P289. Concordance in IBD endoscopic scoring requires expertise and training: Preliminary results of an ongoing IG‑IBD study [on behalf of Italian Group for Inflammatory Bowel Disease (IG‑IBD)]
M. Daperno1, M. Comberlato2, F. Bossa3, L. Biancone4, A. Bonanomi5, A. Cassinotti6, R. Cosintino7, G. Lombardi8, R. Mangiarotti7, A. Orlando9, A. Papa10, R. Pica11, F. Rizzello12, R. D'Incà13
1A.O. Ordine Mauriziano, Gastroenterology Unit, Turin, Italy; 2Ospedale di Bolzano, Gastroenterology Unit, Bolzano, Italy; 3IRCCS Ospedale Casa Sollievo della Sofferenza, Gastroenterology Unit, San Giovanni Rotondo, Italy; 4A.O. Universitaria Policlinico Tor Vergata Roma, Gastroenterology Unit, Rome, Italy; 5A.O.U. Careggi, Gastroenterology Unit, Florence, Italy; 6L. Sacco University Hospital, Department of Gastroenterology, Milan, Italy; 7San Camillo Forlanini Hospitals, Rome, Italy; 8A.O. Cardarelli, Gastroenterology Unit, Naples, Italy; 9V Cervello Hospital, Istituto di Medicina Generale e Pneumologia Reparto di Medicina Interna, Palermo, Italy; 10Catholic University of Rome, Internal Medicine and Gastroenterology Complesso Integrato Columbus, Rome, Italy; 11Università La Sapienza, Gastroenterology Unit, Rome, Italy; 12Policlinico S. Orsola Malpigli-Università Di Bologna, Di Medicina Interna e Gastroenterologia, Bologna, Italy; 13University of Padua, Padua, Italy
Background: Endoscopic scoring is an essential tool required for clinical trials and probably for routine practice. As multicenter studies increasingly include endoscopic evaluation as relevant end-point, reliability of endoscopic scoring may become an issue.
Aim of this multicenter IG-IBD study was to explore reproducibility of endoscopic scoring in the setting of dedicated IBD endoscopist, and to evaluate the effectiveness of a dedicated training program in amelioration of basal agreement.
Methods: 13 expert endoscopists reviewed endoscopic videoclips (6 ulcerative colitis clips, 5 post-surgical Crohn clips and 5 luminal Crohn clips), and blindly assigned endoscopic scores:
- Mayo endoscopic subscore for ulcerative colitis
- Rutgeerts' score for post-surgical Crohn's disease
- CDEIS and SES-CD for luminal Crohn's disease.
At the end of every score assessment, discussion was allowed and reference score was voted unanimously for every clip. The study is still ongoing with a validation phase and an educational program will start early in 2012.
Results: 78 combinations were available for every score (based on 13 observers' scores). Median kappa values were normally distributed (p = 0.1962 and p = 0.0672 for Mayo and Rutgeerts'score, respectively). Median Kappa values for Mayo scores and for Rutgeerts' score were rather unsatisfactory: 0.4405 (95%CI 0.37500.5026) and 0.3750 (95%CI 0.28600.4440) respectively. Intraclass correlation coefficients for total CDEIS and SES-CD attributed and computed by the same 13 observers were on the opposite highly significant and reached excellence: 0.8349 (95%CI 0.54140.9951) and 0.9287 (95%CI 0.75720.9981), respectively.
A second meeting is planned and scores will be re-attributed in a blinded fashion, educational material for increasing endoscopic scoring concordance will be produced at the end of the standardization process.
Conclusions: Endoscopic scoring is relevant for clinical trials, and the use of scores in clinical practice was advocated in order to better identify relevant changes in endoscopic disease activity. In this preliminary experience, simpler endoscopic scores (Mayo and Rutgeerts' score) seem to require a larger amount of education in order to reach adequate agreement, while for more detailed and complex scores (CDEIS and SES-CD) agreement inbetween observers is very high. Reproducibility of endoscopic scores cannot be assumed to be very high, and educational programs aimed to maximize agreement in scoring are mandatory.