You are here

Forward-Viewing Colonoscopy Versus Full-Spectrum Endoscopy (Fuse) for Dysplasia Detection With and Without Chromoendoscopy In Inflammatory Bowel Diseases: A Prospective, Randomized-order, Crossover Tandem Surveillance Colonoscopy Study

Clinical Gastroenterology and Hepatology, Volume 13, Issue 3, March 2015, Pages 548–551

Background

Traditional forward-viewing colonoscopy (FVC) is impeded by a narrow field of view of <170 degrees with adenoma “miss-rates” of 24-42%. This limitation impairs dysplasia surveillance in chronic colitis. Chromoendoscopy improves visualisation of dysplasia but has never been used with FUSE, which provides 330 degree visualisation. This study compares FVC and FUSE with and without chromoendoscopy for the identification of dysplasia in IBD.

Methods

This was a prospective, randomized-order, crossover tandem colonoscopy trial at an academic IBD centre. Inclusion criteria were left sided or extensive colitis for >10 years disease duration, primary sclerosing cholangitis (PSC) or previous dysplasia. Patients underwent same-day, back-to-back tandem colonoscopy with FVC (PCF/CF 180/190, Olympus, Japan) and the FUSE colonoscope (EndoChoice, USA) under propofol sedation. Order randomization was computer-generated. Patients and endoscopist were masked to group allocation until immediately prior to colonoscopy. All lesions detected during initial colonoscopy were removed and any additional lesions detected on subsequent colonoscopy were then removed. Chromoendoscopy (CE) was with methylene blue dye spray, additional random biopsies were performed during the second colonoscopy. The primary endpoint was dysplasia-yield. The secondary endpoint was detection of dysplasia on random biopsy versus CE-targeted biopsies. Logistic regression and Chi square statistics were performed.

Results

Twenty-four participants were randomized (16 males, mean age 37 years, Crohn’s colitis n=16, ulcerative colitis n=8, mean duration of colitis: 13.5 years). Two subjects had prior low-grade dysplasia and two had primary sclerosing cholangitis. Ileal-intubation rate was 100%. 46% were randomized to FVC first and 54% to FUSE first. All patients had cross-over CE as the second procedure. On a per-lesion analysis for lesion detection, FUSE odds ratio (OR) was 4.86 (95% CI: 1.43-16.49) vs FVC OR: 2.33 (95% CI: 0.74-7.13; P<0.01). Dysplasia detection with FUSE had an OR: 7.67 (95% CI: 9.85-69.6) vs FVC OR: 2.90 (95% CI: 0.50-16.67). Combined hyperplastic/ dysplasia detection with FUSE had an OR: 3.80 (95% CI: 1.07-13.52) vs FVC OR: 1.74 (95% CI 0.52-5.74). The mean lesion detection with FUSE vs FVC was 1.62 vs 0.45 (P<0.042), with mean dysplasia detection of 0.30 vs 0.09 respectively (P=0.21). FUSE +/- CE versus FVC +/- CE had a mean dysplasia detection of 0.25 vs 0.04 (P=0.04) and mean lesion detection of 2.25 vs 0.67 respectively (P=0.003). Lesional and dysplasia miss rates are shown in Table 1. Mean caecal intubation times for FUSE and FVC were 4.7 and 4.6 minutes respectively (ns). Dysplasia yield on targeted biopsies with CE yield was 10.8% vs 0% on random biopsies (P<0.0001).

Conclusions

FUSE significantly increased dysplasia identification in IBD surveillance. We confirmed the dysplasia yield of random biopsies in the setting of IBD is extremely low.

Table 1 Miss rates for FUSE and forward-viewing colonoscope (FVC) for lesions and dysplasia

  Lesion miss rate Dysplasia miss rate
FUSE 31% 0%
FVC 78% 66.7%