· · · Repository hosted by TU Delft Library

We investigate the potential to enhance breast cancer event predictors by exploiting subtype information. We do this with a two-stage approach that first determines a sample's subtype using a recent module-driven approach, and secondly constructs a subtype-specific predictor to predict a metastasis event within five years. Our methodology is validated on a large compendium of microarray breast cancer datasets,including 43 replicate array pairs for assessing subtyping stability. Note that stratifying by subtype strongly reduces the training set sizes available to construct the individual predictors, which may decrease performance. Besides sample size, other factors likeunequal class distributions and differences in the number of samplesper subtype, easily obscure a fair comparison between subtype-specific predictors constructed on different subtypes, but also between subtype specific and subtype a-specific predictors. Therefore, we constructed a completely balanced experimental design, in which none ofthe above factors play a role and show that subtype-specific eventpredictors clearly outperform predictors that do not take subtype information into account.

Motivation: In recent years increasing evidence appeared that breastcancer may not constitute a single disease at the molecular level,but comprises a heterogeneous set of subtypes. This suggests that instead of building a single predictor, better predictors might be constructed that solely target samples of a designated subtype. An unavoidable drawback of developing subtype-specific predictors, however,is that a stratification by subtype drastically reduces the numberof samples available for their construction. It is therefore questionable whether the potential benefit of subtyping can outweigh the drawback of a severe loss in sample size. Factors like unequal class distributions and differences in the number of samples per subtype, further complicate comparisons. Results: We present several evaluation strategies that facilitate a comprehensive comparison between subtype-specific predictors and predictors that do not take subtype information into account. Emphasis lies on careful control of sample size as well as class and subtype distributions. The methodology is applied to a large breast cancer compendium involving over 1500 arrays,using a state-of-the-art subtyping scheme. We show that the resulting subtype-specific predictors outperform those that do not take subtype information into account, especially when taking sample size considerations into account.

Current neonatal illness scoring systems are not designed to predictoutcomes for individual patients, but rather can provide an overview of a population of patients for objective comparison when reporting outcomes. Having more patient-specific predictions may help physicians make better treatment decisions in a Neonatal Intensive Care Unit (NICU) environment. We developed neonatal mortality prediction models using C5.0 decision tree software that met criteria for clinically useful results (>50-60% sensitivity, >90% specificity) for individual patients using data from real-time medical measurement devices. The models were evaluated to identify: (1) the model with the bestperformance based on minimizing false positives, and (2) the attributes used most often in the best clinically useful models. Performance results showed that the mortality model using summary data duringthe first 48 hours after NICU admission provided, on average, the highest sensitivity and specificity with the least number of false positives (sensitivity=63%, specificity=94%, positive predictive value=38%), exceeding the performance criteria requested by our clinicalpartners. The attributes used most often in the best models for predicting mortality with our data were: mean blood pressure, serum pH,immature/total neutrophil ratio, serum sodium, serum glucose, respiratory rate, heart rate, and pO2 blood oxygen level.

AbstractMore than half a million people in Europe live on chronic renal re-placement therapy by hemodialysis (HD). Though arteriovenous fistulas are the preferred vascular access (VA) method of connecting a patient’s circulation to the artificial kidney, VA dysfunction is the major cause of morbidity and hospitalization in HD patients. The ARCHproject has been initiated to tackle this problem by delivering clinical decision support for VA surgery.In this report we describe the stakeholders and user scenarios for models and tools being developed by the ARCH project, the consolidated requirements for services of software infrastructure, proposed a portfolio of services to be realized. We also summarize the Europeanvascular access clinical guidelines, and describe four different current VA surgery clinical workflows, and propose a number of ways ofintegrating the computational tools developed by the ARCH consortiuminto the current practice.ConclusionsIn this report, we described the stakeholders and user scenarios formodels and tools being developed by the ARCH project, the consolidated requirements for services of software infrastructure, proposed aportfolio of services to be realized.We defined two main user scenario categories: the ARCH researcher category and the clinical practitioner category. The scenarios withinthe ARCH researcher category illustrate the definition of clinical protocols, development, in-vitro and clinical validation of modelingtools for vascular access (VA), while the scenarios in the clinicalpractitioner category illustrate the use of the validated tools in clinical practice. We defined the Research Information Management Infrastructure (RIMI) to support the users in ARCH researcher user scenarios, and the Clinical Information Management Infrastructure (CIMI)to support the users in clinical practitioner user scenarios.An important development has been the adoption of the VPH euHeart project of the ARCH client/server infrastructure design and XML data representation for data collection, along with the actual software appli-cation to be adapted to the project’s specific requirements.TheRIMI and the CIMI will provide domain-specific services that willbe tailored to the intended users of the infrastructure. These domain-specific services, will be realized through the ARCH technicalcomputational and data basic services.Most RIMI services are already in place and being used by the project members. These services include: a project Twiki web for documentexchange; an ARCH server for storage of large data sets and an ARCHclient to facilitate data storage and exchange; a GForge server forsoft-ware exchange; an OpenClinica server for managing Case Report Forms.Regarding the proposed CIMI services, we are now in the process of validating our design of the clinical graphical interface (CGI) of the clinical application that will assist the clinicians in performingthe VA function prediction simulations. Some screenshots are shownin this report and initial feedback of the clinicians has been favorable. At the same time, we are creating the basic data and computational technical services that will allow the realization of the clinical application itself.