Search

Background Predicting Crohn’s disease (CD) phenotype development has proven challenging due to difficulties in biopsy image interpretation of histologically similar yet biologically distinct phenotypes. At initial diagnosis, mostly CD patients are classified as B1 (inflammatory behavior), they typically either retain B1 phenotype or develop more complicated B2 (stricturing), B3 (internal penetrating), or B2/B3 phenotypes (defined by Montreal Classification). Prediction of phenotype development based on baseline biopsies can radically improve our clinical care by altering disease management. Biopsy-based image analysis via Convolutional Neural Networks (CNNs) has been successful in cancer detection, but investigation into its utility for CD phenotypes is lacking. We applied a machine learning CNN model to classify CD phenotypes and histologically normal ileal controls. Methods Baseline hematoxylin & eosin (H&E) stained ileal biopsy slides were obtained from the Cincinnati Children’s Hospital Medical Center’s RISK validation sub cohort. At University of Virginia, biopsy slides were digitized, and a ResNet101 CNN model was trained. High resolution images were patched into 1000x1000 pixels with a 50% overlap and then resized to 256x256 pixels for training (80-20 split was kept between training and testing sets to ensure same patient patches were not mixed). Gradient Weighted Activating Mappings (GradCAMs) were used to visualize the model’s decision making process. Results We initially trained the model for CD vs. controls where it achieved 97% accuracy in detecting controls. We further trained it for classifying CD phenotypes (n=16 B1, n=16 B2, n=4 B3, n=13 B2/B3; phenotype decision at 5 year). It displayed a higher accuracy in detecting B2 (85%) while there were overlaps in the detection of other phenotypes (Figure 1). For B2, Grad-CAM heatmaps highlighted central pink areas within the lamina propria as the model’s regions of interests which were present when other phenotypes were misclassified as B2 (Figure 2). Conclusions: Here we highlight the potential utility of a machine learning image analysis model for describing CD phenotypes using H&E stained biopsies. Previous studies have shown B2 to be associated with increased activation for extracellular matrix genes (connective tissue component). Our GradCAM results support this finding as the pink central areas utilized by the model for classifying B2 could be connective tissue. Further confirmation via molecular phenotyping including Sirius Red immunohistochemistry is underway. Our work supports prediction of CD phenotypes using baseline biopsies at diagnosis and has potential to influence individualized care for children with CD.

Background: Adalimumab is effective for adults with moderate to severe Crohn's disease (CD). Data in children with inflammatory bowel disease (IBD) is sparse. We aim to evaluate pediatric IBD response to adalimumab. Methods: We conducted a retrospective chart review of pediatric IBD patients treated with adalimumab. We defined response as either a decrease in the disease activity index severity or remission after 3–6 months of treatment. Results: Ten patients were identified. Seven had CD, three had ulcerative colitis (UC). All had colonic disease and disease in either the esophagus, stomach, or duodenum. Eight patients initially responded to infliximab, 5 with CD and 3 with UC; 2 CD patients had incomplete data for evaluation. The mean and median times between initial infliximab and initial adalimumab were 29.5 and 24 months, respectively. Eight patients responded to adalimumab 40 mg or 80 mg biweekly, or 80 mg initial dose followed by 40 mg biweekly. The mean Pediatric Crohn's Disease Activity Index (PCDAI) in CD patients before and after adalimumab were 12 and 4.2, respectively (normal ≤10). The mean Lichtiger Colitis Activity Index (LCAI) in UC patients before and after adalimumab were 9 and 5.1, respectively (normal

Background: Tumor necrosis factor-α (TNF-α) may contribute to the alterations in protein and energy metabolism present in children with Crohn's disease (CD), who frequently suffer from growth disturbance. We hypothesized that anti-TNF-α therapy would reduce protein losses, due to decreased proteolysis and increased protein synthesis, and that anti-TNF-α therapy would decrease resting energy expenditure. Methods: Children with active CD underwent metabolic assessment immediately before and 2 weeks following initial infliximab infusion. Using the stable isotopes [d5] phenylalanine and [1–13C] leucine, 2 independent measures of protein metabolism were determined during fasting and in response to parenteral nutrition. Energy expenditure, determined by indirect calorimetry, was measured in fasting and parenterally fed states. Results: Fifteen children completed the study. Following infliximab therapy, significant reductions in proteolysis (P < 0.05) were noted in the fasting state (8%–11%) and during parenteral nutrition infusion (10%–12%). Phenylalanine utilization for protein synthesis decreased significantly (8%–13%) following infliximab (P < 0.05). Protein balance was not significantly altered. No significant changes in energy expenditure were observed following infliximab in fasting or parenterally fed states. Supplementation with parenteral nutrition resulted in significantly decreased proteolysis (8%–21%; P < 0.05), increased protein synthesis (37%–45%; P < 0.01), and improved protein balance (P < 0.01) compared to the fasting state. Conclusions: Following the initial infliximab infusion in children with CD, proteolysis and protein synthesis were significantly reduced in the fasting and parenterally fed states. Supplementation with parenteral nutrition resulted in significant improvements in protein metabolism compared to the fasting state both before and after infliximab therapy. (Inflamm Bowel Dis 2007)

Background: Our purpose was to study the concordance of serological tests for inflammatory bowel disease with clinical diagnosis established by traditional testing in children. Methods: All children seen in our division who had IBD Diagnostic System (ie, pANCA, ASCA IgA, and ASCA IgG) performed over a 21-month period (June 1998 to February 2000) were identified. Their medical records were reviewed for basic demographics, test results (endoscopy, histology, and radiology), IBD Diagnostic System results, and patient symptoms/medications. Results of the IBD Diagnostic System were compared with several patient characteristics including age, sex, absence/presence of symptoms, medication use, disease activity and duration. Results: One hundred seven patients were divided into 6 groups based on clinical diagnosis and IBD Diagnostic System results. The sensitivity, specificity and +/− predictive values of the IBD Diagnostic System for ulcerative colitis were 69.2, 95.1, 90.0, and 87.1%, respectively, and for Crohn's disease were 54.1, 96.8, 90.9, and 80.8%, respectively. Overall, the results of the IBD Diagnostic System were concordant with the clinical diagnosis in 76 of the 107 (71%) patients. Conclusions: In our experience, the specificity of IBD Diagnostic System is better than the sensitivity; the sensitivity is better for ulcerative colitis than Crohn's disease (69.2% vs 54.1%). The low sensitivity, especially for Crohn's disease, precludes the possibility that the IBD Diagnostic System can replace traditional studies when evaluating for inflammatory bowel disease. Though we do not exclude inflammatory bowel disease solely by IBD Diagnostic System results, it is reassuring to note that all patients without clinical evidence of inflammatory bowel disease also had negative IBD Diagnostic System results.