Gene Expression Analysis Of Diabetic Foot Ulcer Debrided Tissue For Predicting Responsiveness To Treatments
Jessica M. Eager1, Michael S. Weingarten, MD2, Will Dampier, PhD3, Kara L. Spiller, PhD1.
1School of Biomedical Engineering, Science and Health Systems, Drexel University, Philadelphia, PA, USA, 2Department of Surgery, Drexel University College of Medicine, Philadelphia, PA, USA, 3Department of Microbiology and Immunology, Drexel University College of Medicine, Philadelphia, PA, USA.
BACKGROUND Chronically open wounds are a common complication for diabetic patients. These ulcers (DFUs) are notoriously difficult to treat for a variety of reasons, but the major challenge stems from the mystery as to why some wounds respond to treatment and others do not. The goal of our study is to create a diagnostic assay that will determine the likelihood a DFU wound will close by 12 weeks in response to the standard of care. METHODS We used Nanostring™ for gene expression analysis of 227 wound healing and macrophage phenotype related genes in debrided tissue from the first visit to the clinic of 23 patients. The top 10 most highly expressed genes were used to train a neural network algorithm to classify healing outcome at 12 weeks as fully closed, remained open, or required amputation. 10-fold cross validation was performed to determine the best number of hidden units. The dataset was then divided into training and testing cohorts using a 60/40 split. RESULTS After training, the algorithm correctly classified the outcome for 7 of the 10 patients, resulting in 80% specificity and 60% sensitivity when the outcomes are binarized by combining the remained open and required amputation groups. CONCLUSIONS These results, while preliminary, suggest that gene expression analysis of debrided tissue from the first visit can be used as a method of predicting healing outcome for DFUs. Because this method relies solely on the first visit, it has the potential to guide patients and clinicians toward treatment options better suited to the wound environment and more likely to facilitate healing.
Back to 2019 Abstracts