Introduction

A differential diagnosis of inflammatory bowel diseases (IBD) and irritable bowel syndrome (IBS) is one of the unresolved problems in modern gastroenterology. The absence of clear criteria for IBD detection leads to delay in the diagnostic process for months and years.

Aims & Methods

The aim of the study was to identify clinical and laboratory markers that could distinguish between IBD and IBS with a subsequent development of an artificial neural network (ANN).
We retrospectively evaluated the clinical symptoms, anamnesis data and laboratory tests of patients with verified exacerbation of Crohn`s disease (CD) and ulcerative colitis (UC), and IBS. We found parameters that could distinguish IBD and IBS and then CD and UC, and used them to create the ANN.

Results

Of the 1006 examined, 32% were patients with IBS and 68% with IBD. Patients with IBS were significantly more likely to have a tendency to constipation (34.8% compared with 5.0% in IBD patients), loose stool up to 2 times (26.9% compared with 17.0%), the absence of extraintestinal (EIMs) (94.1% compared to 66.3%) and perianal manifestations (PAM) (96.0% compared to 83.5%) (p<0.001). The patients with IBD had diarrheal syndrome ≥ 5 times a day more frequently (40.0% and 6.2% in patients with IBS, respectively), blood admixture (65.3% and 6.8%, respectively) and higher level of fecal calprotectin (p<0.001). All these data were used to create and train the ANN on general and test (15% of all cases) dataset. Multilayer perceptron showed the best results in IBD detection: the sensitivity was 95.8% and the specificity was 90.2%.
To create the ANN for IBD differentiation, we analyzed medical histories of 350 patients (51%) with UC and 333 (49%) with CD. Statistical analysis in the group with UC significantly more often had severe diarrheal syndrome (≥ 5 times a day) (48.3% compared to 42.3% among the patients with CD), blood in the stool (87.1% compared to 79.0%, respectively), the absence of EIMs (73,4% compared to 58,9%, respectively) and PAMs (97,4% compared to 58,9%, respectively) (p<0,001). Patients with CD compared with UC were significantly more likely to have constipation (7.2% compared to 2.9%, respectively) (p=0.002), increased stool up to 2 times (21,6% compared to 12,6%, respectively), subfebrile temperature (25.8% compared to 16.9%, respectively) (p=0.004) and surgical intervention on the small and large intestines in the anamnesis (20.1% in compared to 1.4%, respectively) (p<0.001). The clinical and anamnestic parameters identified in our study were used to create and train the ANN. The best results were shown by a linear set with 5 input neurons: the accuracy of detecting CD was 78.7% and UC - 75.4%, which shows a moderate ability of the ANN to differentiate UC and CD.

Conclusion

As a result of the study, we created the ANNs with high sensitivity and specificity in the differentiation of organic and functional intestinal diseases and with moderate accuracy in IBD differentiation. The developed ANN can help clinicians in the diagnosis of IBD and IBS, but further work for improving its accuracy and the possibility of implementation into clinical practice is required.

Disclosure

All authors have declared no conflict of interest.

Introduction

Chronic mesenteric ischemia (CMI) is a condition in which the mesenteric blood flow fails to meet the increased metabolic demands (1), mostly caused by atherosclerotic stenosis of the mesenteric arteries. This leads to severe postprandial abdominal pain, resulting in fear of eating and subsequent weight loss. According to the European Guidelines(2), a significant stenosis is defined as a luminal diameter reduction of ≥50% in multivessel disease and ≥70% in single-vessel disease. In the absence of a gold standard diagnostic test, treatment decisions rely on multidisciplinary consensus, which is reached based on the patient’s clinical history, presence of a significant stenosis on computed tomography angiography (CTA) and in case of single-vessel disease a functional test. The definitive diagnosis CMI is confirmed after revascularization by patient reported symptom reduction. However, prior research report clinical improvement in only 69-73%(3), suggesting possible overtreatment. Currently, radiologists manually grade a stenosis by measuring the luminal diameter reduction on cross-sectional slices. Yet, other morphological features, such as stenosis length and shape(4-6), also influence blood flow. Three-dimensional (3D) analysis of mesenteric stenosis may better reflect the hemodynamic and clinical relevance of a mesenteric artery stenosis. 3D-analysis using semi-automated segmentation tools, such as 3D Slicer, allows for more comprehensive evaluation, including cross-sectional area reduction. This study aims to compare the positive predictive value (PPV) and negative predictive value (NPV) of diagnosing definitive CMI by using semi-automated 3D segmentation based on area reduction versus the conventional luminal diameter reduction-based grading performed by radiologists.

Aims & Methods

This multicenter, retrospective study analyzed baseline CTA scans from 94 patients (CoBaGI trial 2015-2019) with a consensus diagnosis of atherosclerotic CMI who underwent endovascular treatment. The CoBaGI trial(7) was a randomized controlled trial comparing the patency of bare-metal and covered stents. Of this cohort, 72 patients had a definitive diagnosis CMI compared to 22 non-CMI patients. CTA scans were excluded if there was no arterial phase, low image quality or if centerline extraction failed. Stenoses were segmented semi-automatically by using 3D Slicer. The centerline curves of the celiac artery (CA) and superior mesenteric artery (SMA) were extracted, and minimum and maximum diameters and cross-sectional areas were measured. Diameter reduction and area reduction were calculated and used to compute the PPV and NPV for the definitive diagnosis CMI for each grading method.

Results

Of the 94 baseline CTA scans, 53 CTA scans were eligible for analysis (36 definitive CMI patients, 17 non-CMI patients). The PPV for diameter reduction graded by radiologists was 0.72 (95% CI 0.56 - 0.82), compared to 0.78 (95% CI 0.63 - 0.89) for 3D segmentation. The NPV was 0.50 (95% CI 0.24 - 0.76) for diameter reduction graded by radiologists compared to 0.56 (95% CI 0.33 - 0.77) for 3D segmentation. When applying a threshold of 75% area reduction, the PPV improved to 0.79 (95% CI 0.65 - 0.89), and the NPV to 0.80 (95% CI 0.49 - 0.94).

Conclusion

Grading a mesenteric artery stenosis by using 3D-segmentation with area reduction provides a higher PPV and NPV for diagnosing definitive CMI compared to the conventional diameter reduction based grading. These findings suggest that incorporating 3D-segmentatons could improve patient selection for revascularization and reduce unnecessary procedures.

References

1. Moawad J, Gewertz BL. Chronic mesenteric ischemia. Clinical presentation and diagnosis. Surg Clin North Am. 1997;77(2):357-69.
2. Terlouw LG, Moelker A, Abrahamsen J, Acosta S, Bakker OJ, Baumgartner I, et al. European guidelines on chronic mesenteric ischaemia - joint UEG, EAGEN, ESGAR, NVMDL, HSG, CIRSE and DMIS clinical guidelines on the diagnosis and treatment of patients with chronic mesenteric ischaemia. United European Gastroenterol J. 2020;8(4):371-95.
3. van Dijk LJD, Moons LMG, van Noord D, Moelker A, Verhagen HJM, Bruno MJ, Rouwet EV. Persistent symptom relief after revascularization in patients with single-artery chronic mesenteric ischemia. J Vasc Surg. 2018;68(3):779-85.
4. Mustapha N, Mandal P, Johnston P, Norsarahaida A. A numerical simulation of unsteady blood flow through multi-irregular arterial stenosis. Applied Mathematical Modelling, 34, pp 1559-1573. 2010;34.
5. Sriyab S. The Effect of Stenotic Geometry and Non-newtonian Property of Blood Flow through Arterial Stenosis. Cardiovasc Hematol Disord Drug Targets. 2020;20(1):16-30.
6. Kamangar S, Badruddin I, Ahammad N, Kalimuthu G, Nik Ghazali NN, N.J S, et al. The Influence of Geometrical Shapes of Stenosis on the Blood Flow in Stenosed Artery. Sains Malaysiana. 2017;46:1923-33.

Introduction

Autoantibody detection is a cornerstone in the evaluation of Celiac Disease (CD); however, there is a limited offer of multiplexed testing devices.

Aims & Methods

We assessed the analytical performance of a novel, single-use, multiplexed microarray immunoassay (MosaiQ AiPlex^® CD microarray. AliveDx, Switzerland), used with its fully automated high-throughput proprietary system, for the qualitative detection of total IgA and semiquantitative detection of autoantibodies (IgA and IgG isotypes) against tissue transglutaminase (tTG) and deamidated gliadin peptide (DGP), compared with selected devices. The study was conducted at two sites. Site 1: North East Innovation Lab (The Newcastle upon Tyne Hospitals NHS Foundation Trust, UK), Site 2: AliveDx (Eysins, Switzerland). Included human serum samples were banked, de-identified and tested using the investigational device after characterization as reactive or non-reactive to the respective measurand by Phadia AB assays: tTG IgA (EliA Celikey tTG IgA), tTG IgG (EliA Celikey tTG IgG), DGP IgA (EliA Gliadin^DP IgA) and DGP IgG (EliA Gliadin^DP IgG). Comparator for total IgA was Tina-quant IgA Gen.2 (Roche Diagnostics Gmbh). Positive percent agreement (PPA) and negative percent agreement (NPA) were calculated. Double-sided 95% Confidence Interval (CI) were calculated using Clopper-Pearson Exact Method.

Results

Compared with the respective devices (equivocal results excluded), the investigational microarray showed PPA ranging from 95.7% to 100% and NPA ranging from 87.5 to 98.7%. Table 1 shows the performance of the individual measurands.

Conclusion

The investigational microarray immunoassay showed high level of agreement with the compared singleplex devices, providing laboratorians and clinicians with unified results coming from one multiplexed device. This platform has the potential to assist in the simplification of CD evaluation by simultaneously and automatically analyzing key serological markers.

Disclosure

This study was funded by AliveDx Suisse SA, Eysins, Switzerland. Gerber Gomez, Jose Santiago, Ewa Garyga, Rocio Pasion-Galvan, Helen Laird, Mark Hooper, Susan Seaton, Anthony Yau, Christian Fischer are AliveDx employees.

Introduction

Celiac disease is a chronic disease that requires lifelong treatment with a strict diet, which has an impact on the health-related quality of life of the patients.

Aims & Methods

To analyse the quality of life in patients with celiac disease that are followed in our centre.
A single-center, prospective, observational study based in the Spanish version of Celiac Disease Quality of Life Survey (CD-QOL), a validated questionnaire.

Results

53 patients agreed to complete the questionnaire. Mean age of the patients at the moment of the consultation was 39.15 years, while mean age at diagnosis of the patients who agreed to participate was 33.20 years. Of them, 35.8% had university degree, 28.3% completed professional training and 35.8% only had compulsory basic education.
Of these patients, 69.2% didn’t have any relative diagnosed with celiac disease, and only the 13.2% used to live with another celiac patient. Just 9.4% of the patients who filled in the questionnaire were members of a celiac patient society.
75.5% of the patients claimed that celiac disease affects their social life, 84.9% declared that they comply properly with the gluten-free diet, although for the 32.7% the adherence to the dietary treatment is difficult.
Of the patients who participated, 33.9% declared that they feel quite / a lot limited by the disease, 27.4% feel quite / much stigmatized because of the illness, 22.7% have quite / a lot of problems in their social life, 32.2% have quite / a lot difficulties travelling, and 22.7% of patients declared that they have quite / a lot the impression to not have a normal life.

Conclusion

1. Celiac disease may be quite burdensome even in asymptomatic patients who are not supposed to have small bowel damage in their follow-up. According to our series, up to 75.5% of patients have problems in their social life.
2. Health education during the follow-up providing greater knowledge of the disease, support from specialized nurses, and the recommendation to join patients support groups (patient societies) could be tools that may empower celiac patients to help manage their disease and enjoy a normal life.

References

1. Al-Toma A, Volta U, Auricchio R, et al. European Society for the Study of Coeliac Disease (ESsCD) guideline for coeliac disease and other gluten-related disorders. United European Gastroenterol J 2019;7(5):583-613.
2. Casellas F, Rodrigo L, Molina-Infante J, et al. Transcultural adaptation and validation of the Celiac Disease Quality of Life (CD-QOL) Survey, a specific questionnaire to measure quality of life in patients with celiac disease. Rev Esp Enferm Dig 2013; 105(10):585-93.

Introduction

Detection of autoantibodies directed to tissue transglutaminase (tTG) and deamidated gliadin peptide (DGP) is important in the assessment of individuals under clinical suspicion of celiac disease (CD).

Aims & Methods

In this study, we compared the analytical performance for the detection of autoantibodies to tTG (IgA isotype), of the MosaiQ AiPlex^® CD microarray (AiPlex CD. AliveDx, Switzerland), a CE-marked, novel, single-use, fully automated, multiplexed microarray immunoassay and routine methods. De-identified serum samples, characterized as reactive or non-reactive for tTG IgA as per the routine practice method of the evaluation site (EliA™ Celikey^® IgA, Phadia AB, Sweden), were tested with the evaluated device. Although AiPlex CD features the semi-quantitative measurement of both tTG IgA and IgG and anti-deamidated gliadin peptide (DGP) IgA and IgG antibodies, and qualitative measurement of total IgA in serum; this comparison was limited to tTG IgA as no samples characterized for other CD autoantibodies were available. Positive percent agreement (PPA) and negative percent agreement (NPA) were calculated. Double-sided 95% Confidence Intervals (CI) were calculated using Clopper-Pearson Exact Method.

Results

77 samples characterized as reactive for tTG IgA by the routine method and 90 samples characterized as non-reactive for this marker were included in the analysis. The evaluated device identified 73 out of 77 samples characterized as reactive by the routine method, for a PPA of 95% (95% CI: 87.2, 98.5). Out of the four discordant samples, one also yielded a non-reactive result for tTG IgG and reactive results for both DGP IgA and IgG in the evaluated device and negative anti-endomysial antibodies (EMA) by immunofluorescence; two samples were reactive to DGP IgA or DGP IgG by the evaluated device and one sample was non-reactive to all four markers and was negative to EMA. Regarding the samples characterized as non-reactive by the routine method, the evaluated device identified 86 out of 90, for an NPA of 96% (95% CI: 89.0, 98.7). For the four samples with discordant results, the evaluated device also yielded reactive results for at least one additional CD antibody (tTG IgG, DGP IgA, DGP IgG).

Conclusion

In this study, the evaluated microarray immunoassay showed high concordance with the compared routine method for the serological assessment of tTG IgA, a key marker in the evaluation of CD.

Disclosure

This study was funded by AliveDx Suisse SA, Eysins, Switzerland. Gerber Gomez and Christian Fischer are AliveDx employees.