Towards improved celiac disease diagnosis
Towards improved celiac disease diagnosis
Disciplines
Other Human Medicine, Health Sciences (40%); Computer Sciences (10%); Clinical Medicine (50%)
Keywords
-
Celiac disease,
Pattern recognition,
Endoscopy,
Villous atrophy,
Automated imag analysis,
Children
Celiac disease (CD) has a prevalence of 1%. To confirm the diagnosis, children with serologically suspected CD undergo small bowel biopsy. To avoid biopsies but still reliably prove the diagnosis, a variety of new endoscopic techniques including software for the analysis of endoscopic still images has been developed. If automated analysis-software can detect villous atrophy not only in endoscopic still images but in video clips as well, in such cases, a small bowel biopsy could entirely be avoided. The main question we will try to assess is, if such software can be developed and easily integrated into the routine of pediatric endoscopy. Furthermore, we will examine if the narrow band imaging technique, an optical/digital chromoendoscopy, can further improve the reliability of automated analysis. Various feature extraction and classification strategies will be applied for automated differentiation between presence and absence of villous atrophy in the endoscopic video clips. 1
In pediatric celiac disease, having a prevalence of 1%, a biopsy of the small bowel is recommended in the case of unclear serology to clarify diagnosis. The reliability of the corresponding histology findings is problematic due to two reasons: (1) Celiac disease exhibits a so-called patchy distribution in some cases, which means that not the entire mucosa is affected. Misdiagnosis may happen in case the celiac disease affected mucosa areas are missed and biopsies have been taken from healthy areas. (2) Pathological diagnosis of biopsy specimens suffers from high inter- and intra-observer variability. Thus, there is need for observer-independent diagnostic methods for celiac disease. In this project the computer-based analysis of endoscopy still images allowed to reliably detect villous atrophy and as such confirm the diagnosis of celiac disease. Furthermore, an computer-based analysis was also successfully developed for and used on endoscopic video, thus enhancing the accuracy of biopsy targeting on the one hand, on the other hand abandoning the necessity of conducting small bowel biopsy at all when villous atrophy was detected. This new diagnostic approach increases the patients` safety during endoscopy, but also significantly contributes to cost reduction as the entire biopsy diagnostic work-up can be avoided.
- Andreas Uhl, Universität Salzburg , associated research partner
Research Output
- 208 Citations
- 17 Publications
-
2018
Title Fisher encoding of convolutional neural network features for endoscopic image classification DOI 10.1117/1.jmi.5.3.034504 Type Journal Article Author Wimmer G Journal Journal of Medical Imaging Pages 034504-034504 Link Publication -
2018
Title Automated classification of celiac disease during upper endoscopy: Status quo and quo vadis DOI 10.1016/j.compbiomed.2018.04.020 Type Journal Article Author Gadermayr M Journal Computers in Biology and Medicine Pages 221-226 Link Publication -
2016
Title Computer-aided texture analysis combined with experts' knowledge: Improving endoscopic celiac disease diagnosis DOI 10.3748/wjg.v22.i31.7124 Type Journal Article Author Gadermayr M Journal World Journal of Gastroenterology Pages 7124-7134 Link Publication -
2016
Title Texture Description Using Dual Tree Complex Wavelet Packets DOI 10.1007/978-3-319-48890-5_18 Type Book Chapter Author Liedlgruber M Publisher Springer Nature Pages 181-190 -
2016
Title Fully automated decision support systems for celiac disease diagnosis DOI 10.1016/j.irbm.2015.09.009 Type Journal Article Author Gadermayr M Journal IRBM Pages 31-39 -
2016
Title Narrow Band Imaging Versus White-Light: What is Best for Computer-Assisted Diagnosis of Celiac Disease? DOI 10.1109/isbi.2016.7493282 Type Conference Proceeding Abstract Author Gadermayr M Pages 355-359 -
2016
Title Incorporating Human Knowledge in Automated Celiac Disease Diagnosis DOI 10.1109/ipta.2016.7821009 Type Conference Proceeding Abstract Author Gadermayr M Pages 1-6 -
2016
Title CNN Transfer Learning for the Automated Diagnosis of Celiac Disease DOI 10.1109/ipta.2016.7821020 Type Conference Proceeding Abstract Author Wimmer G Pages 1-6 -
2020
Title Improving CNN training on endoscopic image data by extracting additionally training data from endoscopic videos DOI 10.1016/j.compmedimag.2020.101798 Type Journal Article Author Wimmer G Journal Computerized Medical Imaging and Graphics Pages 101798 Link Publication -
2017
Title Evaluation of Domain Specific Data Augmentation Techniques for the Classification of Celiac Disease Using Endoscopic Imagery DOI 10.1109/mmsp.2017.8122221 Type Conference Proceeding Abstract Author Wimmer G Pages 1-6 -
2017
Title Degradation adaptive texture classification for real-world application scenarios DOI 10.1134/s1054661817010035 Type Journal Article Author Gadermayr M Journal Pattern Recognition and Image Analysis Pages 66-81 -
2017
Title Convolutional Neural Network Architectures for the Automated Diagnosis of Celiac Disease DOI 10.1007/978-3-319-54057-3_10 Type Book Chapter Author Wimmer G Publisher Springer Nature Pages 104-113 -
2016
Title Making texture descriptors invariant to blur DOI 10.1186/s13640-016-0116-7 Type Journal Article Author Gadermayr M Journal EURASIP Journal on Image and Video Processing Pages 14 Link Publication -
2015
Title Survey on computer aided decision support for diagnosis of celiac disease DOI 10.1016/j.compbiomed.2015.02.007 Type Journal Article Author Hegenbart S Journal Computers in Biology and Medicine Pages 348-358 Link Publication -
2015
Title Comparing Endoscopic Imaging Configurations in Computer-Aided Celiac Disease Diagnosis DOI 10.1109/ipta.2015.7367184 Type Conference Proceeding Abstract Author Gadermayr M Pages 446-451 -
2015
Title Dealing with Intra-Class and Intra-Image Variations in Automatic Celiac Disease Diagnosis DOI 10.1007/978-3-662-46224-9_79 Type Book Chapter Author Gadermayr M Publisher Springer Nature Pages 461-466 -
2015
Title Boosting Small-Data Performance of LBP: A Case Study in Celiac Disease Diagnosis DOI 10.1007/978-3-319-19665-7_19 Type Book Chapter Author Gadermayr M Publisher Springer Nature Pages 224-233