The majority of Dr. Goldgof’s work focuses on the identification of novel biomarker strategies using genomics and other emerging technologies to guide clinical decision making and the identification and implementation of molecular imaging approaches to guide clinical decision making.His team’s work is related to several directions: (a) medical image analysis for identification of image features related to cancer detection and progress as well as novel molecular imaging technologies, (b) data mining methods to discover gene features related to cancer, and (c) combination of anatomical imaging features with gene features for improved outcome prediction with recent concentration on (c). Dr. Goldgof and colleagues investigated the impact on classification accuracy of gene selection approaches on filtered-to-200-gene datasets, using four datasets with 3 filters: t-test, information gain, and reliefF. They applied Iterative Feature Perturbation (IFP) and Recursive Feature Elimination (SVM-RFE) for further gene selection. They conducted a statistical analysis of accuracy across the best 50 genes using the Friedman/Holm test, which showed that IFP and SVMRFE were significantly more accurate more often when applied to the t-test-filtered gene sets. Surprisingly, the simple t-test, applied as a filter, results in the best overall SVM accuracy and is at least as accurate as the other, more complicated filter methods.