Integrated Molecular and Clinical Profiling to Optimize Outcome Prediction in Splenic Marginal Zone Lymphoma

The study aims at developing and validating an integrated clinico-molecular model for an accurate survival prognostication in newly diagnosed SMZL.

Already existing and coded tumor biological material and health-related personal data will be retrospectively collected. Mutation analysis will be performed by targeted deep next generation sequencing of tumor genomic DNA. Deletion of 7q will be assessed by FISH on nuclei isolated from tumor tissues. Immunoglobulin gene rearrangement and mutation status will be analyzed on tumor genomic DNA by PCR and Sanger sequencing. The methylation status of target genes will be assessed by methylation specific PCR on tumor genomic DNA.

The adjusted association between exposure variables and OS will be estimated by Cox regression. This approach will provide the covariates independently associated with OS that will be utilized in the development of a hierarchical molecular model to predict OS. The hierarchical order of relevance in predicting OS among covariates will be established by recursive partitioning analysis. An amalgamation algorithm will be used to merge terminal nodes showing homogenous OS. The discrimination capacity of the model will be assessed by calculating the c-index. Relative survival analysis will be used to provide a measure of the excess mortality experienced by patient's subgroups stratified according to the developed hierarchical molecular models, irrespective of whether the excess mortality is directly or indirectly attributable to the disease. The model developed in the training set will be tested in the validation sets and the model performance (c-index) in the validation set will be compared with that in the training set.