Su1816 - Agreement of CT Imaging Features of Crohn's Disease Between Radiologists and Automated Machine Learning Image Analysis

Abstract 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.

Download Full-text

Assessing Small Bowel Stricturing and Morphology in Crohn’s Disease Using Semi-automated Image Analysis

Inflammatory Bowel Diseases ◽

10.1093/ibd/izz196 ◽

2019 ◽

Vol 26 (5) ◽

pp. 734-742 ◽

Cited By ~ 3

Author(s):

Ryan W Stidham ◽

Binu Enchakalody ◽

Akbar K Waljee ◽

Peter D R Higgins ◽

Stewart C Wang ◽

...

Keyword(s):

Image Analysis ◽

Crohn’S Disease ◽

Small Bowel ◽

Crohn's Disease ◽

Structural Damage ◽

Clinical Decision Making ◽

Tertiary Care ◽

Automated Image Analysis ◽

Image Analysis System ◽

Ct Enterography

Abstract Background Evaluating structural damage using imaging is essential for the evaluation of small intestinal Crohn’s disease (CD), but it is limited by potential interobserver variation. We compared the agreement of enterography-based bowel damage measurements collected by experienced radiologists and a semi-automated image analysis system. Methods Patients with small bowel CD undergoing a CT-enterography (CTE) between 2011 and 2017 in a tertiary care setting were retrospectively reviewed. CT-enterography studies were reviewed by 2 experienced radiologists and separately underwent automated computer image analysis using bowel measurement software. Measurements included maximum bowel wall thickness (BWT-max), maximum bowel dilation (DIL-max), minimum lumen diameter (LUM-min), and the presence of a stricture. Measurement correlation coefficients and paired t tests were used to compare individual operator measurements. Multivariate regression was used to model identification of strictures using semi-automated measures. Results In 138 studies, the correlation between radiologists and semi-automated measures were similar for BWT-max (r = 0.724, 0.702), DIL-max (r = 0.812, 0.748), and LUM-min (r = 0.428, 0.381), respectively. Mean absolute measurement difference between semi-automated and radiologist measures were no different from the mean difference between paired radiologists for BWT-max (1.26 mm vs 1.12 mm, P = 0.857), DIL-max (2.78 mm vs 2.67 mm, P = 0.557), and LUM-min (0.54 mm vs 0.41 mm, P = 0.596). Finally, models of radiologist-defined intestinal strictures using automatically acquired measurements had an accuracy of 87.6%. Conclusion Structural bowel damage measurements collected by semi-automated approaches are comparable to those of experienced radiologists. Radiomic measures of CD will become an important new data source powering clinical decision-making, patient-phenotyping, and assisting radiologists in reporting objective measures of disease status.

Download Full-text

Radiologic (radiographic) and imaging features of ulcerative colitis and Crohn’s disease

Inflammatory Bowel Disease: From Bench to Bedside ◽

10.1007/978-1-4020-3477-0_19 ◽

2003 ◽

pp. 371-407

Author(s):

Edward Fitzgerald

Keyword(s):

Ulcerative Colitis ◽

Crohn’S Disease ◽

Crohn's Disease ◽

Imaging Features

Download Full-text

764 MACHINE LEARNING IDENTIFIES NOVEL BLOOD PROTEOMIC PREDICTORS OF PENETRATING AND STRICTURING COMPLICATIONS IN NEWLY DIAGNOSED PEDIATRIC CROHN'S DISEASE

Gastroenterology ◽

10.1016/s0016-5085(20)31082-9 ◽

2020 ◽

Vol 158 (6) ◽

pp. S-158-S-159

Author(s):

Ryan C. Ungaro ◽

Liangyuan Hu ◽

Jiayi Ji ◽

Subra Kugathasan ◽

Marla Dubinsky ◽

...

Keyword(s):

Machine Learning ◽

Crohn’S Disease ◽

Crohn's Disease ◽

Newly Diagnosed

Download Full-text

Machine learning methods to predict presence of intestine damage in patients with Crohn’s disease

Medical Imaging 2020: Computer-Aided Diagnosis ◽

10.1117/12.2549326 ◽

2020 ◽

Author(s):

Binu E. Enchakalody ◽

Brianna Henderson ◽

Stewart Wang ◽

Grace L. Su ◽

Ashish Wasnik ◽

...

Keyword(s):

Machine Learning ◽

Crohn’S Disease ◽

Crohn's Disease ◽

Learning Methods ◽

Machine Learning Methods

Download Full-text

Mo1172 A RADIOMICS MACHINE LEARNING PREDICTOR CAN REPRODUCIBLY DIAGNOSE ACTIVE TERMINAL ILEAL CROHN'S DISEASE PATIENTS ON CT ENTEROGRAPHY SCANS ACROSS VARIATIONS IN DOSE STRENGTHS

Gastroenterology ◽

10.1016/s0016-5085(20)32713-x ◽

2020 ◽

Vol 158 (6) ◽

pp. S-812-S-813

Author(s):

Prathyush Chirra ◽

Alain G. Rizk ◽

Avani Muchhala ◽

Kaustav Bera ◽

Namita S. Gandhi ◽

...

Keyword(s):

Machine Learning ◽

Crohn’S Disease ◽

Crohn's Disease ◽

Ct Enterography

Download Full-text

Clinical-stage Approaches for Imaging Chronic Inflammation and Fibrosis in Crohn’s Disease

Inflammatory Bowel Diseases ◽

10.1093/ibd/izaa218 ◽

2020 ◽

Vol 26 (10) ◽

pp. 1509-1523

Author(s):

Mary-Louise C Greer ◽

Ruth Cytter-Kuint ◽

Li-tal Pratt ◽

Don Soboleski ◽

Gili Focht ◽

...

Keyword(s):

Crohn’S Disease ◽

Magnetic Resonance ◽

Crohn's Disease ◽

Treatment Response ◽

Clinical Stage ◽

Imaging Biomarkers ◽

Mr Enterography ◽

Imaging Features ◽

Anatomic Site ◽

Pelvic Mri

Abstract The number of imaging-based indices developed for inflammatory bowel disease as research tools, objectively measuring ileocolonic and perianal activity and treatment response, has expanded in the past 2 decades. Created primarily to assess Crohn’s disease (CD), there is increasing adoption of these indices into the clinical realm to guide patient care. This translation has been facilitated by validation in adult and pediatric populations, prompted by simplification of score calculations needed for practical application outside the research environment. The majority of these indices utilize magnetic resonance imaging (MRI), specifically MR enterography (MRE) and pelvic MRI, and more recently ultrasound. This review explores validated indices by modality, anatomic site and indication, including for documentation of the presence and extent of CD, disease progression, complications, and treatment response, highlighting those in clinical use or with the potential to be. As well, it details index imaging features used to quantify chronic inflammatory activity, severity, and to lesser extent fibrosis, in addition to their reference standards and any modifications. Validation in the pediatric population of indices primarily developed in adult cohorts such as the Magnetic Resonance Index of Activity (MaRIA), the Simplified Magnetic Resonance Index of Activity (MARIAs), and the MRE global score (MEGS), together with newly developed pediatric-specific indices, are discussed. Indices that may be predictive of disease course and investigational techniques with the potential to provide future imaging biomarkers, such as multiparametric MRI, are also briefly considered.

Download Full-text