Loading...
For patients in the U.S. with ulcerative colitis (UC), standard endoscopic practice during surveillance colonoscopy is to take four quadrant biopsies every 10 cm in addition to biopsies of visibly abnormal areas that could be neoplastic. Patients typically undergo colonoscopy at 2- to 3-year intervals, starting at 7 to 8 years after the onset of symptoms and continuing for 20 years, and then annually thereafter. Adjustments in surveillance are made for primary sclerosing cholangitis and family history of colorectal cancer. Recent results have challenged this approach in at least two areas. First, patients with long-standing UC who have endoscopically normal colons and negative surveillance exams are at considerably lower risk for neoplasia than are other UC patients and likely could undergo surveillance at longer intervals. Second, several studies of chromoendoscopy have found that targeting flat lesions that are endoscopically visible after dye spraying leads to detection of more neoplastic lesions than do random biopsies taken during white-light conventional colonoscopy.
In the present study from Germany, conventional colonoscopy was compared with chromoendoscopy plus endomicroscopy (performed with confocal laser microscopy colonoscopes) in 161 patients with long-standing UC (≥8 years). Confocal laser microscopy colonoscopes provide targeted images of the mucosa that correspond to histologic sections; they can differentiate normal mucosa from inflamed mucosa and neoplastic mucosa from non-neoplastic mucosa. Although they are not effective at differentiating degrees of dysplasia, the importance of this is minimal with regard to targeting biopsies.
All of the patients had endoscopically inactive colitis, but endomicroscopy was more accurate than conventional colonoscopy at identifying patients with histologic activity. Of more interest was the detection of intraepithelial neoplasia, defined using the Vienna classification and without regard to whether lesions were sporadic adenomas or dysplasia-associated lesions or masses. Significantly more intraepithelial neoplasias were diagnosed in the chromoendoscopy group (12 low-grade and 7 high-grade; 14 not seen with conventional viewing) than in the conventional group (3 low-grade and 1 high-grade; P=0.005). Eleven patients in the chromoendoscopy group had intraepithelial neoplasias, compared with 4 in the conventional group (a nonsignificant difference).
The average number of biopsy specimens collected per patient in the conventional group was 42.2, whereas, if only circumscribed lesions had been biopsied in the chromoendoscopy group, the total number of biopsy specimens would have been 3.9 per patient, with no reduction in the number of intraepithelial neoplasias that would have been identified. Confocal laser microscopy was highly accurate in predicting which lesions would be intraepithelial neoplasias, with a sensitivity of 94.7% and a specificity of 98.3% (accuracy, 97.8%). If biopsy had been restricted to circumscribed lesions identified by chromoendoscopy and also identified as intraepithelial neoplasias by confocal laser microscopy, the number of biopsies performed per patient would have been 0.78. Mean procedure times were 42 minutes in the chromoendoscopy group and 31 minutes in the conventional group (P=0.28).
Kiesslich R et al. Chromoscopy-guided endomicroscopy increases the diagnostic yield of intraepithelial neoplasia in ulcerative colitis. Gastroenterology 2007 Mar; 132:874-82.
Comment
These results are impressive, but deciding whether to change a widespread practice is a complex issue. In randomized trials, conventional colonoscopy is not as effective as chromoendoscopy for collecting biopsies to detect intraepithelial neoplasias, but conventional methods work reasonably well in clinical practice. Because repeated use of conventional colonoscopy is very effective on a per-patient basis for identifying individuals with high-grade dysplasia or early cancer, patients who undergo regular surveillance usually are referred for surgery before fatal cancer develops. How much training would be required to convert endoscopists who are skilled in conventional colonoscopy to a program of chromoendoscopy with targeted biopsies? Chromoendoscopy is a relatively simple technique, and experienced endoscopists likely could quickly use it well, but we do not know that with certainty. In addition, quibbling about the safest and most effective dye will come into play.
Whether to convert from conventional colonoscopy with random biopsies to chromoendoscopy-targeted biopsies is a much more practical and important issue than whether to also add confocal laser microscopy. The biggest hindrance to confocal laser microscopy is that it is available only in Pentax colonoscopes, which are used by only a minority of endoscopists. Also, we don’t know how much training is needed to read the confocal laser images accurately. Also at issue is the lack of reimbursement for performance of this procedure; this problem plagues many new endoscopic procedures, particularly those that are ancillary techniques.
Finally, important issues emerge about integrating the overall endoscopic picture (with or without chromoendoscopic findings) and surveillance intervals. For example, can patients with negative chromoendoscopy skip surveillance for 5 or even more years? In my opinion, endoscopists who are trained in chromoendoscopy clearly can substitute targeted biopsies for random biopsies. This approach should detect more intraepithelial neoplasias and is defensible from a medicolegal standpoint in the U.S. Further, such endoscopists could reasonably lengthen surveillance intervals beyond those usually recommended if patients have normal endoscopic findings using these techniques.