An analysis of histopathological features of Crohn’s disease in surgical specimens

Abstract


Background
Gut in ammatory bowel disease (IBD) is a group of chronic gastrointestinal disorders characterized by relapsing and remitting idiopathic in ammation of the gastrointestinal tract.The two most common types of IBDs are ulcerative colitis and Crohn's disease (CD).The diagnosis of CD is con rmed by clinical evaluation and a combination of endoscopic, histological, radiological, and/or biochemical investigations [1].The European Crohn's and Colitis Organisation (ECCO) consensus has proposed a diagnostic criteria for CD in surgical specimens.CD can present as the following microscopic features in surgical specimens: transmural in ammation aggregated in ammatory pattern and transmural lymphoid hyperplasia; submucosal thickening (expansion by brosis-bromuscular obliteration and in ammation); ssure; sarcoid granuloma (including in lymph nodes); abnormalities of the enteric nervous system (submucosal nerve ber hyperplasia and ganglionitis); and relatively unchanged epithelia-mucin preservation [1].Thus it has been suggested that a diagnosis of CD can be made in surgically resected bowel samples when three of the above features are present in the absence of granulomas, or when an epithelioid granuloma is present with one other feature provided that speci c infections are excluded [1,2].However, although this set of diagnostic criteria plays an important role in CD diagnosis in surgically resected bowel specimens, its sensitivity and speci city have not been investigated.The incidence of CD has been steadily increasing in China in recent years [3].Large-scale multicenter studies examining the diagnostic sensitivity and speci city are needed to validate the above diagnostic criteria.
This study aimed to review the histologic features of CD in surgically resected bowel specimens and validate the sensitivity and speci city of a combination of the examined histologic features for CD diagnosis.

Study design and patients
We evaluated 171 and 215 patients who were diagnosed with CD and non-CD, respectively, between 2010 and 2015 and who underwent surgical bowel resection.The patients were identi ed using the pathology database at Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University.Those with tumor or acute intestinal infarction were excluded.All included patients underwent intestinal surgery and were followed up for at least 12 months.

Clinical diagnosis
CD diagnosis was con rmed by clinical, endoscopic, radiological, and histological examinations [1].The clinical diagnosis of infectious colitis was based on positivity of the pathogenic organism, and symptom relief and endoscopy healing after 6 months of antibiotic therapy without recurrence.Other causes of enteritis such as intestinal perforation, diverticulum, chronic obstruction, stula, mesenteric arteriovenous thrombosis, appendicitis, vascular malformation, Behçet's disease, abdominal cocoon, ectopia, mucosal prolapse, eosinophilic enteritis, and necrotizing enterocolitis were diagnosed based on clinicohistologic features.

Microscopic analysis
Surgical pathology reports and slides were retrieved and reviewed by two pathologists for the following histological features: transmural in ammation, transmural lymphoid aggregates, ssures and related structures, granulomas (including granulomas in lymph nodes) and granuloma-like nodules, abnormalities of the enteric nervous system, muscularis mucosae thickening, mucosal atrophy, and pseudopyloric gland metaplasia.They were de ned as follows: 1. Transmural in ammation: the presence of varying degree of chronic in ammatory cells such as lymphocytes and plasma cells in all layers of the bowel wall (Figure 1A).
2. Transmural lymphoid aggregates: the presence of lymphoid aggregates, with or without germinal centers, anywhere in the bowel wall, from the mucosa to the subserosa (Figure 1A).The linear distribution of lymphoid aggregates were referred to as "the string of beads" (or "Crohn's rosary") (Figure1B) [4,5,6].Transmural lymphoid aggregates were divided into two types according to the distribution: submucosa and beyond submucosa (involving muscularis propria or subserosa).Further, frequent lymphoid aggregates were de ned as 3 or more per low power eld [7].
3. Granulomas and granuloma-like nodules: a granuloma was de ned a discrete collection of at least ve epithelioid cells (activated histiocytes with homogeneous eosinophilic cytoplasm), with or without accompanying multinucleate giant cells, in the bowel wall or lymph nodes [8,9] (Figure 2A-B).
Caseation should be absent.Histiocytic reaction around a ruptured crypt or isolated multinucleate giant cell were excluded.The granuloma-like nodules were composed of non-epithelioid histiocytes without multinucleated giant cells and with clear boundaries.The granuloma-like nodules in the mucosa were excluded.(Figure 2C-F).
4. Fissures and related structures: ssures were identi ed as deep, ask-shaped, or knife-like ulcers that extended into the submucosa or muscularis propria and were lined by plump broblasts, many neutrophils, histocytes and even foreign body giant cells (Figure 3A-B).Abscess was a localized collection of pus surrounded by in amed tissue (Figure 3C).Sinuses were formed by ssuring ulcers extending into or through the colonic wall and usually communicating with other ssuring ulcers extending laterally to produce a complex network.Sinuses that did not demonstrate an opening to intestinal lumen in slides, but have long or complex con guration were referred to as "sinus-like structures" (Figure 3D) [4].Epithelium may be found in the sinus-like structures (Figure 3E) [4,10].
5. Abnormalities of the enteric nervous system: this was de ned as the presence of large, abnormal, irregular nerve bundles throughout the submucosa, muscularis propria or subserosa (Figure 4A-C).
Perineural chronic in ammation may be seen.
. Muscularis mucosae thickening: this was de ned as increased thickness of muscularis mucosae due to hyperplasia and/or brosis.In some cases, the submucosa was present (Figure 5A), whereas it was obliterated in some cases (Figure 5B).
7. Mucosal atrophy: this was de ned as chronic mucosal change such as villous atrophy and crypt distortion with chronic in ammatory cells (especially lymphocytes and plasma cells) in the lamina propria (Figure 5C).
. Pseudopyloric gland metaplasia: this was de ned as the presence of small clusters of small round glands in deep mucosa with clear neutral mucinous cytoplasm (Figure 5D).

Statistical analysis
The c2 and Fisher's exact probability tests were used to evaluate differences in the frequency of the histological parameters between the CD and the non-CD group.Statistical analysis comparing trends between different parameters was performed using Kendall-tau.All statistical analyses were performed using IBM SPSS Statistics Version 19.0 (IBM Corporation, Armonk, NY, USA).A p value of <0.05 was considered statistically signi cant.
Further analysis revealed that the presence of at least 3 of the 6 studied features (transmural in ammation, lymphoid aggregates in the muscularis propria or subserosa, granulomas or granulomalike nodules, ssures or sinus-like structures, abnormalities of the enteric nervous system, and mucosa structure alterations (muscularis mucosae thickening or mucosal atrophy with pseudopyloric gland metaplasia)) had a sensitivity of 98.8% and a speci city of 93.5% for distinguishing CD from non-CD cases (Table 3).Then, we simpli ed the 5 studied histological features other than transmural in ammation into two categories: (1) chronic in ammatory changes and (2) architectural abnormalities.Chronic in ammatory changes were characterized by the presence of granulomas/granuloma-like nodules, lymphoid aggregates in the muscularis propria or subserosa, and/or ssures/sinus-like structures.Meanwhile, architectural abnormalities were characterized by the presence of abnormalities of the enteric nervous system and/or mucosa structure alterations (muscularis mucosae thickening or mucosal atrophy with pseudopyloric gland metaplasia).A combination of transmural in ammation, chronic in ammatory changes, and architectural abnormalities had a sensitivity of 92.4% and speci city of 97.7% for distinguishing CD from non-CD cases.

Discussion
The diagnosis of CD is based on a combination of clinical, endoscopic, radiological, and histological parameters.Although the ECCO consensus had proposed diagnostic recommendations for Crohn's surgical specimens, no study has evaluated its diagnostic value [1].Although previous studies have reported the histological features in Crohn's surgical specimens [11,12], thus far, no study has compared the histological features between CD and non-CD.In the present study, we investigated the combination of histological features that can be useful for differential diagnosis of CD from other forms of enteritis.
Univariate analysis showed that transmural in ammation, lymphoid aggregates in the muscularis propria or subserosa, granulomas or granuloma-like nodules, ssures or sinus-like structures, abnormalities of the enteric nervous system, mucosa structure alterations (e.g., muscularis mucosae thickening or mucosal atrophy with pseudopyloric gland metaplasia) are diagnostic for CD, consistent with previous studies [11,12].Other features such as abscesses and submucosal lymphoid aggregates were of little value in distinguishing CD and non-CD.Granulomas are characteristic ndings in CD and have been reported to be present in 52-75% of resection specimens [11,13,14,15].In line with the literature, granulomas were found in 70.8% of CD cases (including 6 cases whose granulomas were only in the lymph nodes) in this study.
We also compared the frequency and characteristics of granuloma-like nodules between CD and non-CD cases and found that they were only present in CD, although the overall frequency was low (11.7%).
These results indicate that the presence of granuloma-like nodules in the submucosa and beyond in surgically resected specimens could indicate CD and thus the specimen should be thoroughly investigated for further evidence.
Hypertrophy of nerve bers accompanied by in ltration of chronic in ammatory cells is the most primary abnormality of the enteric nervous system.Although nerve hypertrophy is found in less than 50% of CD cases, it has high speci city as it only occurs in 3.3% of non-CD cases.Thickening of the muscularis mucosae by hyperplasia or brosis with or without submucosal obliteration is also common in surgical specimens of CD.The results showed that regardless of whether the submucosa was obliterated or not, muscularis mucosae thickening in a surgically resected bowel specimen should prompt careful search for more evidence of CD.
We further examined two combinations of the six histologic features examined in this study, namely, (1) transmural in ammation; (2) lymphoid aggregates in the muscularis propria or subserosa; (3) granulomas or granuloma-like nodules; (4) ssures or sinus-like structures; (5) abnormalities of the enteric nervous system; and (6) mucosa structure alterations (muscularis mucosae thickening or mucosal atrophy with pseudopyloric gland metaplasia).The results showed that the presence of at least 3 of the above 6 features in the surgically resected bowel specimens had a sensitivity of 98.8% and speci city of 93.5% for distinguishing CD from non-CD cases.
Further, we evaluated the diagnostic value of the combination of transmural in ammation, one chronic in ammatory change (i.e., granulomas/granuloma-like nodules, lymphoid aggregates in the muscularis propria or subserosa, or ssures or sinus-like structures), and one structural abnormality (i.e., abnormalities of the enteric nervous system or mucosa structure alterations that include muscularis mucosae thickening or mucosal atrophy with pseudopyloric gland metaplasia).This simpli ed combination had a sensitivity of 92.4% and a speci city of 97.7% in diagnosing CD.The 5 non-CD cases misclassi ed as CD using this combination included 2 cases of tuberculosis, 1 case of ulcerative colitis, 1 case of chronic obstruction, and 1 case of uncertain etiology.Chronic active Epstein-Barr virus infective enteritis may present with CD-like features such as transmural in ammation, ssuring ulcers, and lymphoid aggregates in intestinal wall [16].However, all ve cases of chronic active Epstein-Barr virus infective enteritis in this study were correctly classi ed as non-CD cases using this simpli ed combination.
The strengths of our study include the relatively large sample size from a single tertiary medical center with expertise in IBD care.This allowed for a relatively uniform tissue sampling of the surgically resected specimens.Further, all patients were followed up for at least 12 months.In addition, all slides were independently reviewed in detail by two pathologists experienced in IBD pathology.However, our study also had some limitations that need to be considered when interpreting the ndings.First, data availability and quality were limited to those available within the medical records.For example, some cases lacked detailed information on medical treatment prior to surgical bowel resection.In addition, macroscopic features were not thoroughly documented in the surgical pathology report and thus could not be included in this study.Further, all patients were from a tertiary medical center, and this may have introduced referral bias.Accordingly, conclusions from this study may not be applicable to other practice settings.Finally, slides were only reviewed by two pathologists.Despite these limitations, we believe that this study is valuable because it provides important information for CD diagnosis in surgically resected bowel specimens.Large-scale prospective studies are needed to con rm our ndings and to standardize the diagnosis of CD in surgically resected bowel specimens, which would be important and particularly relevant in developing countries where the incidence of CD is increasing but the pathological expertise needs to be improved.

Conclusion
Our study highlights the histological features prevalent in surgically resected CD bowel diseases.In addition to clinical, endoscopic, and radiographic information, a combination of histologic features including transmural in ammation, chronic in ammatory changes, and architectural alterations in surgically resected bowel specimens helps distinguish CD from non-CD.Granulomas and granuloma-like nodules.Granuloma is a discrete collection of epithelioid cells (activated histiocytes with homogeneous eosinophilic cytoplasm), with accompanying multinucleate giant cells HE×200 (A), or without accompanying multinucleate giant cells HE×40 (B).Granuloma-like nodule is a well-de ned nodule of non-epithelioid histiocytes in the muscularis propria (C HE×100, D HE×400) and subserosa (E HE×100, F HE×400). Fissures and related structures.Fissures are deep, ask-shaped, or knife-like ulcers that extend into the submucosa HE×20 (A) and is lined by numerous neutrophils HE×100 (B).Abscess is a localized collection of pus surrounded by in amed tissue HE×20 (C).Some sinuses do not demonstrate an opening to intestinal lumen on slides, but those with long or complex con gurations are considered a sinus-like structure HE×20 (D).Sometimes, epithelium lines the sinus-like structure HE×20 (E).

Figures
Figures

Table 1 .
Clinicodemographics and clinical diagnosis in CD and non-CD patientsThe histological features in CD and non-CD are shown in Table2.The frequency of transmural in ammation was signi cantly higher in CD resection specimens than in non-CD resection specimens

Table 2 .
Histological features of the surgically resected bowel specimens by group Crohn's disease; non-CD: other forms of enteritis except Crohn's disease Granulomas were signi cantly more present in CD cases than in non-CD cases (121/171 (70.8%) vs.
a : Fisher's exact probability tests Abbreviations: CD:

Table 3 .
Diagnostic value of two composite histological features in distinguishing CD from non-CD cases The chronic in ammatory changes were as follows: granulomas or granuloma-like nodules; lymphoid aggregates in the muscularis propria or subserosa; ssures or sinus-like structures b : c : The architectural abnormalities were as follows: abnormalities of the enteric nervous system; mucosa structure alterations (muscularis mucosae thickening or mucosal atrophy with pseudopyloric gland metaplasia) Abbreviations: CD: Crohn's disease; non-CD: other forms of enteritis except Crohn's disease; No.: number; Se: sensitivity; Sp: speci city; PPV: positive predictive value; NPV: negative predictive value