SP 4
Combining MRI, SNPs, FISH and GEP/RNAseq in improving risk prediction and treatment decision making
Multiple Myeloma is a very heterogeneous disease with complex treatment regimens. For that reason it would be very beneficial for the individual patients to be able to predict survival, response to treatment or therapy-limiting side effects, and hence to adapt therapy accordingly. The current clinical practice in multiple myeloma to stratify patients into different risk groups uses different data sources, like patient characteristics, laboratory parameters, cytogenetic changes and gene expression markers. However, precisely because multiple myeloma is very complex, each of those sources of information on its own is quite limited in predicting the course of disease.
The major goal of this part of the project is to combine prognostic information from all available data sources to provide a more complete characterization of the individual patient and the predicted course of disease and response to therapy. For that purpose, in a first step all known risk factors are considered in a joint prognostic model. Then, with all available data an integrative data analysis is conducted, while accounting for biological and statistical relations between different data sources. Methods specifically developed for the analysis of high-dimensional data, like regularized regression models for survival time data, are applied. The integration of several different high-throughput datasets is especially complex. Further, we will look for treatment specific factors relevant for side effects and therapeutic response. This is only possible, since the majority of patients was treated in large randomized clinical trials.
Importantly, with the increasing amount of information available to clinicians, providing all information in a comprehensive report is essential to help medical decision making. Therefore, an analysis protocol will be developed, that combines statistical results and text processing in an automated way and arranges all relevant information within an individual CLIOMMICS patient report.
Keywords: multiple myeloma, prediction of course of disease, integrative analysis