· · · 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.

The rapidly growing field of targeted diagnostics is serving an increasing spectrum of mature clinical applications, supported by the continuous development of new facilitating equipment. The wire grid microarray is a novel concept for targeted nucleic acid and protein analysis, developed at Philips Research. This thesis demonstrates theversatility of the wire grid microarray regarding its compatibilitywith a variety of optical imaging systems, in particular: A compactwide-field fluorescence imaging setup, composed of off-the-shelf components, a sophisticated wide-field imaging setup with a highly sensitive CCD camera, and a confocal laser scanning fluorescence imagingsetup. Using these systems, it is shown that the wire grid microarray allows the detection of genomic nucleic acids, and a real-time monitoring of the hybridization of synthetic nucleic acids, on a surface. This provides valuable kinetic information about the hybridization process. The main enabling feature of the wire grid microarray isthe strong suppression of the background fluorescence. The suppression of the background is shown to be at least 5 times higher than for conventional confocal scanning.

Real Time Array PCR is a recently developed biochemical technique that measures amplification curves (like quantitative real time Polymerase Chain Reaction (qPCR)) of a multitude of different templates ina sample. It combines two different techniques to profit from theadvantages of both techniques, namely qPCR (real time quantitative detection) with microarrays (high multiplex capability). This enablesthe quantitative detection of many more target sequences than can be done by qPCR. Thereby, the concentration of the many different target molecules originally present in a sample can be measured. Labeled primers are used that are first elongated to form labeled amplicons in the bulk and these can hybridize to capture probes immobilizedon the surface of the microarray. During each PCR cycle, there is atime window available during which the formed labeled amplicons canhybridize to the target sequences on the microarray surface. By detection of the fluorescence of the spots on the microarray, amplification curves comparable to real time PCR can be obtained, which can be used to deduce the information needed on the presence and the amount of targets originally present in the sample. We present a mathematical model that provides fundamental insights in the different steps of Real Time Array PCR and that can be used to optimize the different biochemical processes taking place. At the microarray surface specific molecules are captured and taken away from the solution, causing a concentration gradient that powers a material flow towards themicroarray surface. Only the labeled strand of the amplicon is captured by the probes on the microarray surface and as a result locallythe PCR process is not symmetric anymore. Moreover, in course of the process more and more ssDNA renatures, leaving relatively less strands and complexes available for hybridization. We found that to a large extent, however, the surface fluorescence scales with the bulkconcentration. Important parameters to optimize are the enzyme concentration and degradation, the primer concentration and the capture probe decay rate. Also the surface hybridization time is critical since the time to reach a steady state is at least one order of magnitude longer compared to the timing of the bulk processes in qPCR.

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.