS) among colon cancer (CC) patients. Patients with CC diagnosed between 1969 and 2006 were identified from the Surveillance Epidemiology and End Results (SEER) registry. Low- and high-risk cohorts were defined. The tenfold cross-validation assessed predictive utility of the machine-learned Bayesian belief network (ml-BBN) model for clinical decision support (CDS).

METHODS

A data set consisting of 146,248 records was analyzed using ml-BBN models to provide CDS in estimating OS based on prognostic factors at 12-, 24-, 36-, and 60-month post-treatment follow-up.

RESULTS

Independent prognostic factors in the ml-BBN model included age, race; primary tumor histology, grade and location; Number of primaries, AJCC T stage, N stage, and M stage. The ml-BBN model accurately estimated OS with area under the receiver-operating-characteristic curve of 0.85, thereby improving significantly upon existing AJCC stage-specific OS estimates. Significant differences in OS were found between low- and high-risk cohorts (odds ratios for mortality: 17.1, 16.3, 13.9, and 8.8 for 12-, 24-, 36-, and 60-month cohorts, respectively).

CONCLUSIONS

A CDSS was developed to provide individualized estimates of survival in CC. This ml-BBN model provides insights as to how disease-specific factors influence outcome. Time-dependent, individualized mortality risk assessments may inform treatment decisions and facilitate clinical trial design.

A probabilistic analysis of completely excised high-grade soft tissue sarcomas of the extremity: an application of a Bayesian belief network

BACKGROUND

It is important to understand the relative importance of prognostic variables in patients with soft tissue sarcomas. The purpose of this study was to describe the hierarchical relationships between features inherent to completely excised, localized high-grade soft tissue sarcomas of the extremity and compare the associations to those previously reported.

METHODS

Data were collected from the Memorial Sloan-Kettering Cancer Center Sarcoma Database. All adult patients with high-grade extremity soft tissue sarcomas who underwent complete excision (R0 margins) at our institution between 1982 and 2010 were included in the analysis. Bayesian belief network (BBN) modeling software was used to develop a hierarchical network of features trained to estimate the likelihood of disease-specific survival. Important relationships depicted by the BBN model were compared to those previously reported.

RESULTS

The records of 1318 consecutive patients met the inclusion criteria, and all were included in the analysis. First-degree associates of disease-specific survival were the primary tumor size; presence of and time to distant recurrence; and presence of and time to local recurrence. On cross-validation, the BBN model was sufficiently robust, with an area under the curve of 0.94 (95 % confidence interval 0.93-0.96).

CONCLUSIONS

We successfully described the hierarchical relationships between features inherent to patients with completely excised high-grade soft tissue sarcomas of the extremity. The relationships defined by the BBN model were similar to those previously reported. Cross-validation results were encouraging, demonstrating that BBN modeling can be used to graphically illustrate the complex hierarchical relationships between prognostic features in this setting.

DW4TR: A Data Warehouse for Translational Research

The linkage between the clinical and laboratory research domains is a key issue in translational research. Integration of clinicopathologic data alone is a major task given the number of data elements involved. For a translational research environment, it is critical to make these data usable at the point-of-need. Individual systems have been developed to meet the needs of particular projects though the need for a generalizable system has been recognized. Increased use of Electronic Medical Record data in translational research will demand generalizing the system for integrating clinical data to support the study of a broad range of human diseases. To ultimately satisfy these needs, we have developed a system to support multiple translational research projects. This system, the Data Warehouse for Translational Research (DW4TR), is based on a light-weight, patient-centric modularly-structured clinical data model and a specimen-centric molecular data model. The temporal relationships of the data are also part of the model. The data are accessed through an interface composed of an Aggregated Biomedical-Information Browser (ABB) and an Individual Subject Information Viewer (ISIV) which target general users. The system was developed to support a breast cancer translational research program and has been extended to support a gynecological disease program. Further extensions of the DW4TR are underway. We believe that the DW4TR will play an important role in translational research across multiple disease types.

Probabilistic (Bayesian) modeling of gene expression in transplant glomerulopathy

Transplant glomerulopathy (TG) is associated with rapid decline in glomerular filtration rate and poor outcome. We used low-density arrays with a novel probabilistic analysis to characterize relationships between gene transcripts and the development of TG in allograft recipients. Retrospective review identified TG in 10.8% of 963 core biopsies from 166 patients; patients with stable function were studied for comparison. The biopsies were analyzed for expression of 87 genes related to immune function and fibrosis by using real-time PCR, and a Bayesian model was generated and validated to predict histopathology based on gene expression. A total of 57 individual genes were increased in TG compared with stable function biopsies (P < 0.05). The Bayesian analysis identified critical relationships between ICAM-1, IL-10, CCL3, CD86, VCAM-1, MMP-9, MMP-7, and LAMC2 and allograft pathology. Moreover, Bayesian models predicted TG when derived from either immune function (area under the curve [95% confidence interval] of 0.875 [0.675 to 0.999], P = 0.004) or fibrosis (area under the curve [95% confidence interval] of 0.859 [0.754 to 0.963], P < 0.001) gene networks. Critical pathways in the Bayesian models were also analyzed by using the Fisher exact test and had P values <0.005. This study demonstrates that evaluating quantitative gene expression profiles with Bayesian modeling can identify significant transcriptional associations that have the potential to support the diagnostic capability of allograft histology. This integrated approach has broad implications in the field of transplant diagnostics.