Endoscopic Ultrasound in Crohn's Disease, Ulcerative Colitis
Endoscopic Ultrasound in Crohn's Disease, Ulcerative Colitis
Apart from transabdominal high frequency ultrasound, there is no method available to date to reliably assess the thickness of the colonic wall and its different layers to allow differentiation of CD and UC. In this prospective, blinded and comparative EUS study using a forward-viewing radial echoendoscope in the sigmoid colon, we were able to clearly differentiate between inflamed and non-inflamed colonic wall, irrespective of the underlying aetiology. Furthermore, the amount of thickening of the single wall layers in combination with absence/presence of paracolonic lymphnodes allowed differentiation between active UC and active CD. It was, however, not possible to differentiate IBD in remission from healthy controls.
The ability to image inflamed and non-inflamed colonic wall has already been shown in transabdominal US. In a meta-analysis, in which the diagnostic accuracy of different cut-off values in CD were compared, sensitivity and specificity of 88% and 93% were achieved when a TWT threshold >3 mm was applied.
In contrast to CD, TWT in UC could not be correlated with clinical disease activity in some previous studies. None of the currently published studies using transabdominal US were able to differentiate the different wall layers from each other and make a diagnosis between CD and UC.
In our study using the following criteria: (i) increased TWT, increased mucosal (ii) and submucosal (iii) thickness and (iv) presence of paracolonic lymphnodes, we were able to differentiate active UC from active CD with a sensitivity of 92.3%.
Even at the early stages of EUS, its ability to visualise the depth of the large bowel wall from the endoluminal side was proven. But, since the oblique-viewing echoendoscopes did not allow to be passed further than rectum or at best the lower sigmoid, EUS studies focussed on rectal cancer staging.
If the more proximal parts of the colon were to be evaluated, mechanical miniprobes were used through the accessory channel of a colonoscope, which allowed examination at the same time as colonoscopy with the advantage of 'one stop shopping', making further bowel preparation, sedation and another examination unnecessary. However, miniprobes were constructed in the 80s and early 90s, are mechanical and their crystals and imaging resolution have never updated. Since they did not allow EUS-FNA, they were not used much any more. The current electronic scanners are highly superior in resolution, ability to allow colour Doppler etc., so that miniprobes are largely mothballed. However, if isolated right sided colitis or ileitis should be investigated, it represents the only viable alternative.
If only the rectum needs to be evaluated, the commercially available stiff rectal probe will offer very good views for the same evaluation.
For this first feasibility study, we included only patients with active disease in the left colon as well as those in remission with history of left colon disease to ensure homogenous inclusion criteria in both groups. The echoendoscope used is by nature a 'gastroscope' used in the colon and a skilled endoscopist may be able to perform a full colonoscopy but this is certainly not true for an average examiner. But, the left colon is affected in many patients with CD and nearly all of the UC patients, which makes this method easily accessible for many of the IBD patients. No complications were observed in this study regarding the use of EUS in patients with active IBD. We therefore assume that complication rates would not differ from those doing colonoscopy in the same patients. This prototype front-view echoendoscope used in our study has an angle of view of 140 ° and similar forward propelling characteristics comparable to standard endoscopes ensuring a comparable view inside the colon throughout the examination. This is also true in regards of the stiffness of the tip of the scope and tip flexion. Consequently, safety and potential complications of EUS in patients with active IBD are unlikely to be different from standard colonoscopy. A study conducted by Navaneethan et al. showed that the prevalence of perforation in IBD patients was 0.3% (19/5295 patients).
Shimizu et al. related the increased depth of the different wall layers determined by EUS to active inflammation of IBD. Using miniprobe EUS, Higaki et al. explored the correlation between inflammatory changes of mucosa and submucosa in the rectum and relapse in 23 patients with UC in remission, compared to 14 healthy controls. The thickness of the mucosal/submucosal layers was significantly increased in patients with later relapse when compared to those remaining in remission.
However, initial efforts to demonstrate the ability of EUS to differentiate CD from UC were largely disappointing. As a consequence, EUS was not developed further for the evaluation of IBD. Meanwhile, forward-viewing instruments have become available, which provide high resolution imaging and might allow negotiation of colon loops. In our study using such an echoendoscope, we were able to precisely differentiate and measure all the wall layers individually. Patients with active UC had a prominent thickening of the mucosal layer, whereas patients with active CD had a significant thickening of the submucosal layer.
In our study, histology was used as reference for proof of disease and spread of inflammation. This showed a strong correlation between TWT, and mucosal and submucosal thickness in patients with IBD and HC and proved that our EUS diagnoses were viable using the thickness of the different wall layers as a measure. Other authors proved that the vertical spread of inflammation of UC was consistent with histopathological findings in 90% of the investigated cases. Hurlstone et al. evaluated the ability to predict the clinical severity of disease by the wall thickness in rectal disease and provided a strong correlation between inflammatory scores and wall thickness. The authors concluded that EUS of the colon could become a useful adjunctive examination for the transmural assessment of the colonic wall in UC. This might justify expensive and time consuming EUS developments since 'mucosal healing' has become a major target of interest especially for the evaluation of failure/success rates of the expensive TNF alpha therapy. Several transabdominal studies have already shown a reduction in TWT and mural vascularity after TNF alpha therapy in CD patients. Maintenance of wall thickness and increased mural vascularity were seen in all nonresponders.
Discussion
Apart from transabdominal high frequency ultrasound, there is no method available to date to reliably assess the thickness of the colonic wall and its different layers to allow differentiation of CD and UC. In this prospective, blinded and comparative EUS study using a forward-viewing radial echoendoscope in the sigmoid colon, we were able to clearly differentiate between inflamed and non-inflamed colonic wall, irrespective of the underlying aetiology. Furthermore, the amount of thickening of the single wall layers in combination with absence/presence of paracolonic lymphnodes allowed differentiation between active UC and active CD. It was, however, not possible to differentiate IBD in remission from healthy controls.
The ability to image inflamed and non-inflamed colonic wall has already been shown in transabdominal US. In a meta-analysis, in which the diagnostic accuracy of different cut-off values in CD were compared, sensitivity and specificity of 88% and 93% were achieved when a TWT threshold >3 mm was applied.
In contrast to CD, TWT in UC could not be correlated with clinical disease activity in some previous studies. None of the currently published studies using transabdominal US were able to differentiate the different wall layers from each other and make a diagnosis between CD and UC.
In our study using the following criteria: (i) increased TWT, increased mucosal (ii) and submucosal (iii) thickness and (iv) presence of paracolonic lymphnodes, we were able to differentiate active UC from active CD with a sensitivity of 92.3%.
Even at the early stages of EUS, its ability to visualise the depth of the large bowel wall from the endoluminal side was proven. But, since the oblique-viewing echoendoscopes did not allow to be passed further than rectum or at best the lower sigmoid, EUS studies focussed on rectal cancer staging.
If the more proximal parts of the colon were to be evaluated, mechanical miniprobes were used through the accessory channel of a colonoscope, which allowed examination at the same time as colonoscopy with the advantage of 'one stop shopping', making further bowel preparation, sedation and another examination unnecessary. However, miniprobes were constructed in the 80s and early 90s, are mechanical and their crystals and imaging resolution have never updated. Since they did not allow EUS-FNA, they were not used much any more. The current electronic scanners are highly superior in resolution, ability to allow colour Doppler etc., so that miniprobes are largely mothballed. However, if isolated right sided colitis or ileitis should be investigated, it represents the only viable alternative.
If only the rectum needs to be evaluated, the commercially available stiff rectal probe will offer very good views for the same evaluation.
For this first feasibility study, we included only patients with active disease in the left colon as well as those in remission with history of left colon disease to ensure homogenous inclusion criteria in both groups. The echoendoscope used is by nature a 'gastroscope' used in the colon and a skilled endoscopist may be able to perform a full colonoscopy but this is certainly not true for an average examiner. But, the left colon is affected in many patients with CD and nearly all of the UC patients, which makes this method easily accessible for many of the IBD patients. No complications were observed in this study regarding the use of EUS in patients with active IBD. We therefore assume that complication rates would not differ from those doing colonoscopy in the same patients. This prototype front-view echoendoscope used in our study has an angle of view of 140 ° and similar forward propelling characteristics comparable to standard endoscopes ensuring a comparable view inside the colon throughout the examination. This is also true in regards of the stiffness of the tip of the scope and tip flexion. Consequently, safety and potential complications of EUS in patients with active IBD are unlikely to be different from standard colonoscopy. A study conducted by Navaneethan et al. showed that the prevalence of perforation in IBD patients was 0.3% (19/5295 patients).
Shimizu et al. related the increased depth of the different wall layers determined by EUS to active inflammation of IBD. Using miniprobe EUS, Higaki et al. explored the correlation between inflammatory changes of mucosa and submucosa in the rectum and relapse in 23 patients with UC in remission, compared to 14 healthy controls. The thickness of the mucosal/submucosal layers was significantly increased in patients with later relapse when compared to those remaining in remission.
However, initial efforts to demonstrate the ability of EUS to differentiate CD from UC were largely disappointing. As a consequence, EUS was not developed further for the evaluation of IBD. Meanwhile, forward-viewing instruments have become available, which provide high resolution imaging and might allow negotiation of colon loops. In our study using such an echoendoscope, we were able to precisely differentiate and measure all the wall layers individually. Patients with active UC had a prominent thickening of the mucosal layer, whereas patients with active CD had a significant thickening of the submucosal layer.
In our study, histology was used as reference for proof of disease and spread of inflammation. This showed a strong correlation between TWT, and mucosal and submucosal thickness in patients with IBD and HC and proved that our EUS diagnoses were viable using the thickness of the different wall layers as a measure. Other authors proved that the vertical spread of inflammation of UC was consistent with histopathological findings in 90% of the investigated cases. Hurlstone et al. evaluated the ability to predict the clinical severity of disease by the wall thickness in rectal disease and provided a strong correlation between inflammatory scores and wall thickness. The authors concluded that EUS of the colon could become a useful adjunctive examination for the transmural assessment of the colonic wall in UC. This might justify expensive and time consuming EUS developments since 'mucosal healing' has become a major target of interest especially for the evaluation of failure/success rates of the expensive TNF alpha therapy. Several transabdominal studies have already shown a reduction in TWT and mural vascularity after TNF alpha therapy in CD patients. Maintenance of wall thickness and increased mural vascularity were seen in all nonresponders.