Evaluation of Optimal Biopsy Location for Assessment of Histological Activity, Transcriptomic and Immunohistochemical Analyses in Patients With Active Crohn's Disease

Gregor Novak; Toer Stevens; Tanja Van Viegen; Peter Bossuyt; Borut Štabuc; Jenny Jeyarajah; Guangyong Zou; Ingrid C. Gaemers; Trevor D. McKee; Fred Fu; Lisa M. Shackelton; Reena Khanna; Gijs R. van den Brink; William J. Sandborn; Brian G. Feagan; Rish K. Pai; Vipul Jairath; Niels Vande Casteele; Geert D'Haens


Aliment Pharmacol Ther. 2019;49(11):1401-1409. 

In This Article


Although the first histological index for assessment of disease activity in IBD was described more than 60 years ago,[31] there is surprisingly no consensus on the optimal location for biopsy procurement. To our knowledge, this is the first study to assess the relationship between biopsy location relative to ulcers and histological disease activity scores, proinflammatory gene expression levels and number of inflammatory cells in patients with Crohn's disease. Our main findings were that biopsies from the ulcer edge in both the colon and ileum yield the highest histological disease activity scores and proinflammatory gene expression levels, and that a gradient of inflammation exists from the ulcer edge to the most distant location evaluated (21-24 mm) for these assessments. Furthermore, no differences in histological disease activity scores or proinflammatory gene mRNA expression levels were found between the colon and ileum based on biopsy location. Although a similar pattern of results was observed for the number of myeloperoxidase-positive cells in the lamina propria, which mostly represent neutrophils[32,33] and are associated with active inflammation in IBD,[20] CD3+ (T lymphocytes)[34] and CD68+ (macrophages/monocytes)[35] cells were not more abundant at the ulcer edge in either the colon or ileum, and we generally did not observe consistent differences between locations. The implications for these findings are unknown and require additional research on the effect of biopsy location on cellular composition in Crohn's disease.

The medical management of Crohn's disease has been advanced in the past decade by the addition of multiple new agents[36,37] and treatment algorithms[3,38,39] that have allowed consideration of more rigorous definitions of remission beyond endoscopic mucosal healing, including histological remission.[9] However, the STRIDE consensus guidelines do not currently consider histological remission a potential treatment target for Crohn's disease in clinical practice due to insufficient evidence.[7] One explanation for this circumstance is the uncertainty regarding the optimal approach to outcome measurement. Specifically, no valid definition of histological remission has been established,[7,9] likely due to an incomplete understanding of Crohn's disease pathophysiology and the resultant lack of fully validated histological disease activity scoring indices.[10,11] For assessment of histological disease activity in this study, we included the partially validated and widely used GHAS,[10,11] and with modification to assess the percentage of affected surface in one biopsy rather than the number of biopsies affected. Although this is a substantial modification of the GHAS, we believe it has a minor impact on the total score. We also utilised two histological indices recently developed and validated for the assessment of ulcerative colitis disease activity, the RHI and NHI,[19,20] although neither index has been validated for use in patients with Crohn's disease. However, disease activity in both Crohn's disease and ulcerative colitis is defined by the degree of neutrophilic inflammation and surface injury (ulcers and erosions). The RHI also separately evaluates the degree of increased lamina propria lymphoplasmacytic inflammation, another feature common to both Crohn's disease and ulcerative colitis. Further research will be required to assess the reliability and responsiveness of these indices in Crohn's disease and this study provides a preliminary basis for these investigations.

Standardisation and optimisation of biopsy sampling protocols that take the potential anatomic heterogeneity and segmental nature of intestinal inflammation in Crohn's disease into account have been emphasised as important for reducing sampling error.[11] Historically, biopsies have been procured from the most macroscopically inflamed area, but empiric data to support this approach are lacking.[11] Inconsistencies in biopsy sampling are evident in recent trials that included histological disease activity assessment.[12,13] In the Fitzroy phase two trial of filgotinib for Crohn's disease, 12 biopsies were procured from the least and worst affected site in six intestinal segments. In segments with endoscopically normal mucosa, biopsies were obtained at the investigator's discretion.[13] Similarly, in the adalimumab EXTEND study, up to 10 biopsy specimens were collected (two from each segment). If areas of active disease were present, the samples were procured from those locations.[12] Our study shows that biopsies from the edge of an ulcer in the colon or ileum yield significantly higher histological disease activity scores than biopsies from clearly non-ulcerated mucosa (P < 0.0001). These results might have been expected as erosions/ulcers are included as component items of all three histological indices used in our study and add considerable weight to the total index scores. However, histological disease activity scores were also significantly different in biopsies taken from non-ulcerated mucosa in the ileum, and nonsignificant trends were similarly observed in the colon in non-ulcerated mucosa. Although our study was not designed to address the optimal biopsy location for the assessment of changes in inflammatory activity after a therapeutic intervention, it is plausible that responsiveness to change might be greater in areas with low histological disease activity (non-ulcerated mucosa) relative to those observed at the ulcer edge, or indeed, vice-versa. Additional research is needed to determine appropriate biopsy location in non-ulcerated mucosa.

Of note, we also observed no difference in proinflammatory gene mRNA expression levels in the colon or ileum according to biopsy location. On initial consideration, this observation seems contradictory to the findings of a previous study that showed different genetic signatures based on disease location.[40] However, these divergent observations may be reconciled by the existence of a final common pathway once the inflammatory cascade is initiated in the colon and ileum despite the existence of different initial molecular substrates at the time of disease initiation.

Analysis of proinflammatory gene expression may become a powerful objective tool in clinical practice and trials of patients with Crohn's disease. In addition to our work, this notion is supported by Boland et al's work, who confirmed that measuring proinflammatory gene expression from mucosal biopsies is a feasible and reproducible method to measure site-specific inflammation in Crohn's disease, and that expression of some inflammatory genes correlates with endoscopic disease activity.[14,41] Decreased mucosal expression of proinflammatory genes has been observed with successful treatment with infliximab and adalimumab,[22,37,41] and normalisation of proinflammatory gene expression has been shown to predict long-term remission with successful treatment with TNF antagonists.[25] Boland et al have also suggested that it may be possible to analyse gene expression in biopsies sampled from the distal colon as a surrogate proxy for more proximal disease activity, potentially avoiding complete colonoscopy, especially in patients with Crohn's colitis and rectal inflammation.[14]

Our study has both strengths and limitations. To our knowledge, this is the first study of its nature to be undertaken in Crohn's disease, and our results add important information for both standardisation of biopsy procurement procedures, as well as novel findings regarding the use of the RHI and NHI for the assessment of Crohn's histological disease activity that support further validation for their use in this disease. The multi-centre nature of our study increases the generalisability of the findings. Also of note, ileocolonoscopies were video-recorded and a blinded central reader confirmed correct biopsy procurement before biopsies were included in the analyses. Some limitations should also be acknowledged. First, we did not account for concomitant medication in our analysis. Although all patients had confirmed endoscopically active disease, the results for certain genes could be influenced by drug-specific mechanisms of action (eg expression of IL-23 with ustekinumab). Second, data for ileal and colonic biopsies were pooled from patients with different historical disease location according to Montreal classification (ie ileal biopsies from patients with ileal [L1] and ileocolonic [L3] disease and colonic biopsies from patients with colonic [L2] and ileocolonic [L3] disease). However, we did not find any statistical difference in any of the parameters assessed in this study between ileal or colonic samples procured from patients with ileocolonic disease (L3) vs those procured from patients with purely ileal (L1) or purely colonic (L2) disease (data not shown). Third, data from patients with post-operative recurrence of Crohn's disease at the neoterminal ileum were pooled with data from patients with ileal Crohn's disease who had not previously undergone surgery; populations with potentially different disease aetiopathogenesis. Of the 33 patients who had ileal biopsies, 13 had samples taken from the neoterminal ileum. No differences were observed for any of the study parameters in biopsies procured from the ileum or neoterminal ileum (Table S4). However, these findings should be interpreted with caution as the analyses are post hoc with limited sample size. Fourth, it is unknown whether taking multiple biopsies from the edge of ulcers in patients with Crohn's disease adds to the risk of the procedure with respect to bleeding and perforation. Although we demonstrated that two to four biopsies from the edge of ulcers were safe in this study, the safety of obtaining more than two to four biopsies from the edge of ulcers in patients with Crohn's disease has not been studied. Finally, ulcers are dynamic entities that may alternately exist at any time in states of progression (worsening) and/or healing. Our results may not necessarily be generalisable to these different disease states. Additional research is needed to evaluate histopathological processes and biomarkers involved in tissue repair.

In conclusion, biopsy procurement from the edge of an ulcer in Crohn's disease consistently yielded the highest histological disease activity scores and mRNA expression levels of the proinflammatory genes of interest in this study. We believe these results will help to standardise histological disease activity assessment in Crohn's disease, both for use in clinical trials and clinical practice.