The Paddington International Virtual Chromoendoscopy Score in ulcerative colitis exhibits very good inter-rater agreement after computerized module training: a multicenter study across academic and community practice (with video)

Abstract

Background and aims: Electronic virtual chromoendoscopy (EVC) can demonstrate ongoing disease activity in ulcerative colitis (UC), even when Mayo subscores suggest healing. However, applicability of EVC technology outside the expert setting has yet to be determined.

Methods: Fifteen participants across 5 centers reviewed a computerized training module outlining high-definition and EVC (iScan) colonoscopy modes. Interobserver agreement was then tested (Mayo score, Ulcerative Colitis Endoscopic Index of Severity [UCEIS], and the Paddington International Virtual Chromoendoscopy Score [PICaSSO] for UC), using a colonoscopy video library (30 cases reviewed pretraining and 30 post-training). Knowledge sustainability was retested in a second round (42 cases; 9/15 participants), 6 months after training provision.

Results: Pretraining intraclass correlation coefficients (ICC) were good for the Mayo endoscopic subscore (ICC, .775), UCEIS scoring erosions/ulcers (ICC, .770), and UCEIS overall (ICC, .786) and for mucosal (ICC, .754) and vascular components of PICaSSO (ICC, .622). For the vascular components of UCEIS, agreement was only moderate (ICC, .429) and did not enhance post-training (ICC, .417); conversely, use of PICaSSO improved post-training (mucosal ICC, .848; vascular, .746). Histologic correlation using the New York Mt. Sinai System was strong for both PICaSSO components (Spearman's ρ for mucosal: .925; vascular, .873; P < .001 for both). Moreover, accuracy in specifically discriminating quiescent from mild histologic strata was strongest for PICaSSO (area under the receiver operating characteristic curve [AUROC] for mucosal, .781; vascular, .715) compared with Mayo (AUROC, .708) and UCEIS (AUROC for UCEIS overall, .705; vascular, .562; bleeding, .645; erosions/ulcers, .696). Inter-rater reliability for PICaSSO was sustained by round 2 participants (round 1 and 2 ICC for mucosal, .873 and .869, respectively; vascular, .715 and .783, respectively), together with histologic correlation (ρ mucosal, .934; vascular, .938; P < .001 for both).

Conclusions: PICaSSO demonstrates good interobserver agreement across all levels of experience, providing excellent correlation with histology. Given the ability to discriminate subtle endoscopic features, PICaSSO may be applied to refine stratified treatment paradigms for UC patients.