CAD-CAP: a 25,000-image database serving the development of artificial intelligence for capsule endoscopy.

Background and study aims Capsule endoscopy (CE) is the preferred method for small bowel (SB) exploration. With a mean number of 50,000 SB frames per video, SBCE reading is time-consuming and tedious (30 to 60 minutes per video). We describe a large, multicenter database named CAD-CAP (Computer-Assisted Diagnosis for CAPsule Endoscopy, CAD-CAP). This database aims to serve the development of CAD tools for CE reading. Materials and methods Twelve French endoscopy centers were involved. All available third-generation SB-CE videos (Pillcam, Medtronic) were retrospectively selected from these centers and deidentified. Any pathological frame was extracted and included in the database. Manual segmentation of findings within these frames was performed by two pre-med students trained and supervised by an expert reader. All frames were then classified by type and clinical relevance by a panel of three expert readers. An automated extraction process was also developed to create a dataset of normal, proofread, control images from normal, complete, SB-CE videos. Results Four-thousand-one-hundred-and-seventy-four SB-CE were included. Of them, 1,480 videos (35 %) containing at least one pathological finding were selected. Findings from 5,184 frames (with their short video sequences) were extracted and delimited: 718 frames with fresh blood, 3,097 frames with vascular lesions, and 1,369 frames with inflammatory and ulcerative lesions. Twenty-thousand normal frames were extracted from 206 SB-CE normal videos. CAD-CAP has already been used for development of automated tools for angiectasia detection and also for two international challenges on medical computerized analysis.

Efficacy and Safety of Tumor Necrosis Factor Antagonists in Treatment of Internal Fistulizing Crohn's Disease.

BACKGROUND & AIMS: Few data are available on the effects of tumor necrosis factor (TNF) antagonist therapy for patients with internal fistulizing Crohn's disease (CD) and there is debate regarding the risk of abscess. We aimed to assess the long-term efficacy and safety of anti-TNF therapy for patients with internal fistulas. METHODS: We performed a retrospective study of data collected from the Groupe d'Etude Thérapeutique des Affections Inflammatoires Digestives trial, from January 1, 2000, through December 31, 2017. Our final analysis included 156 patients who began treatment with an anti-TNF agent for CD with internal fistula (83 men; median disease duration, 4.9 y). The primary end point was the onset of a major abdominal surgery. Secondary analysis included disappearance of the fistula tract during follow-up evaluation and safety. The Kaplan-Meier method was used for statistical analysis. RESULTS: After a median follow-up period of 3.5 years, 68 patients (43.6%) underwent a major abdominal surgery. The cumulative probabilities for being surgery-free were 83%, 64%, and 51% at 1, 3, and 5 years, respectively. A concentration of C-reactive protein >18 mg/L, an albumin concentration <36 g/L, the presence of an abscess at the fistula diagnosis, and the presence of a stricture were associated independently with the need for surgery. The cumulative probabilities of fistula healing, based on imaging analyses, were 15.4%, 32.3%, and 43.9% at 1, 3, and 5 years, respectively. Thirty-two patients (20.5%) developed an intestinal abscess and 4 patients died from malignancies (3 intestinal adenocarcinomas). One patient died from septic shock 3 months after initiation of anti-TNF therapy. CONCLUSIONS: In a retrospective analysis of data from a large clinical trial, we found that anti-TNF therapy delays or prevents surgery for almost half of patients with CD and luminal fistulas. However, anti-TNF therapy might increase the risk for sepsis-related death or gastrointestinal malignancies.

Confocal laser endomicroscopy: a new gold standard for the assessment of mucosal healing in ulcerative colitis.

BACKGROUND AND AIM: Endoscopic assessment of mucosal healing in ulcerative colitis (UC) is increasingly accepted as a measure of disease activity, therapeutic goal, and the key prognostic indicator. While regular endoscopy evaluates appearance of the mucosal surface, confocal laser endomicroscopy (CLE) enables in vivo visualization of subepithelial mucosa at 1000× magnification during ongoing endoscopy. Our aims were to determine using CLE whether endoscopically normal appearing colonic mucosa in patients with UC in remission (UC-IR) has fully regenerated mucosal structures, resolved inflammation, and to identify the mechanisms. METHODS: Twelve patients (six controls and six with UC-IR) underwent colonoscopy using CLE and intravenous fluorescein infusion. During colonoscopy, CLE images of colonic mucosa and conventional mucosal biopsies were obtained and evaluated using image-analysis systems. We quantified; (i) regeneration of colonic crypts and blood microvessels; (ii) cyclooxygenase 2 (COX2) expression; (iii) mitochondrial DNA (mtDNA) mutations; (iv) inflammatory infiltration; and (v) vascular permeability (VP). RESULTS: In control subjects, CLE demonstrated normal colonic crypts and microvasculature. COX2 expression was minimal, and < 7% crypts showed mtDNA mutations. Colonic mucosa of UC-IR patients had impaired and distorted crypt regeneration, increased COX2, 69% crypts with mtDNA mutations, persistent inflammation, and abnormal vascular architecture with increased VP (all P < 0.001 vs normal mucosa). CONCLUSIONS: (i) Endoscopically normal appearing colonic mucosa of patients with UC-IR remains abnormal: CLE demonstrates impaired crypt regeneration, persistent inflammation, distinct abnormalities in angioarchitecture and increased vascular permeability; molecular imaging showed increased COX2 and mtDNA mutations; (ii) CLE may serve as a new gold standard for the assessment of mucosal healing in UC.

Multidrug resistance gene-1 polymorphisms and resistance to cyclosporine A in patients with steroid resistant ulcerative colitis.

BACKGROUND: Cyclosporine A (CsA) is inconstantly effective in inducing remission in acute attacks of ulcerative colitis (UC) not responding to steroids. This study aimed to establish whether multidrug resistance gene (MDR)1 polymorphisms would be associated with CsA failure. PATIENTS AND METHODS: The distribution of the different genotypes of single nucleotide polymorphisms (SNP) G2677T/A and C3435T of MDR1 exons 21 and 26, respectively, was studied in 154 patients (mean age, 44 yr) who had received CsA to treat severe attacks of steroid resistant UC in 11 centers in France and Belgium. Patients were classified as CsA failure (n = 50) when they needed colectomy within 30 days after CsA initiation. The SNPs were detected by use of a 5' nuclease allelic discrimination assay. RESULTS: There was a significant association between the G2677T/A polymorphism distribution (exon 21) and the risk for CsA failure (P = 0.0001). The TT genotype of exon 21 was significantly associated with the risk compared with the two other genotypes (odds ratio, 3.77; 95% confidence interval, 1.42-9.97, P = 0.007). There was no significant association between the genotype C3435T distribution (exon 26) and the risk of CsA failure (P = 0.23). CONCLUSION: The TT genotype of exon 21 MDR1 polymorphisms is associated with a higher risk of CsA failure in patients with steroid resistant UC. Further studies should be performed to establish whether other treatments could be more efficient to avoid surgery in this subset of patients.