Efficacy of probe-based confocal laser endomicroscopy for surveillance in ulcerative colitis endomicroscopy for ulcerative colitis surveillance

We read with interest the study by van den Broek and colleagues published in the recent issue of Endoscopy [1]. The authors examined whether the use of probe-based confocal laser endomicroscopy (pCLE) may increase the efficiency of surveillance of patients with ulcerative colitis. In the study, 48 visible lesions and 87 random areas in 22 patients with ulcerative colitis were evaluated by real-time narrow-band imaging (NBI) and high-definition endoscopy (HDE). Fluorescein-enhanced pCLE was performed prior to biopsy. The outcome measures were the feasibility and the diagnostic accuracy of pCLE. They showed that the sensitivity, specificity, and accuracy of pCLE were 65 %, 82 %, and 81 %, whereas these figures were 100 %, 89%, and 92 % for real-time endoscopic diagnosis with NBI and HDE. From these results, the authors concluded that pCLE for ulcerative colitis surveillance is feasible and has reasonable diagnostic accuracy.

We agree with the authors that taking random biopsies is time-consuming and of doubtful efficiency. We have previously shown that pit pattern diagnosis might be useful in improving the efficiency of surveillance for ulcerative colitis [2]. Considering that pCLE is, in fact, in vivo histology and has shown high agreement with true histopathology, it may be possible that pCLE substitutes random biopsies and obviates targeted biopsies. However, in the van den Broek study, there are major issues that need to be discussed before their conclusion can be drawn.

The first point concerns the management of lesions that are indefinite for dysplasia. Van den Broek used the Vienna criteria for histopathological classification [3]. The Vienna criteria were established to develop common worldwide terminology for gastrointestinal epithelial neoplasia, including esophageal, gastric, and colorectal lesions. Neoplastic lesions are categorized as Category 3, Category 4 or Category 5. The van den Broek study used these categories to classify biopsied lesions. This classification is used for neoplastic lesions in general; however, it is well known that histopathological classification of specimens in ulcerative colitis needs special attention because histopathological diagnosis is sometimes very difficult due to the inflammatory changes. This makes it difficult to make a distinction between low-grade dysplasia and non-neoplastic lesions with inflammation. This is why Riddell et al. divided the classification of “indefinite for dysplasia” into three different categories – probably negative, unknown, and probably positive, each of which needs a different follow-up strategy [4]. Riddell reported that patients with “unknown” or “probably positive for dysplasia” need short-interval follow-up [4]. Therefore, it is important not to miss lesions with “indefinite for dysplasia (unknown or probably positive)” during surveillance. As the histological distinction between low-grade dysplasia and indefinite for dysplasia is sometimes difficult, it seems that pCLE can hardly distinguish between these two lesion types. However, in the van den Broek study, lesions that were “indefinite for dysplasia” were excluded because the study focused only on neoplastic lesions. It would be of interest to determine whether or not pCLE can identify lesions classified as “indefinite for dysplasia.”

The next point is the outcome measure of feasibility. Van den Broek et al. expressed the feasibility of pCLE as pCLE imaging time required, percentage of imaging time with clear pCLE histology, and pCLE video quality as rated by two endoscopists. However, the pCLE imaging time required may not reflect the total time of examination directly and we believe that the most important issue is the total time needed to complete the surveillance colonoscopy in each individual. To compare the efficacy of surveillance between target biopsy and random biopsy, a nationwide randomized controlled study is ongoing in Japan [5]. In this trial, the total time of the examination is included as the outcome measure. We believe this prospective trial will clarify the difference between target biopsy and random biopsy in terms of total time needed for the examination.

We believe that the van den Broek study, which was undertaken by a leading group in the world in this field, has shown very important findings. However, to evaluate the efficacy of pCLE, we believe that the above issues need to be discussed further.

References

• 1 van den Broek F J, van Es J A, van Eeden S et al. Pilot study of probe-based confocal laser endomicroscopy during colonoscopic surveillance of patients with longstanding ulcerative colitis.  Endoscopy. 2011;  43 116-122
• 2 Hata K, Watanabe T, Motoi T, Nagawa H. Pitfalls of pit pattern diagnosis in ulcerative colitis-associated dysplasia.  Gastroenterology. 2004;  126 374-376
• 3 Schlemper R J, Riddell R H, Kato Y et al. The Vienna classification of gastrointestinal epithelial neoplasia.  Gut. 2000;  47 251-255
• 4 Riddell R H, Goldman H, Ransohoff D F et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications.  Hum Pathol. 1983;  14 931-968
• 5 Watanabe T, Ajioka Y, Matsumoto T et al. Target biopsy or step biopsy? Optimal surveillance for ulcerative colitis: a Japanese nationwide randomized controlled trial.  J Gastroenterol. 2011;  46 (Suppl 1) 11-16