Chromoendoscopy improves dysplasia detection in UC

The standard of care for patients with long-standing ulcerative colitis (UC) calls for frequent surveillance colonoscopies with multiple random biopsies. But this method, which is time-consuming and expensive, results in low diagnostic yields when performed with standard white light endoscopy (WLE).

Studies have shown that chromoendoscopy (CE) — the spray application of dye to the colon during endoscopy — can improve dysplasia detection rates as much as fourfold. Methylene blue staining, for instance, distinguishes dysplastic and non-dysplastic lesions with 93 percent sensitivity and specificity. Still, despite convincing data and endorsements from the American Gastroenterological Association and the Crohn's and Colitis Foundation of America, acceptance of CE has been slow.

Michael F. Picco, M.D., of Mayo Clinic in Jacksonville, Fla., attributes this in part to the perception that CE is difficult to master. Other stumbling blocks to widespread adoption of the technique include potentially longer procedure times and reliability of image interpretation.

To answer these objections, Dr. Picco led a study to determine whether CE could be learned easily with acceptable accuracy and procedure length. Investigators included six skilled endoscopists, two from each Mayo Clinic site, who were inexperienced with chromoendoscopy in UC.

The endoscopists performed surveillance colonoscopies on 57 patients with long-standing ulcerative colitis, first visualizing each colon segment with WLE, then with CE. The amount of time the colon was examined was recorded for each patient, and photos were taken of polyp abnormalities along with high-resolution images of normal colon tissue.

The study had three stages:

  • The first phase determined that interobserver agreement of images from the first 15 patients was acceptable enough for the study to continue.
  • The second phase, conducted months after the first, was the primary assessment of interobserver agreement. Investigators analyzed nearly 600 white light and indigo carmine images from 57 patients, including images of 160 polyps with 29 dysplastic lesions.
  • In the final phase, which focused on dysplasia detection, enrollment was increased to 75 patients to estimate the incremental increase in dysplasia detection for CE compared with WLE alone.

Study results

Interobserver agreement was excellent, with kappa scores of 0.91 percent and 0.86 percent for WLE and CE, respectively. Dr. Picco notes that these scores are particularly impressive in UC, where lesions may be obscured by background inflammation.

Of 29 dysplastic lesions, 10 were identified by WLE and an additional 12 by CE, leading to a relative incremental yield for CE of 120 percent. The relative incremental yield was highest for flat polyps. WLE identified just one patient with flat dysplasia, whereas CE identified seven. Missed polyps were generally flat lesions less than 0.2 inches (5 millimeters) in size.

Colonoscopy withdrawal time, including WLE visualization, indigo carmine staining and random biopsy, improved from 31 minutes for endoscopists performing fewer than five procedures to 18 minutes for those performing five to 14 procedures. Proficiency was generally attained after 10 exams. Overall, CE added about 10 minutes to standard WLE with random biopsies — similar to times in the published literature.

Dr. Picco says, "Our findings show that CE is easy to learn and verifiable, has good observer agreement, and doesn't add much time to surveillance colonoscopy. One limitation is that although the investigators were unfamiliar with CE in patients with long-standing UC, they were all experienced endoscopists working in referral centers. Still, we think our results are encouraging for the adoption of CE into general endoscopic practice."

Other study investigators include: