P222 Raman Spectroscopy in the diagnosis of pIBD: experimental study
Viola, I.(1);Costa, S.(1);Pellegrino , S.(1);Acri, G.(2);Romano, C.(1);
(1)University of Messina - "G.Martino" Hospital, Pediatric Gastroenterology and Cystic Fibrosis Unit- Department of Human Pathology in Adulthood and Childhood, Messina, Italy;(2)University of Messina - "G.Martino" Hospital, Department of BIOMORF, Messina, Italy;
Background
Diagnostic delay is a major issue among the paediatric Inflammatory Bowel Disease (pIBD) population because of the association with higher risk of complications and growth failure. At present, non-invasive serological and faecal biomarkers lack sufficient sensibility and/or sensitivity to be used as a diagnostic tool in clinical practice and none can discriminate between Crohn's Disease (CD) and Ulcerative Colitis (UC). In this setting, new non-invasive diagnostic tools with better diagnostic performance are needed.Raman Spectroscopy (RS) is a spectroscopic technique that relies on the inelastic scattering process of monochromatic light. Through the analysis of spectra molecular changes can be identified.The overall objective of this study was to explore RS performance as a non-invasive diagnostic test in IBD diagnosis.
Methods
RS was applied to the analysis of stool protein extracts in pIBD patients and a control group. Being directly in contact with the intestine stool can provide information on the bacterial and inflammatory background and represents an excellent sample material to study. At our Paediatric Gastroenterology Unit 55 patients (age 0-18) were enrolled, 32 cases (20 MC, 2 Very Early Onset-IBD, 9 CU, 2014 Porto criteria), 23 controls (functional constipation, irritable bowel syndrome, Rome IV criteria). Protein extract from faecal samples (obtained by the same kit test used to evaluate faecal calprotectin) were analysed with a DXR- Smart Raman Spectrometer (thermos Fisher Scientific, Waltham, MA, USA) with a 785 nm diode laser. After obtaining a spectrum the analysis focused on the band in the wavenumber range of about 1550-1750 cm-1 which defines the amide-I vibration modes, deconvolution of this band was performed. Ratio of the area of the sub-bands, parameter R, was calculated. Receiver Operating Characteristic (ROC) Curve and Youden Index (J) were used for statistical analysis.
Results
As represented by the ROC curve R showed a great diagnostic accuracy in differentiating between cases and controls, AUC 0,9389 (0,87-1,00) figure 1. A cut-off value of 1,2 distinguished cases from controls with sensibility 0,9 (0,78-0,95), specificity 1 (0,85-1), figure 2. 
In the group of cases R distinguished CD from UC, AUC 0,83 (0,65-1) figure 3. A cut of value of 1,4 distinguished CD from UC with sensibility 1 (0,69-1), specificity 0,83 (0,58-0,96) figure 4.
Conclusion
This is the first application of RS as a non-invasive diagnostic tool on stool samples. Preliminary data show that RS may recognize IBD from non-IBD features, and, what is completely new, it can discriminate the two main forms of IBD, CD and UC. Future prospective studies including patients with suspected IBD will help defining the diagnostic performance.