Intestinal tuberculosis with small bowel stricture and hemorrhage as the predominant manifestation: Three case reports

BACKGROUND Intestinal tuberculosis is a chronic disease caused by Mycobacterium tuberculosis that mainly affects the ileum and cecum.Small bowel tuberculosis,characterized by predominant involvement of the small intestine,is an extremely rare condition with highly atypical clinical presentations,making diagnosis even more challenging.CASE SUMMARY We report three cases of small intestinal tuberculosis,two of the patients pre-sented primarily with abdominal pain,and one presented with gastrointestinal bleeding.All patients underwent blood tests and imaging examinations.Small bowel endoscopy(SBE)revealed that the main lesions in these patients were in-testinal stenosis or gastrointestinal bleeding caused by small intestinal ulcers.One patient ultimately underwent surgical treatment.Following a complex diagnostic process and comprehensive analysis,all patients were confirmed to have small intestinal tuberculosis and received standard antituberculosis treatment,leading to an improvement in their condition.CONCLUSION Patients with SBTs present with nonspecific symptoms such as abdominal pain,weight loss,and occasional gastrointestinal bleeding.Accurate diagnosis requires a thorough evaluation of clinical symptoms and various tests to avoid misdiagno-sis and complications.

Clinical profile,diagnostic yield,and procedural outcomes of single balloon enteroscopy:A tertiary care hospital experience

BACKGROUND Single balloon enteroscopy(SBE)allows ease of access for small bowel visualization and has multiple diagnostic and therapeutic indications.It provides the advantage of performing various therapeutic interventions alongside the diagnostic procedure.SBE has also been considered a relatively safe procedure with no major complications.AIM To investigate the indications,safety,and clinical yield of SBE,and determine its effect on disease outcome.METHODS A retrospective,descriptive study was conducted at a tertiary care hospital in Karachi,Pakistan.Medical records of 56 adult patients(≥18 years)who underwent SBE between July 2013 and December 2021 were reviewed and data were collected using a structured proforma.A descriptive analysis of the variables was performed using Statistical Package of Social Sciences Version 19.Results are reported as the mean±SD for quantitative variables and numbers and percentages for qualitative variables.Missing data are reported as unknown.RESULTS A total of 56 patients who underwent 61 SBE procedures were included.The mean age was 50.93±16.16 years,with 53.6%of them being males.Hypertension(39.3%)and diabetes mellitus(25.0%)were the most common pre-existing comorbidities.Obscure gastrointestinal bleed(39.3%)was the most common indication for enteroscopy,followed by chronic diarrhea(19.7%)and unexplained anemia(16.4%).The majority of procedures were performed in the endoscopy suite(90.2%)under monitored anaesthesia care(93.4%).Most procedures were diagnostic(91.8%)and completed without complications(95.1%).The depth of examination ranged from 95 cm to 500 cm with a mean of 282.05±90.04 cm.The most common findings were inflammation and ulcerations(29.5%),followed by masses(19.7%)and vascular malformations(14.8%).As a result of the findings,a new diagnosis was made in 47.5%of the cases and a previous one was ruled out in 24.6%of them;65.6%of the cases had a change in management.CONCLUSION SBE is a suitable modality for investigating diseases in the small bowel.It is shown to be technically efficient and reasonably safe and is associated with high diagnostic and therapeutic yield.

Detection of Angioectasias and Haemorrhages Incorporated into a Multi-Class Classification Tool for the GI Tract Anomalies by Using Binary CNNs

The proposed deep learning algorithm will be integrated as a binary classifier under the umbrella of a multi-class classification tool to facilitate the automated detection of non-healthy deformities, anatomical landmarks, pathological findings, other anomalies and normal cases, by examining medical endoscopic images of GI tract. Each binary classifier is trained to detect one specific non-healthy condition. The algorithm analyzed in the present work expands the ability of detection of this tool by classifying GI tract image snapshots into two classes, depicting haemorrhage and non-haemorrhage state. The proposed algorithm is the result of the collaboration between interdisciplinary specialists on AI and Data Analysis, Computer Vision, Gastroenterologists of four University Gastroenterology Departments of Greek Medical Schools. The data used are 195 videos (177 from non-healthy cases and 18 from healthy cases) videos captured from the PillCam(R) Medronics device, originated from 195 patients, all diagnosed with different forms of angioectasia, haemorrhages and other diseases from different sites of the gastrointestinal (GI), mainly including difficult cases of diagnosis. Our AI algorithm is based on convolutional neural network (CNN) trained on annotated images at image level, using a semantic tag indicating whether the image contains angioectasia and haemorrhage traces or not. At least 22 CNN architectures were created and evaluated some of which pre-trained applying transfer learning on ImageNet data. All the CNN variations were introduced, trained to a prevalence dataset of 50%, and evaluated of unseen data. On test data, the best results were obtained from our CNN architectures which do not utilize backbone of transfer learning. Across a balanced dataset from no-healthy images and healthy images from 39 videos from different patients, identified correct diagnosis with sensitivity 90%, specificity 92%, precision 91.8%, FPR 8%, FNR 10%. Besides, we compared the performance of our best CNN algorithm versus our same goal algorithm based on HSV colorimetric lesions features extracted of pixel-level annotations, both algorithms trained and tested on the same data. It is evaluated that the CNN trained on image level annotated images, is 9% less sensitive, achieves 2.6% less precision, 1.2% less FPR, and 7% less FNR, than that based on HSV filters, extracted from on pixel-level annotated training data.