Motivation: It has been proposed that clustering clinical markers, such as blood test results, can be used to stratify patients.\r\nHowever, the robustness of clusters formed with this approach to data pre-processing and clustering algorithm choices has not\r\nbeen evaluated, nor has clustering reproducibility. Here, wemade use of the NHANES survey to compare clusters generated with\r\nvarious combinations of pre-processing and clustering algorithms, and tested their reproducibility in two separate samples.\r\nMethod: Values of 44 biomarkers and 19 health/life style traits were extracted from the National Health and Nutrition\r\nExamination Survey (NHANES). The 1999ââ?¬â??2002 survey was used for training, while data from the 2003ââ?¬â??2006 survey was\r\ntested as a validation set. Twelve combinations of pre-processing and clustering algorithms were applied to the training set.\r\nThe quality of the resulting clusters was evaluated both by considering their properties and by comparative enrichment\r\nanalysis. Cluster assignments were projected to the validation set (using an artificial neural network) and enrichment in\r\nhealth/life style traits in the resulting clusters was compared to the clusters generated from the original training set.\r\nResults: The clusters obtained with different pre-processing and clustering combinations differed both in terms of cluster\r\nquality measures and in terms of reproducibility of enrichment with health/life style properties. Z-score normalization, for\r\nexample, dramatically improved cluster quality and enrichments, as compared to unprocessed data, regardless of the\r\nclustering algorithm used. Clustering diabetes patients revealed a group of patients enriched with retinopathies. This could\r\nindicate that routine laboratory tests can be used to detect patients suffering from complications of diabetes, although\r\nother explanations for this observation should also be considered.\r\nConclusions: Clustering according to classical clinical biomarkers is a robust process, which may help in patient\r\nstratification. However, optimization of the pre-processing and clustering process may be still required.
Loading....