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 novel nomogram for peritoneal mesothelioma predicts survival.

BACKGROUND: Malignant peritoneal mesothelioma (MPM) is a rare disease treated with cytoreductive surgery (CRS) and hyperthermic intraperitoneal chemotherapy (HIPEC). Estimation of personalized survival times can potentially guide treatment and surveillance. METHODS: We analyzed 104 patients who underwent CRS and cisplatin-based HIPEC for MPM. By means of 25 demographic, laboratory, operative, and histopathological variables, we developed a novel nomogram using machine-learned Bayesian belief networks with stepwise training, testing, and cross-validation. RESULTS: The mean peritoneal carcinomatosis index (PCI) was 15, and 66 % of patients had a completeness of cytoreduction (CC) score of 0 or 1. Eighty-seven percent of patients had epithelioid histology. The median follow-up time was 49 (1-195) months. The 3- and 5-year overall survivals (OS) were 58 and 46 %, respectively. The histological subtype, pre-CRS PCI, and preoperative serum CA-125 had the greatest impact on OS and were included in the nomogram. The mean areas under the receiver operating characteristic curve for the 10-fold cross-validation of the 3- and 5-year models were 0.77 and 0.74, respectively. The graphical calculator or nomogram uses color coding to assist the clinician in quickly estimating individualized patient-specific survival before surgery. CONCLUSIONS: Machine-learned Bayesian belief network analysis generated a novel nomogram predicting 3- and 5-year OS in patients treated with CRS and HIPEC for MPM. Pre-CRS estimation of survival times may potentially individualize patient care by influencing the use of systemic therapy and frequency of diagnostic imaging, and might prevent CRS in patients unlikely to achieve favorable outcomes despite surgical intervention.

Time-dependent estimates of recurrence and survival in colon cancer: clinical decision support system tool development for adjuvant therapy and oncological outcome assessment.

Unanswered questions remain in determining which high-risk node-negative colon cancer (CC) cohorts benefit from adjuvant therapy and how it may differ in an equal access population. Machine-learned Bayesian Belief Networks (ml-BBNs) accurately estimate outcomes in CC, providing clinicians with Clinical Decision Support System (CDSS) tools to facilitate treatment planning. We evaluated ml-BBNs ability to estimate survival and recurrence in CC. We performed a retrospective analysis of registry data of patients with CC to train-test-crossvalidate ml-BBNs using the Department of Defense Automated Central Tumor Registry (January 1993 to December 2004). Cases with events or follow-up that passed quality control were stratified into 1-, 2-, 3-, and 5-year survival cohorts. ml-BBNs were trained using machine-learning algorithms and k-fold crossvalidation and receiver operating characteristic curve analysis used for validation. BBNs were comprised of 5301 patients and areas under the curve ranged from 0.85 to 0.90. Positive predictive values for recurrence and mortality ranged from 78 to 84 per cent and negative predictive values from 74 to 90 per cent by survival cohort. In the 12-month model alone, 1,132,462,080 unique rule sets allow physicians to predict individual recurrence/mortality estimates. Patients with Stage II (N0M0) CC benefit from chemotherapy at different rates. At one year, all patients older than 73 years of age with T2-4 tumors and abnormal carcinoembryonic antigen levels benefited, whereas at five years, all had relative reduction in mortality with the largest benefit amongst elderly, highest T-stage patients. ml-BBN can readily predict which high-risk patients benefit from adjuvant therapy. CDSS tools yield individualized, clinically relevant estimates of outcomes to assist clinicians in treatment planning.

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.

Application of multivariate probabilistic (Bayesian) networks to substance use disorder risk stratification and cost estimation.

INTRODUCTION: This paper explores the use of machine learning and Bayesian classification models to develop broadly applicable risk stratification models to guide disease management of health plan enrollees with substance use disorder (SUD). While the high costs and morbidities associated with SUD are understood by payers, who manage it through utilization review, acute interventions, coverage and cost limitations, and disease management, the literature shows mixed results for these modalities in improving patient outcomes and controlling cost. Our objective is to evaluate the potential of data mining methods to identify novel risk factors for chronic disease and stratification of enrollee utilization, which can be used to develop new methods for targeting disease management services to maximize benefits to both enrollees and payers. METHODS: For our evaluation, we used DecisionQ machine learning algorithms to build Bayesian network models of a representative sample of data licensed from Thomson-Reuters' MarketScan consisting of 185,322 enrollees with three full-year claim records. Data sets were prepared, and a stepwise learning process was used to train a series of Bayesian belief networks (BBNs). The BBNs were validated using a 10 percent holdout set. RESULTS: The networks were highly predictive, with the risk-stratification BBNs producing area under the curve (AUC) for SUD positive of 0.948 (95 percent confidence interval [CI], 0.944-0.951) and 0.736 (95 percent CI, 0.721-0.752), respectively, and SUD negative of 0.951 (95 percent CI, 0.947-0.954) and 0.738 (95 percent CI, 0.727-0.750), respectively. The cost estimation models produced area under the curve ranging from 0.72 (95 percent CI, 0.708-0.731) to 0.961 (95 percent CI, 0.95-0.971). CONCLUSION: We were able to successfully model a large, heterogeneous population of commercial enrollees, applying state-of-the-art machine learning technology to develop complex and accurate multivariate models that support near-real-time scoring of novel payer populations based on historic claims and diagnostic data. Initial validation results indicate that we can stratify enrollees with SUD diagnoses into different cost categories with a high degree of sensitivity and specificity, and the most challenging issue becomes one of policy. Due to the social stigma associated with the disease and ethical issues pertaining to access to care and individual versus societal benefit, a thoughtful dialogue needs to occur about the appropriate way to implement these technologies.