Association for Academic Surgery Rapid detection of Clostridium difficile toxins in serum by Raman spectroscopy S. Kiran Koya, PhD, a Michelle Brusatori, PhD, a Jonathan V. Martin, MD, b Sally Yurgelevic, MS, a Changhe Huang, PhD, a David M. Liberati, MS, b Gregory W. Auner, PhD, a and Lawrence N. Diebel, MD b, * a Smart Sensors and Integrated Microsystems, Michael and Marian Ilitch Department of Surgery, School of Medicine, Wayne State University, Detroit, Michigan b Michael and Marian Ilitch Department of Surgery, School of Medicine, Wayne State University, 6C University Health Center, Detroit, Michigan article info Article history: Received 2 March 2018 Received in revised form 5 May 2018 Accepted 8 June 2018 Available online xxx Keywords: Clostridium difficile toxin A and B Systemic CDI Raman spectroscopy Rapid detection Diagnostic test abstract Background: Clostridium difficile infection (CDI) is due to the effects of toxins, toxin A and toxin B on the host. Severe CDI is associated with systemic signs of infection. Animal models of CDI demonstrate a strong correlation between systemic toxemia and the occurrence of severe disease. However, current technologies have low sensitivity to detect C difficile toxemia in human subjects. Raman spectroscopy (RS) is an upcoming technology that is used to detect bacteria and their toxins. We speculate that RS may be a sensitive method to detect clinically relevant concentrations of C difficile toxins in serum. Materials and methods: Serum samples were spiked with varying concentrations of toxin A, toxin B, and both. RS was performed on an air-dried serum drop that was placed on a mirror-polished stainless steel slide. Raman spectra were obtained, background corrected, vector normalized, and analyzed by Partial Least Square Linear Discriminant Analysis and Support Vector Machine for Classification. Model accuracy was measured by cross- validation and bootstrap methods. Results: Toxin-spiked sera of various concentrations (1 ng/mL, 1 pg/mL, and 0.1 pg/mL) were distinguished from control serum 100% with cross-validation error rate ranging from 0% to 18% and bootstrap error rate ranging from 0% to 12% for various concentrations. The sensitivity ranged from 87% to 100% and specificity ranged from 77% to 100% for various concentrations of toxin-spiked serum. Conclusions: We conclude that RS may be a sensitive method to detect clinically relevant concentrations of C difficile toxins in serum and thus to help diagnose severe CDI in patients in real-time at the point of care. ª 2018 Published by Elsevier Inc. * Corresponding author. Michael and Marian Ilitch Department of Surgery, 6C University Health Center, 4201 Saint Antoine, Detroit, MI 48201. Tel.: 313-577-5005; fax: 313-577-5310. E-mail address: ldiebel@med.wayne.edu (L.N. Diebel). Available online at www.sciencedirect.com ScienceDirect journal homepage: www.JournalofSurgicalResearch.com journal of surgical research  december 2018 (232) 195 e201 0022-4804/$ e see front matter ª 2018 Published by Elsevier Inc. https://doi.org/10.1016/j.jss.2018.06.031