Andreas Vecsei

We report the case of an infant in whom congenital syphilis was diagnosed at the age of 5 weeks. The case is remarkable because of (a) the negative venereal disease laboratory test from the cord blood, (b) the incidental diagnosis of the disease in the fifth week of life, (c) pneumonia alba being one of the symptoms, (d) the occurrence of a mild Jarisch-Herxheimer reaction after initiation of penicillin therapy and (e) the successful treatment of infection related anaemia with recombinant human erythropoietin.

This Guideline refers to infants, children and adolescents aged 0-18 years. The areas covered include: indications for diagnostic and therapeutic esophagogastroduodenoscopy and ileo-colonoscopy; endoscopy for foreign body ingestion; corrosive ingestion and stricture/stenosis endoscopic management; upper and lower gastrointestinal bleeding; endoscopic retrograde cholangio-pancreatography and endoscopic ultrasonography. Percutaneous endoscopic gastrostomy and endoscopy specific to inflammatory bowel disease (IBD) has been dealt with in other Guidelines [1-3] and are therefore not mentioned in this Guideline. Training and ongoing skill maintenance are to be dealt with in an imminent sister publication to this.

ABSTRACT We evaluate different cross-validation (CV) protocols for an automated classification of colonic polyps. For this purpose we select six previously developed methods which achieved promising results already in the past. We then evaluate the methods using the cross-validation protocols leave-one-image-out (LOO-CV), leave-one-parent-image-out (LOPIO-CV), leave-one-lesion-out (LOLO-CV), and leave-one-patient-out (LOPO-CV). We show that, in general, the more restrictive cross-validation protocols lead to high results drops. While in case of LOO-CV the accuracies are rather high across all methods evaluated, the picture changes the more strictness a cross-validation mode imposes on the set of training images.

SPECIFIC AUTHOR CONTRIBUTIONS: Andreas Vécsei, MD, and Ulrike Graf wrote the manuscript. All authors contributed to study design, data collection and analysis, and approved the final draft for submission. The corresponding author declares that the manuscript is submitted on behalf of all authors. The best mode of follow-up in celiac disease has not yet been established. The intention of this study was to clarify which noninvasive follow-up investigation - serological tests or intestinal permeability test (IPT) - correlates best with histology and whether the interval between diagnosis and follow-up affects the accuracy of these tests. Data from adult patients with celiac disease (diagnosed between December 1989 and July 2006) followed up with biopsy, IPT, and serological tests [IgG anti-gliadin antibodies (AGA-IgG), AGA-IgA, and endomysial antibodies (EMA)] were retrieved from a computerized database. Results of noninvasive tests were compared with the persistence of villous atrophy on biopsy. Patients were divided into groups A, which comprised patients followed up within 2 years after diagnosis, and B, comprising patients followed up later than 2 years. Forty-seven patients were evaluable. The lactulose/mannitol (L/M) ratio had a sensitivity of 85 % and a specificity of 46.2 % for mucosal atrophy, whereas saccharose excretion showed a sensitivity of 60 % and a specificity of 52.6 %. The sensitivities of AGA-IgA and AGA-IgG were 15 % and 20 %, respectively, while specificity was 100 % for both. Validity of AGA was limited due to low number of positive results. EMA assay was 50 % sensitive and 77.8 % specific. In group A (n = 23) L/M ratio performed best in terms of sensitivity (88.9 %), whereas EMA achieved a higher specificity (71.4 %). In group B, the sensitivity of the L/M ratio decreased to 85.7 %, while the specificity of EMA increased to 91.7 %. In this study, none of the noninvasive tests was an accurate substitute for follow-up biopsy in detecting severe mucosal damage.

Distortion correction is applied to endoscopic duodenal imagery to improve automated classification of celiac disease affected mucosa patches. In a set of six edge- and shape-related feature extraction techniques, only a single one is able to consistently benefit from distortion correction, while for others, even a decrease of classification accuracy is observed. Different types of distortion correction do not lead to significantly different behaviour in the observed application scenario.