Pathologist-trained machine learning classifiers developed to quantitate celiac disease features differentiate endoscopic biopsies according to modified marsh score and dietary intervention response.

BACKGROUND: Histologic evaluation of the mucosal changes associated with celiac disease is important for establishing an accurate diagnosis and monitoring the impact of investigational therapies. While the Marsh-Oberhuber classification has been used to categorize the histologic findings into discrete stages (i.e., Type 0-3c), significant variability has been documented between observers using this ordinal scoring system. Therefore, we evaluated whether pathologist-trained machine learning classifiers can be developed to objectively quantitate the pathological changes of villus blunting, intraepithelial lymphocytosis, and crypt hyperplasia in small intestine endoscopic biopsies. METHODS: A convolutional neural network (CNN) was trained and combined with a secondary algorithm to quantitate intraepithelial lymphocytes (IEL) with 5 classes on CD3 immunohistochemistry whole slide images (WSI) and used to correlate feature outputs with ground truth modified Marsh scores in a total of 116 small intestine biopsies. RESULTS: Across all samples, median %CD3 counts (positive cells/enterocytes) from villous epithelium (VE) increased with higher Marsh scores (Type 0%CD3 VE = 13.4; Type 1-3%CD3 VE = 41.9, p < 0.0001). Indicators of villus blunting and crypt hyperplasia were also observed (Type 0-2 villous epithelium/lamina propria area ratio = 0.81; Type 3a-3c villous epithelium/lamina propria area ratio = 0.29, p < 0.0001), and Type 0-1 crypt/villous epithelial area ratio = 0.59; Type 2-3 crypt/villous epithelial area ratio = 1.64, p < 0.0001). Using these individual features, a combined feature machine learning score (MLS) was created to evaluate a set of 28 matched pre- and post-intervention biopsies captured before and after dietary gluten restriction. The disposition of the continuous MLS paired biopsy result aligned with the Marsh score in 96.4% (27/28) of the cohort. CONCLUSIONS: Machine learning classifiers can be developed to objectively quantify histologic features and capture additional data not achievable with manual scoring. Such approaches should be further investigated to improve biopsy evaluation, especially for clinical trials.

Duodenal mast cells and eosinophils in children with celiac disease: occurrence and distribution pattern.

OBJECTIVE: The aim of this study was to characterize duodenal mast cell (MC) and eosinophil (EO) numbers, their distribution within the lamina propria and possible impact on disease severity of paediatric celiac patients compared to children without celiac disease (CD). METHODS: We analysed duodenal samples of 215 children (109 CD, 106 controls) who underwent esophagogastroduodenoscopy from 2010 to 2018. After immunohistochemical staining, average MC and EO counts were histologically examined in ten high-power-fields. Additionally, cell-distribution within the lamina propria was analysed. Possible influence of relevant clinical parameters was evaluated. STATISTICS: Student's-t-test, Mann-Whitney U-test, Chi-square-test, ANOVA, significance-level <.05. Trial registration-number: DRKS00024669. RESULTS: MC-density was higher in CD-patients compared to the control-group (23.7 (±12.1)/HPF versus 19.7 (±9.1)/HPF; p = .008), varying in number interindividually. Eosinophils were also increased in the duodenum of celiac patients (23.3 (±9.3)/HPF versus 12.2 (±6.3)/HPF; p= <.001). MCs were distributed more often homogenously in all parts of CD lamina propria (44 biopsies (40.4%), residing more distant from the intestinal lumen in controls (0 biopsies with homogenous distribution-pattern (0%); p= <.001). Regarding EOs no polarity was observable. Atopic diseases did not occur significantly more often in patients with elevated EO-counts. CONCLUSION: MC- and EO-numbers were increased in the duodenum of CD-patients and MCs showed a different distribution-pattern in the lamina propria of celiac patients. These findings support the concept that both cell-types contribute to disease-pathogenesis. However, functional studies highlighting both cell-types' and their mediators' role regarding mucosal alterations during the course of the inflammatory process in celiac patients are needed. TRIAL REGISTRATION NUMBER AND URL: DRKS00024669; https://www.drks.de/drks_web/.

Serum Level and Gene Expression of Interleukin-15 Do Not Correlate with Villous Atrophy in Celiac Disease Patients.

Background and Aims: Interleukin-15 (IL-15) is a key player in the pathogenesis of celiac disease (CD). We investigated the functional role of IL-15 in the process of epithelial cell phenotypic modification at different stages of CD. Materials and Methods: In this study, we looked for correlations between the IL-15 mRNA levels in duodenal tissue and serum protein levels in a cohort of Iranian patients affected by CD based on the degree of histopathology. Ninety-five formalin-fixed, paraffin-embedded duodenal tissue specimens were collected: 23 with a Marsh I value; 30 with a Marsh II value; 32 with a Marsh III value; and 10 normal controls. The expression levels of the IL-15 gene in these biopsy specimens were determined by real-time quantitative polymerase chain reaction (qPCR), and IL-15 serum protein concentrations were determined by enzyme-linked immunosorbent assay and compared to tissue expression. Results: The IL-15 mRNA levels were higher in patients with Marsh II compared with the control group, and the Marsh I, and Marsh III groups. The differences between the Marsh II and Marsh I patients were statistically significant (p = 0.03). Similarly, the serum concentration of IL-15 was higher in Marsh II patients compared to those with Marsh I and Marsh III lesions, although the differences were not statistically significant (p = 0.221). Conclusions: Our results demonstrate that IL-15 gene expression might be elevated only in the early stages of CD onset (and histological damage) and that IL-15 serum levels do not significantly correlate with its tissue expression whatever the degree of histopathology.

Utility of narrow band imaging in predicting histology in celiac disease.

BACKGROUND: Narrow band imaging (NBI) with magnification better visualizes the duodenal microsurface and mucosal vascularity. NBI delineates villous atrophy better than conventional white light endoscopy. AIMS: This study was conducted to evaluate the diagnostic accuracy of narrow band imaging with magnification (NBI-ME) in celiac disease (CD). METHODS: In this prospective study, consecutive patients of suspected CD and controls were subjected to tissue transglutaminase antibody test and endoscopic evaluation initially with white light followed by NBI-ME, and biopsies were taken from duodenum. Duodenal villous patterns on NBI were interpreted as normal, blunted distorted, and absent. Severity of villous atrophy was reported according to the modified Marsh criteria. RESULTS: One hundred and twenty-two patients (mean age of 27.53 ± 13.37 years and a male to female ratio of 1:1.26) and 40 controls were studied. The sensitivity and specificity of NBI-ME in predicting villous atrophy were found to be 95.54% and 90%, respectively. The specificity and negative predictive value of NBI-ME in predicting villous atrophy amongst controls was 100% and 97.5%, respectively. Abnormal findings (blunted and absent villous patterns) combined with elevated transglutaminase antibody (> 5-fold) were found to have high accuracy in predicting villous atrophy. CONCLUSION: NBI with magnification has high sensitivity and specificity in predicting villous atrophy in patients with celiac disease.

Correlation of TTG IgA Level with Small Intestinal Histopathological Changes for Celiac Disease among Adult Saudi Patients.

BACKGROUND AND OBJECTIVES: According to recent guidelines, a diagnosis of celiac disease (CD) can be made without a biopsy, especially in children. There are no enough studies despite high prevalence and differences in genetic, race, and cultures. Therefore, we examined the correlation between tissue transglutaminase (TTG) and duodenal biopsy changes in our region because we are identical and different from others in culture, environment, and habits, and the correlation is same as that in different regions. METHODS: A retrospective cohort study at the Ministry of National Guard Health Affaires (NGHA) health care facilities that are distributed throughout kingdom of Saudi Arabia from April 19, 2015, till March 29, 2018. This study used the BESTCARE system that includes data from all NGHA facilities; data from 513 patients with CD were collected. All patients diagnosed with celiac disease aged 15 years or more, confirmed by improvement on gluten-free diet (GFD), and were not on GFD before endoscopy or serology test or both of them were included in the study, and the TTG IgA level was measured at the same time or within 2-3 months of biopsy date. The exclusion criteria were negative duodenal biopsy, which is less than 2; patients with negative biopsy and negative serology; patients who were on GFD before testing, and any patients known to have immunity diseases or illness causing mucosal changes. The TTG IgA level was measured in IU/ mL and was labeled as negative (<20 IU/mL) and positive (≥ 20 IU/mL) based on the cutoff value. However, Intestinal biopsy findings were identified as Marsh classification groups. RESULTS: One hundred thirty-four patients who met the inclusion criteria were included in the study. Median age of our sample was 24 years (16-37 years). Among these, 99 (73.88%) were female patients, whereas male patients were only 35 (26.12%). Histopathologic investigation of intestinal biopsy were Marsh 0 group was 16 cases (11.9%), Marsh 1 group was 8 cases (6%), Marsh 2 group was 4 cases (3%), Marsh 3a group was 32 cases (23.9%), Marsh 3b group was 64 cases (47.8%), and Marsh 3c group was 10 cases (7.5%). The TTG IgA antibody serology groups were <20 IU/mL in 13 cases (9.7%) and ≥20 IU/mL in 121 cases (90.3%). Among all patients with CD who had negative biopsy (Marsh 0 group), 16 (100%) of them had positive TTG IgA antibody. However, among patients with Marsh 1 group biopsy, 5 (62.5%) cases had negative TTG IgA antibody compared with 3 (37.5%) positive cases. Of the four cases (100%) with Marsh 2 group, all of them had positive TTG IgA antibody. However, in Marsh 3a group biopsy, 3 (9.4%) cases had negative TTG IgA antibody compared with 29 (90.6%) cases with positive TTG IgA antibody. Furthermore, among the patients with Marsh 3b group biopsy, 5 (7.8%) had negative antibody and 59 (92.2%) had positive serology. Of all biopsies of Marsh 3c group, 10 (100 %) had positive TTG IgA antibody. CONCLUSIONS: In perspective of high prevalence of CD in KSA, even more than western countries, we can pretend that positive TTG antibody tests can be applied for the diagnosis of CD without biopsy, particularly in symptomatic patients along with high titer, that is, 5-10 times the upper limit of normal (ULN). However, to validate it further, we need larger prospective studies in which duodenal biopsies should be taken according to recommended protocol and should be interpreted by experienced pathologist. Furthermore, biopsy is still needed in patients who do not show clinical improvement on a gluten-free diet and in cases with mildly or moderately elevated TTG IgA.

Helicobacter pylori infection in children with celiac disease: Multi-center, cross-sectional study.

BACKGROUND: An inverse association has been suggested between celiac disease (CD) and Helicobacter pylori (Hp) infection in children; however, there are inconsistent data. The purpose of this multi-center study is to evaluate the association between Hp and CD in childhood. METHODS: Children who underwent endoscopy between July 2016 and November 2017 in four pediatric gastroenterology centers were included in the study. Patients with a history of previous Hp eradication, antibiotic or acid-suppressive drug therapy in the last 4 weeks, and any underlying chronic disease were excluded. The presence of Hp infection was confirmed by both histopathology and the rapid urease test. The ones who had the diagnosis of CD were compared with the children who underwent endoscopy during the same period and had another diagnosis. Duodenal histopathology of children with CD was categorized according to the modified Marsh classification. RESULTS: Of 3056 endoscopies performed in the study period, 2484 cases were eligible for the study. A total of 482 CD patients (mean age: 9.71 ± 4.63 years, 58.5% girls) and 2060 controls (mean age: 9.92 ± 4.66 years, 54.6% girls) were included in the study. The rate of Hp infection was significantly lower in CD group (26.3% vs 50.1%, P < .01). The difference was prominent even in children younger than 6 years old (P < .01). There was no correlation between Hp infection and the modified Marsh scores in CD (P > .05). CONCLUSION: In this cross-sectional study, where Hp infection is common even in the pediatric population, the frequency of Hp infection was significantly lower in children with CD compared with the controls. Systematic cohort studies are necessary to clarify causal association between Hp infection and the development of celiac disease.

Diagnostic pitfalls of celiac disease.

Celiac disease is a very common autoimmune disorder caused by the ingestion of dietary gluten products in genetically susceptible persons. Its global prevalence is estimated around 1 %. However, the most cases are not diagnosed. Clinical presentation is widely variable with the involvement of various human systems. Besides a clinical picture (that is often non characteristic), the diagnosis is based on positivity of serological testing (tissue transglutaminase autoantibodies) and histological evaluation of small intestinal mucosa. The article presents a rational diagnostic approach to celiac disease. Key words: celiac disease - Marsh classification - tissue transglutaminase autoantibodies - villous atrophy.

Relationships between Clinical Presentation, Serology, Histology, and Duodenal Deposits of Tissue Transglutaminase Antibodies in Pediatric Celiac Disease.

BACKGROUND: The clinical, histological, and serological spectrum of celiac disease (CD) vary widely. We aimed to examine relationships between symptoms, serum anti-tissue transglutaminase antibodies (tTG) levels, mucosal damage, and mucosal anti-tTG deposits in pediatric CD. METHODS: A retrospective single-center, cohort study of children referred for endoscopy with suspected CD during 2011-2014. We retrieved the clinical data, blindly reviewed duodenal biopsies, and performed immunohistochemical staining for anti-tTG deposits. Patients were classified as monosymptomatic or polysymptomatic. Mucosal anti-tTG deposits were classified according to the location of deposits, dominant intensity, maximal intensity, and percentage of stained area. RESULTS: Of 252 patients with confirmed CD, complete data were available for 100: 37 males in the age range 1.3-16.7 with median 4.0 years. Monosymptomatic patients (n = 54) presented at an older age than polysymptomatic patients (1.3-15.5, median 8.1 vs. 1.3-16.7, median 6.3 years, p = 0.026). Marsh 2-3c was more prevalent in polysymptomatic patients (93 vs. 78%, p = 0.028). The intensity of mucosal anti-tTG deposits correlated with serum anti-tTG levels but not with the clinical presentation. CONCLUSIONS: Multiple symptoms and high serum anti-tTG antibody levels correlated with mucosal damage in children with CD. The role of immunohistochemical staining for intestinal anti-tTG mucosal deposits in the diagnosis of borderline CD is not yet established.

Evolutionary Developments in Interpreting the Gluten-Induced Mucosal Celiac Lesion: An Archimedian Heuristic.

The evolving history of the small intestinal biopsy and its interpretation-and misinterpretations-are described in this paper. Certain interpretative errors in the technical approaches to histological assessment are highlighted-even though we may never be rid of them. For example, mucosal "flattening" does not reduce individual villi to their cores, as still seems to be widely believed. Neither is the mucosa undergoing an atrophic process-since it can recover structurally. Rather, the intestinal mucosa manifests a vast hypertrophic response resulting in the formation of large plateaus formed from partially reduced villi and their amalgamation with the now increased height and width of the inter-villous ridges: this is associated with considerable increases in crypt volumes. Sections through mosaic plateaus gives an erroneous impression of the presence of stunted, flat-topped villi which continues to encourage both the continued use of irrelevant "atrophy" terminologies and a marked failure to perceive what random sections through mosaic plateaus actually look like. While reviewing the extensive 40+ year literature on mucosal analysis, we extracted data on intraepithelial lymphocytes (IEL) counts from 607 biopsies, and applied receiver-operating characteristic (ROC)-curve analysis. From that perspective, it appears that counting IEL/100 enterocyte nuclei in routine haematoxylin and eosin (H&E) sections provides the most useful discriminator of celiac mucosae at histological level, with an effective cut-off of 27 IEL, and offering a very high sensitivity with few false negatives. ROC-curve analysis also revealed the somewhat lesser accuracies of either CD3+ or γδ+ IEL counts. Current official guidelines seem to be somewhat inadequate in clearly defining the spectrum of gluten-induced mucosal pathologies and how they could be optimally interpreted, as well as in promoting the ideal manner for physicians and pathologists to interact in interpreting intestinal mucosae submitted for analysis. Future trends should incorporate 3-D printing and computerised modelling in order to exemplify the subtle micro-anatomical features associated with the crypt-villus interzone. The latter needs precise delineation with use of mRNA in-section assays for brush border enzymes such as alkaline phosphate and esterase. Other additional approaches are needed to facilitate recognition and interpretation of the features of this important inter-zone, such as wells, basins and hypertrophic alterations in the size of inter-villous ridges. The 3-D computerised models could considerably expand our understandings of the microvasculature and its changes-in relation both to crypt hypertrophy, in addition to the partial attrition and subsequent regrowth of villi from the inter-villous ridges during the flattening and recovery processes, respectively.

Validity of histology for the diagnosis of paediatric coeliac disease: a Swedish multicentre study.

OBJECTIVE Histological evaluation of intestinal biopsies for the diagnosis of coeliac disease can be challenging and compatible with risk of misdiagnosis. The aim was to evaluate the agreement of pathological diagnosis for coeliac disease in children investigated at four major paediatric university hospitals in Sweden. MATERIALS AND METHODS Intestinal duodenal biopsies were collected from 402 children at median 9.7 years (1.4-18.3 years). A pathologist at each hospital performed the primary evaluation. A designated pathologist, blinded to the primary evaluation, performed a second Marsh classification of biopsies (M0 to M3c) taken from the bulb and duodenum separately. Kappa (κ) scores between first and second evaluation determined the agreement. Plasma samples were collected at the day of intestinal biopsy and analysed for tissue transglutaminase autoantibodies (tTGA) using radioligand-binding assays. RESULTS Marsh scores were concordant in 229/356 biopsies (64%, κ = 0.52, p < 0.0001). Among discordant results, 15/127 (12%) showed M0 in distal duodenum but ≥ M2 in the bulb, whereas the opposite was true for 8/127 (6%) of the biopsies. There were fewer collected duodenal biopsies, more missing bulb biopsies and missing CD3 staining among discordant evaluations. The second evaluation revealed a Marsh score compliant with coeliac disease in 22 children of whom seven children were tTGA positive. CONCLUSIONS The variation between university hospitals on the pathological evaluation of biopsies may lead to misdiagnosis of coeliac disease in paediatric patients. Access to clinical and endoscopic information as well as tTGA levels may be useful for the pathologist to complement the evaluation in dubious cases.

Immunohistochemical CD3 staining detects additional patients with celiac disease.

AIM: To investigate whether performing immunohistochemical CD3 staining, in order to improve the detection of intra-epithelial lymphocytosis, has an additional value in the histological diagnosis of celiac disease. METHODS: Biopsies obtained from 159 children were stained by hematoxylin and eosin (HE) and evaluated using the Marsh classification. CD3 staining was subsequently evaluated separately and independently. RESULTS: Differences in evaluation between the routine HE sections and CD3 staining were present in 20 (12.6%) cases. In 10 (6.3%) patients the diagnosis of celiac disease (Marsh II and III) changed on examination of CD3 staining: in 9 cases, celiac disease had initially been missed on the HE sections, while 1 patient had been over-diagnosed on the routine sections. In all patients, the final diagnosis based on CD3 staining, was concordant with serological results, which was not found previously. In the other 10 (12.3%) patients, the detection of sole intra-epithelial lymphocytosis (Marsh I) improved. Nine patients were found to have Marsh I on CD3 sections, which had been missed on routine sections. Interestingly, the only patient with negative serology had Giardiasis. Finally, in 1 patient with negative serology, in whom Marsh I was suspected on HE sections, this diagnosis was withdrawn after evaluation of the CD3 sections. CONCLUSION: Staining for CD3 has an additional value in the histological detection of celiac disease lesions, and CD3 staining should be performed when there is a discrepancy between serology and the diagnosis made on HE sections.

The Value of Ensari's Proposal in Evaluating the Mucosal Pathology of Childhood Celiac Disease: Old Classification versus New Version.

OBJECTIVE: Small intestinal biopsy remains the gold standard in diagnosing celiac disease (CD); however, the wide spectrum of histopathological states and differential diagnosis of CD is still a diagnostic problem for pathologists. Recently, Ensari reviewed the literature and proposed an update of the histopathological diagnosis and classification for CD. MATERIALS AND METHODS: In this study, the histopathological materials of 54 children in whom CD was diagnosed at our hospital were reviewed to compare the previous Marsh and Modified Marsh-Oberhuber classifications with this new proposal. RESULTS: In this study, we show that the Ensari classification is as accurate as the Marsh and Modified Marsh classifications in describing the consecutive states of mucosal damage seen in CD. CONCLUSIONS: Ensari's classification is simple, practical and facilitative in diagnosing and subtyping of mucosal pathology of CD.