P297 Novel serum soluble markers of histological healing in ulcerative colitis
O.M. Nardone1, L. Jeffery1, A. Bazarova2, A. Acharjee3, G. Gkoutos3, D. Zardo4, U.N. Shivaji1, S.C. Smith1, S. Ghosh1, M. Iacucci1
1Institute of Translational Medicine and Institute of Immunology and Immunotherapy, University of Birmingham, Birmingham, UK, 2Institute for Biological Physics, University of Cologne, Cologne, Germany, 3Cancer and Genomic Science, University of Birmingham, Birmingham, UK, 4University Department of Pathology, Hospitals Birmingham NHS Foundation Trust, Birmingham, UK
Background
The recent introduction of a commercial panel of soluble biomarkers, that can predict endoscopic healing, may support the monitoring and management of Crohn’s disease reducing the need of repeated endoscopy.1 Achieving histological remission (HR) is currently considered to be an important target in ulcerative colitis (UC).2 The aim of this study was to identify serological markers of HR in patients with UC and determine their diagnostic accuracy individually and in combination to predict HR.
Methods
We enrolled 40 UC patients referred to a tertiary academic centre for colonoscopy as part of standard of care. Demographic data and histologic data (RHI, Nancy) were collected. Concentrations of these biomarkers were compared with the histology at the time that the patients underwent colonoscopy with advanced enhancement technologies and targeted biopsies. Histological healing was defined as Nancy ≤1 and RHI ≤3. A total of 55 soluble analytes, relevant to inflammation or shown to be altered in IBD, were measured in serum using Procartaplex luminex assays (ThermoFisher). Finally, 24 analytes that were detected in more than 40% of patients were selected in downstream modelling and used to train a logistic regression model.
Results
We identified Brain-Derived Neurotrophic Factor (BDNF) and Macrophage Inflammatory Proteins (MIP-1 α) that, when combined together, showed the highest predictive power for predicting RHI≤3 and Nancy ≤ 1 with an area under the receiver operating characteristic curve (AUROC) of 0.82 (95% CI: 0.69–0.97). When we investigated analytes separately for predicting the same outcome, univariate logistic regression was significant for two other markers: Leukemia Inhibitory Factor (LIF) and Vascular Cell Adhesion Molecule 1 (sVCAM1) showing higher values associated with the lack of healing. The AUROC was 0.74 for LIF (95% CI: 0.591–0.89) and 0.72 for sVCAM1 (95% CI: 0.53–0.90).
Conclusion
BDNF, MIP-1α, LIF and sVCAM may be promising markers to assess histological healing in UC patients. Replication in larger cohorts of patients is now required.
*OMN and LJ contributed equally.
D’Haens G,
Nardone OM,