Meet Inspiring Speakers and Experts at our 3000+ Global Conference Series LLC LTD Events with over 1000+ Conferences, 1000+ Symposiums
and 1000+ Workshops on Medical, Pharma, Engineering, Science, Technology and Business.

Explore and learn more about Conference Series LLC LTD : World’s leading Event Organizer

Back

25th World Congress on Cancer Science and Therapy

Baltimore, USA

Frédéric Baribaud

Frédéric Baribaud

Johnson & Johnson, USA

Title: Molecular surrogates of histologic activity in crohn’s disease

Biography

Biography: Frédéric Baribaud

Abstract

Objective markers of disease severity in inflammatory bowel disease that support clinical decision-making are still needed. We hypothesized that novel objective markers of tissue inflammation are best identified at the site of disease with a tissue-level assessment of disease activity. Biopsy samples were obtained from participants in the UNITI trials of ustekinumab in moderate-to-severe Crohn’s disease. Pairs of adjacent biopsies were taken from the rectum, splenic flexure and ileum. One biopsy from each pair was assessed by global histology disease activity score (GHAS) while the other was submitted for microarray analysis. Partial least squares regression and random forest were used to identify biomarkers associated with histological severity and robustness of the resulting models was assessed using cross-validation. A single multivariate model comprising 16 genes was identified that predicted histological activities in rectum or splenic flexure biopsies. This model was characterized by R2=0.78 for the training set, and R2=0.59, 0.54 and 0.32 on external validation sets. A separate 14-gene model capturing histological activity in ileal biopsies was characterized by R2=0.5 for the training set and R2=0.45 in the external validation set. In general, both models contained genes related to tissue degradation, barrier function and immune regulation, including CXCL11 (I-TAC). Both models retained performance in external validation datasets from UNITI-2 but exhibited lower performance. Our analysis supports the ability of biopsy transcriptomics combined with machine learning approaches, to capture disease-relevant variability in Crohn’s disease and more importantly, supports the use of similar approaches to identify additional surrogate markers.