Search in the Abstract Database

Abstracts Search 2019

P281 Endoscopic healing assessed by advanced optical enhancement techniques combined with faecal calprotectin (FCP) can accurately assess histological healing in ulcerative colitis patients

R. Cannatelli*1, U. N. Shivaji2,3, S. C. Smith1, D. Zardo4, A. Bazarova1, G. Gkoutos1, S. Ghosh1,2,3, M. Iacucci1,2,3,5

1University of Birmingham, Institute of Translational Medicine, Birmingham, UK, 2National Institute for Health Research (NIHR) Birmingham Biomedical Research Centre, Birmingham, UK, 3University of Birmingham, Institute of Immunology and Immunotherapy, Birmingham, UK, 4University Hospitals Birmingham NHS Foundation Trust, Department of Histopathology, Birmingham, UK, 5University of Calgary, IBD Unit, Calgary, Canada

Background

Mucosal healing (MH) is considered a key target of therapy in ulcerative colitis (UC) but there is debate about endoscopic healing, histological healing, and surrogate marker of MH using faecal calprotectin (FC). We have recently described and validated endoscopic MH using high-definition electronic chromoendoscopy. In this study, we aimed to investigate MH using multiple endoscopic scorings, FC, and validated histological scores.

Methods

We prospectively obtained clinical data, endoscopic scores [Mayo Endoscopic Score (MES), Ulcerative Colitis Endoscopic Index of Severity (UCEIS) PICaSSO score (Paddington International virtual ChromoendoScopy ScOre)] and FC for UC patients undergoing colonoscopy using high-definition (Pentax) iScan optical enhancement (OE) or NBI near focus (Olympus). Histological scorings were assessed using Robarts Histological Index (RHI) and Nancy index (NI). Receiver-operating characteristics (ROC) curves were plotted to determine operating characteristics of FC alone or in combination with endoscopic scores to predict histological healing.

Results

In total, 44 patients (mean age 45 years, 52% men) were included. By partial Mayo score <2, 30 patients (68.2%) were in remission; however, endoscopic remission was seen in only 21(47.7%) with MES = 0 and UCEIS ≤ 1 and 18 (40.9%) with PICaSSO ≤ 2. The mean ± sd of FC was 465.5 ± 703.3 μg/g and 20 (45.5%) patients had FC ≤ 100 μg/g. The histological healing, defined as RHI ≤ 6 was seen in 21 (47.7%) patients and NI ≤ 1 was seen in 19 (43.2%). The threshold for FC alone as a predictor of histological healing using RHI was 313 μg/g with an accuracy of 84.1% (95% CI 69.9–93.4%) and AUROC of 87% (95% CI 75–98%), whilst for NI it was 112 μg/g, with accuracy of 81.8% (95% CI 67.3–91.8%) and AUROC 85% (95% CI 73–96%). The accuracy of predicting histological healing using a combination of PICaSSO and FC(≤100 μg/g) is 93.2% (95% CI 81.3–98.6%) with AUROC 96% (95% CI 91–100%) for both RHI and NI (formula used for NI=FC+1.5*Picasso). The combination of UCEIS and FC (≤100 μg/g) had an accuracy of 90.9% (95% CI 78.3–97.5%) in predicting histological healing for both RHI and NI, with an AUROC of 95% (95% CI 89–100%) and 94% (95% CI 87–100%), respectively. MES was not modelled in combination with FC as the best 2 endoscopy scores were modelled further.

Abstract P281 – ROC curves predicting histological healing.

Conclusion

The combination of PICaSSO and FC could help to identify UC patients with histological healing accurately than FC alone. PICaSSO with FC had better operating characteristics for prediction of histological healing than UCEIS and FC when using advanced endoscopic imaging with either iscan OE or NBI near Focus.