Classification of premalignant pancreatic cancer mass-spectrometry data using decision tree ensembles

This study explores the utility of three different feature selection schemas (Student t test, Wilcoxon rank sum test and genetic algorithm) to reduce the high dimensionality of a pancreatic cancer proteomic dataset. Using the top features selected from each method, we compared the prediction performances of a single decision tree algorithm C4.5 with six different decision-tree based classifier ensembles (Random forest, Stacked generalization, Bagging, Adaboost, Logitboost and Multiboost). We show that ensemble classifiers always outperform single decision tree classifier in having greater accuracies and smaller prediction errors when applied to a pancreatic cancer proteomics dataset.In our cross validation framework, classifier ensembles generally have better classification accuracies compared to that of a single decision tree when applied to a pancreatic cancer proteomic dataset, thus suggesting its utility in future proteomics data analysis. Additionally, the use of feature selection method allows us to select biomarkers with potentially important roles in cancer development, therefore highlighting the validity of this method.Pancreatic cancer is one of the most lethal types of cancer. In United States, there are ~30,000 new cases being diagnosed each year. The mortality rate of pancreatic cancer patients is approaching 100%. Only 4% of the patients survive 5 years or more after being diagnosed. The grim statistics of pancreatic cancer necessitates the urgent development of methods to facilitate their early detection and prevention [1]. Despite the advancement of our knowledge in recent years regarding the pathophysiology of pancreatic cancer [2,3], we still lack an effective method to diagnose this cancer type early enough to impact the treatment outcomes.Recently, there has been substantial interests in applying proteomics technology to identify clinically useful biomarkers for early-stage pancreatic cancer [4-11]. In a more general sense, many investigators hav