Search in the Abstract Database

Abstracts Search 2014

* = Presenting author

P255. Computed virtual chromoendoscopy with FICE in the definition of raised dysplastic lesions and pseudopolyps in long-standing ulcerative colitis: a preliminary, prospective, study towards a new classification

A. Cassinotti1, F. Buffoli2, P. Fociani3, V. Villanacci4, T. Staiano2, M. Fichera1, R. Grassia2, G. Manes1, M. Salemme4, M. Lombardini1, P. Molteni1, G. Sampietro5, D. Foschi5, G. Maconi1, R. de Franchis1, S. Ardizzone1, 1Luigi Sacco University Hospital, Gastroenterology and IBD Unit, Milan, Italy, 2Istituti Ospitalieri Cremona, Endoscopy Unit, Cremona, Italy, 3Luigi Sacco University Hospital, Pathology Unit, Milan, Italy, 4Spedali Civili, Pathology Unit, Brescia, Italy, 5Luigi Sacco University Hospital, General Surgery, Milan, Italy

Background

The surface pit pattern is an aid to the differential diagnosis of colorectal polyps in the general population. We evaluated for the first time the diagnostic accuracy for neoplastic and non-neoplastic colorectal polyp differentiation with detection of surface patterns by Flexible spectral Imaging Color Enhancement (FICE) in patients with long-standing ulcerative colitis (UC).

Methods

A prospective study was performed on consecutive UC patients with raised colorectal lesions found during surveillance colonoscopy for long-standing (>8 years) disease. Targeted biopsies of all raised lesions as well as polypectomy of lesions suspected for neoplasia were performed, as well as additional random biopsies of flat mucosa were taken every 10-cm. Kudo classification of surface pit-pattern (I–II = normal/ovalar, not suspect for neoplasia, III–V = tubular/gyrus-like, suspect for neoplasia; 0 = unclassifiable) was used to predict the histology by FICE without magnification (setting 4), as well as any other endoscopic feature, including disease activity. We determined the accuracy (ACC), sensitivity (SE), specificity (SP), positive (PPV) and negative (NPV) predictive value of FICE in correlating diagnosis with histology.

Results

A total of 122 lesions (mean size 8, range 2–30 mm) was analysed from 30 patients (6 females; mean age 57 years). A library of 538 pictures was recorded. 9 lesions (7%) were found to be neoplastic (low-grade dysplasia). No dysplasia was found in random biopsies from flat mucosa. Table 1 shows the correlation between surface patterns and histology.

The ACC, SE, SP, PPV and NPV of FICE using the Kudo classification alone were 83%, 89%, 83%, 33% and 99%, respectively. NPV was higher while PPV was lower than those reported in the literature about FICE in non-IBD patients [1,2]. The performance of FICE significantly improved after the addition of specific endoscopic features (such as fibrin cap and/or hyperaemic mucosa) as a marker of inflammatory activity in a new, modified III–IV Kudo pit-pattern (94%, 89%, 95%, 67% and 99%, respectively).

Table 1
Surface patternNumber (%) of lesionsDysplasic lesions, n (%)
0 (unclassifiable)23 (19%)0 (0%)
Kudo I alone54 (44%)0 (0%)
Kudo II alone24 (20%)1 (4%)
Kudo III-L alone18 (15%)5 (28%)
Kudo III-S alone0 (0%)0 (0%)
Kudo IV3 (2%)3 (100%)
Kudo V0 (0%)0 (0%)
Mixed, Kudo I and II78 (64%)1 (1%)
Mixed, Kudo III–V21 (17%)8 (38%)

Conclusion

FICE without magnification can help to predict the histology of raised lesions in UC and to guide sampling lesions. Further controlled studies are needed to validate new classifications of mucosal patterns in the setting of IBD.

1. Pohl J, (2008), Computed virtual chromoendoscopy for classification of small colorectal lesions: a prospective comparative study., Am J Gastroenterol.

2. Yoshida N, (2012), The detection of surface patterns by flexible spectral imaging color enhancement without magnification for diagnosis of colorectal polyps, Int J Colorectal Dis.