Magnification imaging in inflammatory bowel disease: a useful aid for the prediction of disease relapse?

Predicting the evolution and outcome of inflammatory bowel disease presents us with a challenging problem, and its solution could potentially have a great impact on our therapeutic strategy in this condition. Disease flares seem to occur randomly. It would be greatly beneficial to patients if physicians could predict a relapse and could treat patients selectively according to their risk of relapse. Unfortunately, no gold standard test exists to predict relapse in ulcerative colitis or in inflammatory bowel disease in general.

Much effort has been put into the identification of biochemical markers that might predict the risk of disease relapse [1]. In patients with Crohn’s disease, an index made up of a set of biochemical markers (erythrocyte sedimentation rate, α₁ glycoprotein, and α₂ globulin) has some usefulness, allowing the calculation of a prognostic index - a threshold of 0.35 in this index was found to be predictive of relapse over a period of 18 months [2]. In Crohn’s disease, patients with increased C-reactive protein were found to have a higher likelihood of relapse in the following 2 years than patients with a normal C-reactive protein [3]. The combination of a normal C-reactive protein and a normal erythrocyte sedimentation rate was found in another study to be highly predictive of the absence of relapse in the following 6 weeks [4].

Biochemical markers are less helpful for the prediction of disease course in ulcerative colitis. Fecal calprotectin may be a promising marker in both Crohn’s disease and ulcerative colitis, but exact cutoff levels for predicting relapse are not clearly defined and the test still lacks specificity [5]. In ulcerative colitis the focus has been on the use of histological characteristics of remission at the mucosal level for the prediction of disease course.

In this issue of “Endoscopy,” Hurlstone et al. [6] report a large prospective study on the value of high-magnification chromoscopic colonoscopy (HMCC) for in vivo prediction of histopathological changes in patients with ulcerative colitis. The authors found a good correlation between the Saitoh criteria for magnification imaging and Matts’ histopathological criteria. They performed a biphasic examination of all colonic segments in 325 consecutive patients with a known history of ulcerative colitis. First, conventional white-light endoscopy was used, followed by HMCC with indigo carmine dye. Magnification endoscopy was significantly better at predicting the extent of disease and histopathological grade than conventional endoscopy and the Baron scores for all parameters.

The question remains, however, of whether endoscopic assessment of histological changes can predict clinical outcome. Several authors have claimed that histological changes in the absence of clinical signs of disease activity can predict a high risk of relapse [7] [8], and Tanaka et al. [9] demonstrated the value of a combination of five histological criteria and endoscopic assessment of the extent of the disease for the prediction of surgical risk in patients with ulcerative colitis. Recently, Nishio et al. [10] prospectively studied the role of HMCC in predicting the clinical course in 113 patients with quiescent ulcerative colitis. Only patients with an inactive mucosal appearance, obscure vascular patterns, granularity of the mucosa, or erythema on routine high-resolution colonoscopy were included; patients with contact bleeding, spontaneous bleeding, erosions, or ulceration were excluded. The authors classified the rectal pit pattern appearance into four groups according to its irregularity: grade 1, small, round, and regularly arranged pits; grade 2, rather large, oval pits, somewhat irregular in arrangement; grade 3, pits of various shapes and sizes, and irregularly arranged; grade 4, dispersed pits varying in morphology, associated with small erosions. The authors found a good correlation between HMCC and both histopathological findings and mucosal interleukin 8 activity. Using this classification, they could predict a relapse risk of 60 % in the following 12 months in patients who had a grade 4 rectal pit pattern, but who had only mild or no endoscopic features of disease activity at the time of endoscopy. Early studies using HMCC had already shown that the presence of normal crypt openings and visible network patterns in the rectum correlated with a lower grade of histological inflammation [11].

Higaki et al. [12] used a high-frequency endosonographic catheter to assess the thickness of the mucosal and submucosal layer of the colonic wall in 23 patients with ulcerative colitis, who were followed prospectively for 1 year. Patients who had a disease relapse were found to have a thicker colonic wall (mucosa + submucosa) in comparison with patients who remained in remission. It is not known how these findings relate to data derived from HMCC, however.

The advantage of a method of endoscopic estimation of histological inflammation and risk of relapse lies in the fact that physicians can make an instant assessment of the histological disease activity during an examination which is widely performed in patients with ulcerative colitis. This would allow better assessment of the efficacy of different treatment strategies for the induction of lasting remission. Further therapy can then be guided by the endoscopic findings. Apparently, this goal can be achieved more accurately using the six HMCC “optical biopsies” described in the study by Hurlstone et al. than with conventional white-light endoscopy [6]. Fujiya et al. [13] demonstrated that in patients with ulcerative colitis the Saitoh criteria for HMCC correlated more accurately with the histopathological findings than did the white-light endoscopic findings. A small proportion of the patients in their study (n = 18) was studied prospectively and it was shown that the presence of minute defects on HMCC, such as small depressions surrounded by edema, was associated with a high risk of disease relapse (7/9 patients). Although the data we already have are very interesting, more studies are necessary to prospectively assess the specificity and the sensitivity of this risk assessment tool, so the prospective 2-year follow-up study of the 325 patients by the Hurlstone group is essential to address these issues.

The application of HMCC in ulcerative colitis is impeded by the use of different classification schemes and the focus on different parts of the colon in the few studies that are available. Matsumoto et al. [11] and Nishio et al. [10] confined themselves solely to the rectal mucosa, whereas, according to the studies by Fujiya et al. [13], Kunihiro et al. [14], and Hurlstone et al. [6], any part of the colon can be used for predicting the histopathological findings and disease course. It should be emphasized, however, that although ulcerative colitis involves the mucosa in a diffuse manner, healing can be a more segmental process in the early stages, and also that the HMCC image can probably vary between segments. All the authors used different classification methods. Whereas Matsumoto et al. [11] mainly focused on the presence of normal rectal crypt openings and visible network patterns, Nishio et al. [10] developed a four-grade classification of the rectal pit patterns. Hurlstone et al. simplified the modified Saitoh criteria, previously also slightly adapted by Kunihiro et al. [6] [13] [14]. This lack in uniformity merely reflects the difficulties inherent in the quest for an ideal endoscopic marker for the assessment of disease activity and prediction of disease outcome in patients with ulcerative colitis.

Another problem with the implementation of these new techniques in the day-to-day practice of gastroenterologists is inter- and intraobserver agreement. Although Hurlstone et al. [6] report good inter- and intraobserver kappa coefficients for the HMCC Saitoh grades, the expertise and experience of this tertiary referral center is not representative of endoscopists not trained at interpreting HMCC images. Apart from the need to purchase additional endoscopic equipment (although this would have other uses, including screening for dysplasia), a learning curve for the assessment and interpretation of these images is inevitable. It is therefore critical that these results are confirmed in larger, multicenter trials that include secondary referral centers staffed by gastroenterologists with less experience in HMCC.

The ideal marker for the prediction of disease course in inflammatory bowel disease needs to be simple, quick and easy to perform, minimally invasive, cheap, and highly reproducible between different individuals and centers [1]. HMCC clearly does not meet these criteria; and neither do any of the currently available biochemical markers. Nevertheless, HMCC findings correlate well with histopathological findings and, in experienced hands, the technique represents an additional aid for predicting the course of ulcerative colitis. Studies of the technique are also very useful in promoting a better understanding of the disease and its evolution. However, it remains to be seen if HMCC can be usefully implemented in general gastroenterological practice.

Competing interests: None

References

• 1 Vermeire S, Van Assche G, Rutgeerts P. Laboratory markers in IBD: useful magic or unnecessary toys?.  Gut. 2006;  55 426-431
• 2 Brignola C, Campieri M, Bazzocchi G. et al . A laboratory index for predicting relapse in asymptomatic patients with Crohn’s disease.  Gastroenterology. 1986;  91 1490-1494
• 3 Boirivant M, Leoni M, Tariciotti D. et al . The clinical significance of serum C reactive protein levels in Crohn’s disease: results of a prospective longitudinal study.  J Clin Gastroenterol. 1988;  10 401-405
• 4 Consigny Y, Modigliani R, Colombel J F. et al . Biological markers of short-term relapse in Crohn’s disease (CD).  Gastroenterology. 2001;  20 (Suppl) A53
• 5 Costa F, Mumolo M G, Ceccarelli L. et al . Calprotectin is a stronger predictive marker of relapse in ulcerative colitis than in Crohn’s disease.  Gut. 2005;  54 364-368
• 6 Hurlstone D, Sanders D, McAlindon M. et al . High-magnification chromoscopic colonoscopy in ulcerative colitis: a valid tool for in vivo optical biopsy and assessment of disease extent.  Endoscopy. 2006;  38 1213-1217
• 7 Jewell D P. Ulcerative Colitis. In: Feldman M, Friedman LS, Sleisenger MH, editors Sleisinger and Fordtran’s gastrointestinal and liver disease: pathophysiology, diagnosis, management. 7th edn. Philadelphia; WB Saunders 2002
• 8 Riley S, Mani V, Goodman M. et al . Why do patients with ulcerative colitis relapse?.  Gut. 1990;  31 179-183
• 9 Tanaka M, Kusumi T, Oshitani N. et al . Validity of simple mucosal biopsy criteria combined with endoscopy predicting patients with ulcerative colitis ultimately requiring surgery: a multicenter study.  Scand J Gastroenterol. 2003;  38 594-598
• 10 Nishio Y, Ando T, Maeda O. et al . Pit patterns in rectal mucosa assessed by magnifying colonoscopy are predictive of relapse in patients with quiescent ulcerative colitis.  Gut. 2006;  May 8 [online]
• 11 Matsumoto T, Kuroki F, Mizuno M. et al . Application of magnifying chromoscopy for the assessment of severity in patients with mild to moderate ulcerative colitis.  Gastrointest Endosc. 1997;  46 400-405
• 12 Higaki S, Nohara H, Saitoh Y. et al . Increased rectal wall thickness may predict relapse in ulcerative colitis: a pilot follow-up study by ultrasonographic colonoscopy.  Endoscopy. 2002;  34 212-219
• 13 Fujiya M, Saitoh Y, Nomura M. et al . Minute findings by magnifying colonoscopy are useful for the evaluation of ulcerative colitis.  Gastrointest Endosc. 2002;  56 535-542
• 14 Kunihiro M, Tanaka S, Sumii M. et al . Magnifying colonoscopic features of ulcerative colitis reflect histologic inflammation.  Inflamm Bowel Dis. 2004;  10 737-744