Assessment of the validity of a multigene analysis in the diagnostics of inflammatory bowel disease

• Published in: Journal of internal medicine 2014-05, Vol.275 (5), p.484-493
• Main Authors: Bjerrum, J. T., Nyberg, C., Olsen, J., Nielsen, O. H.
• Format: Article
• Language: English
• Subjects: Adult; Amidohydrolases - genetics; Amino Acid Transport Systems, Neutral - genetics; Anion Transport Proteins - genetics; Area Under Curve; Chemokine CXCL1 - genetics; Colitis, Ulcerative - diagnosis; Colitis, Ulcerative - genetics; colonic mucosal biopsy; Crohn Disease - diagnosis; Crohn Disease - genetics; Crohn's disease; diagnostics; Female; gene expression; Gene Expression - genetics; Genetic Markers - genetics; GPI-Linked Proteins - genetics; Humans; Lectins, C-Type - genetics; Male; Matrix Metalloproteinase 7 - genetics; Membrane Proteins - genetics; Pancreatitis-Associated Proteins; Prospective Studies; Real-Time Polymerase Chain Reaction; ulcerative colitis
• ISSN: 0954-6820; 1365-2796
• DOI: 10.1111/joim.12160

Objectives The findings of a previous multigene study indicated that the expression of a panel of seven specific genes had strong differential power regarding inflammatory bowel disease (IBD) versus non‐IBD, as well as ulcerative colitis (UC) versus Crohn's disease (CD). This prospective confirmatory study based on an independent patient cohort from a national Danish IBD centre was conducted in an attempt to verify these earlier observations. Design, setting and participants A total of 119 patients were included in the study (CD, UC and controls). Three mucosal biopsies were retrieved from the left side of the colon of each patient. RNA was extracted, and RT‐PCR was performed to retain expression profiles from the seven selected genes. Expression data from the training set (18 CD, 20 UC and 20 controls) were used to build a classification model, using quadratic discriminant analysis, and data from the test set (20 CD, 21 UC and 20 controls) were used to test the validity of the model. Results The present investigation did not confirm the previous observation that a panel of seven specific genes is able to distinguish between patients with CD and UC, whereas the discriminative power for IBD versus control subjects was substantiated. Conclusion Our results fail to demonstrate that the previously identified seven‐gene classification model is able to discriminate between CD and UC but suggest that the gene panel merely discriminates between inflamed and noninflamed colonic tissue. Thus, a reliable and simple diagnostic tool is still warranted for optimal diagnosis and treatment of patients with IBD, especially the subgroup with unclassified disease.