P254 Re-defining the concept of endoscopic and histological healing by using electronic virtual chromoendoscopy and probe confocal endomicroscopy in ulcerative colitis

Background
The treatment goal of UC has shifted from symptomatic remission alone to endoscopic and recently histological healing.1 The new validated Virtual Chromoendoscopy (VCE) score, PICaSSO (Paddington International virtual ChromoendoScopy ScOre)1 offering detailed mucosal and vascular assessment, and probe confocal laser endomicroscopy (pCLE) as real time in vivo histology, aimed to re-define the concept of mucosal healing (MH). We specifically explored the magnitude of difference between endoscopy and histology defined MH using refined endoscopic assessments.

Methods
In total, 82 UC, 8 controls, male 65.6%; mean age 49.9, SD 14.8 were prospectively enrolled at endoscopy unit, University of Calgary. The endoscopic activity was evaluated by Mayo Endoscopic Score (MES) and PICaSSO mucosal and vascular pattern 1 and thereafter with pCLE (Cellvizio, Paris) after IV fluorescein. The pCLE findings were graded as (A) crypt architecture (Grades 1–4); (B) leakage of fluorescein (Grades 1–4); (C) vessel architecture (Grades 1–4); (D) blood flow (Grades 1–4). Histological score (Robarts histological index, RHI) was used to score histological inflammation. Receiver-operating Characteristic (ROC) curves were plotted to calculate the best cut-off threshold of PICaSSO and pCLE scores to predict histological healing.

Results
For overall PICaSSO score, the optimum cut-off threshold for predicting histological healing defined as RHI ≤ 6 was 4, with sensitivity of 90.0% (95% CI 75.6–96.2) specificity 100% (95% CI 84.6–100), and accuracy of 92.7% (95% CI 84.8–97.3). The overall PICaSSO score of 4 or less was associated with all patients having an RHI ≤ 6. The best cut-off threshold for pCLE score was 10, with sensitivity 95.0% (95% CI 86.0%–99.0%), specificity 95.5% (95% CI 77.2–99.8) and accuracy of 95.1% (95% CI 88.0–98.7). The accuracy of predicting histological healing using PICaSSO or pCLE were superior to MES 0, which had sensitivity of 80% (95% CI 67.6–89.2), specificity 95.5% (95% CI 77.2–99.9), and accuracy of 84.2% (95% CI 74.4–91.3).

graphic
ROC curve of PICaSSO and pCLE for predicting histological healing.

Conclusions
The new VCE PICaSSO score and pCLE score can predict histological healing defined by RHI accurately. Advances in endoscopy enable close approximation to histology and can accurately re-define in real-time MH. Overall PICaSSO score of 4 or less was associated with RHI ≤ 6 in all patients. Large prospective studies are necessary to ascertain whether, with new endoscopic technologies such as readily available VCE, histology can still provide additional information about course of UC.