Artificial-intelligence-based decision support tools for the differential diagnosis of colitis

Artificial-intelligence-based decision support tools for the differential diagnosis of colitis. Guimarães, P., Finkler, H., Reichert, M., Zimmer, V., Gruenhage, F., Krawczyk, M., Lammert, F., Keller, A., & Casper, M. European journal of clinical investigation, 53:e13960, 01, 2023.

Paper doi abstract bibtex

Background: Whereas Artificial Intelligence (AI) based tools have recently been introduced in the field of gastroenterology, application in inflammatory bowel disease (IBD) is in its infancies. We established AI-based algorithms to distinguish IBD from infectious and ischemic colitis using endoscopic images and clinical data. Methods: Firstly, we trained and tested a Convolutional Neural Network (CNN) using 1,796 real-world images from 494 patients, presenting with three diseases (IBD [n=212], ischemic colitis [n=157], infectious colitis [n=125]). Moreover, we evaluated a Gradient Boosted Decision Trees (GBDT) algorithm using five clinical parameters as well as a hybrid approach (CNN+GBDT). Patients and images were randomly split into two completely independent datasets. The proposed approaches were benchmarked against each other and three expert endoscopists on the test set. Results: For the image-based CNN, the GBDT algorithm and the hybrid approach global accuracies were 0.709, 0.792, and 0.766, respectively. Positive predictive values were 0.602, 0.702, and 0.657. Global areas under the receiver operating characteristics (ROC) and precision recall (PR) curves were 0.727/0.585, 0.888/0.823, and 0.838/0.733, respectively. Global accuracy did not differ between CNN and endoscopists (0.721), but the clinical parameter-based GBDT algorithm outperformed CNN and expert image classification. Conclusions: Decision support systems exclusively based on endoscopic image analysis for the differential diagnosis of colitis, representing a complex clinical challenge, seem not yet to be ready for primetime and more diverse image datasets may be necessary to improve performance in future development. The clinical value of the proposed clinical parameters algorithm should be evaluated in prospective cohorts.

@article{gui123,
	author = {Guimarães, Pedro and Finkler, Helen and Reichert, Matthias and Zimmer, Vincent and Gruenhage, Frank and Krawczyk, Marcin and Lammert, Frank and Keller, Andreas and Casper, Mary},
	year = {2023},
	month = {01},
	abstract = "{Background: Whereas Artificial Intelligence (AI) based tools have recently been introduced in the field of gastroenterology, application in inflammatory bowel disease (IBD) is in its infancies. We established AI-based algorithms to distinguish IBD from infectious and ischemic colitis using endoscopic images and clinical data. Methods: Firstly, we trained and tested a Convolutional Neural Network (CNN) using 1,796 real-world images from 494 patients, presenting with three diseases (IBD [n=212], ischemic colitis [n=157], infectious colitis [n=125]). Moreover, we evaluated a Gradient Boosted Decision Trees (GBDT) algorithm using five clinical parameters as well as a hybrid approach (CNN+GBDT). Patients and images were randomly split into two completely independent datasets. The proposed approaches were benchmarked against each other and three expert endoscopists on the test set. Results: For the image-based CNN, the GBDT algorithm and the hybrid approach global accuracies were 0.709, 0.792, and 0.766, respectively. Positive predictive values were 0.602, 0.702, and 0.657. Global areas under the receiver operating characteristics (ROC) and precision recall (PR) curves were 0.727/0.585, 0.888/0.823, and 0.838/0.733, respectively. Global accuracy did not differ between CNN and endoscopists (0.721), but the clinical parameter-based GBDT algorithm outperformed CNN and expert image classification. Conclusions: Decision support systems exclusively based on endoscopic image analysis for the differential diagnosis of colitis, representing a complex clinical challenge, seem not yet to be ready for primetime and more diverse image datasets may be necessary to improve performance in future development. The clinical value of the proposed clinical parameters algorithm should be evaluated in prospective cohorts.}",
	pages = {e13960},
	title = {Artificial-intelligence-based decision support tools for the differential diagnosis of colitis},
	volume = {53},
	journal = {European journal of clinical investigation},
    	url = {https://doi.org/10.1111/eci.13960},
	doi = {10.1111/eci.13960}
}

Downloads: 0

{"_id":"m346F6TF8rddHcP9s","bibbaseid":"guimares-finkler-reichert-zimmer-gruenhage-krawczyk-lammert-keller-etal-artificialintelligencebaseddecisionsupporttoolsforthedifferentialdiagnosisofcolitis-2023","author_short":["Guimarães, P.","Finkler, H.","Reichert, M.","Zimmer, V.","Gruenhage, F.","Krawczyk, M.","Lammert, F.","Keller, A.","Casper, M."],"bibdata":{"bibtype":"article","type":"article","author":[{"propositions":[],"lastnames":["Guimarães"],"firstnames":["Pedro"],"suffixes":[]},{"propositions":[],"lastnames":["Finkler"],"firstnames":["Helen"],"suffixes":[]},{"propositions":[],"lastnames":["Reichert"],"firstnames":["Matthias"],"suffixes":[]},{"propositions":[],"lastnames":["Zimmer"],"firstnames":["Vincent"],"suffixes":[]},{"propositions":[],"lastnames":["Gruenhage"],"firstnames":["Frank"],"suffixes":[]},{"propositions":[],"lastnames":["Krawczyk"],"firstnames":["Marcin"],"suffixes":[]},{"propositions":[],"lastnames":["Lammert"],"firstnames":["Frank"],"suffixes":[]},{"propositions":[],"lastnames":["Keller"],"firstnames":["Andreas"],"suffixes":[]},{"propositions":[],"lastnames":["Casper"],"firstnames":["Mary"],"suffixes":[]}],"year":"2023","month":"01","abstract":"Background: Whereas Artificial Intelligence (AI) based tools have recently been introduced in the field of gastroenterology, application in inflammatory bowel disease (IBD) is in its infancies. We established AI-based algorithms to distinguish IBD from infectious and ischemic colitis using endoscopic images and clinical data. Methods: Firstly, we trained and tested a Convolutional Neural Network (CNN) using 1,796 real-world images from 494 patients, presenting with three diseases (IBD [n=212], ischemic colitis [n=157], infectious colitis [n=125]). Moreover, we evaluated a Gradient Boosted Decision Trees (GBDT) algorithm using five clinical parameters as well as a hybrid approach (CNN+GBDT). Patients and images were randomly split into two completely independent datasets. The proposed approaches were benchmarked against each other and three expert endoscopists on the test set. Results: For the image-based CNN, the GBDT algorithm and the hybrid approach global accuracies were 0.709, 0.792, and 0.766, respectively. Positive predictive values were 0.602, 0.702, and 0.657. Global areas under the receiver operating characteristics (ROC) and precision recall (PR) curves were 0.727/0.585, 0.888/0.823, and 0.838/0.733, respectively. Global accuracy did not differ between CNN and endoscopists (0.721), but the clinical parameter-based GBDT algorithm outperformed CNN and expert image classification. Conclusions: Decision support systems exclusively based on endoscopic image analysis for the differential diagnosis of colitis, representing a complex clinical challenge, seem not yet to be ready for primetime and more diverse image datasets may be necessary to improve performance in future development. The clinical value of the proposed clinical parameters algorithm should be evaluated in prospective cohorts.","pages":"e13960","title":"Artificial-intelligence-based decision support tools for the differential diagnosis of colitis","volume":"53","journal":"European journal of clinical investigation","url":"https://doi.org/10.1111/eci.13960","doi":"10.1111/eci.13960","bibtex":"@article{gui123,\n\tauthor = {Guimarães, Pedro and Finkler, Helen and Reichert, Matthias and Zimmer, Vincent and Gruenhage, Frank and Krawczyk, Marcin and Lammert, Frank and Keller, Andreas and Casper, Mary},\n\tyear = {2023},\n\tmonth = {01},\n\tabstract = \"{Background: Whereas Artificial Intelligence (AI) based tools have recently been introduced in the field of gastroenterology, application in inflammatory bowel disease (IBD) is in its infancies. We established AI-based algorithms to distinguish IBD from infectious and ischemic colitis using endoscopic images and clinical data. Methods: Firstly, we trained and tested a Convolutional Neural Network (CNN) using 1,796 real-world images from 494 patients, presenting with three diseases (IBD [n=212], ischemic colitis [n=157], infectious colitis [n=125]). Moreover, we evaluated a Gradient Boosted Decision Trees (GBDT) algorithm using five clinical parameters as well as a hybrid approach (CNN+GBDT). Patients and images were randomly split into two completely independent datasets. The proposed approaches were benchmarked against each other and three expert endoscopists on the test set. Results: For the image-based CNN, the GBDT algorithm and the hybrid approach global accuracies were 0.709, 0.792, and 0.766, respectively. Positive predictive values were 0.602, 0.702, and 0.657. Global areas under the receiver operating characteristics (ROC) and precision recall (PR) curves were 0.727/0.585, 0.888/0.823, and 0.838/0.733, respectively. Global accuracy did not differ between CNN and endoscopists (0.721), but the clinical parameter-based GBDT algorithm outperformed CNN and expert image classification. Conclusions: Decision support systems exclusively based on endoscopic image analysis for the differential diagnosis of colitis, representing a complex clinical challenge, seem not yet to be ready for primetime and more diverse image datasets may be necessary to improve performance in future development. The clinical value of the proposed clinical parameters algorithm should be evaluated in prospective cohorts.}\",\n\tpages = {e13960},\n\ttitle = {Artificial-intelligence-based decision support tools for the differential diagnosis of colitis},\n\tvolume = {53},\n\tjournal = {European journal of clinical investigation},\n \turl = {https://doi.org/10.1111/eci.13960},\n\tdoi = {10.1111/eci.13960}\n}\n","author_short":["Guimarães, P.","Finkler, H.","Reichert, M.","Zimmer, V.","Gruenhage, F.","Krawczyk, M.","Lammert, F.","Keller, A.","Casper, M."],"key":"gui123","id":"gui123","bibbaseid":"guimares-finkler-reichert-zimmer-gruenhage-krawczyk-lammert-keller-etal-artificialintelligencebaseddecisionsupporttoolsforthedifferentialdiagnosisofcolitis-2023","role":"author","urls":{"Paper":"https://doi.org/10.1111/eci.13960"},"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://www.ccb.uni-saarland.de/wp-content/uploads/2024/01/references.bib_.txt","dataSources":["Pny5E4E9kc7C8gG8g","SiGP46KPWizw6ihLJ"],"keywords":[],"search_terms":["artificial","intelligence","based","decision","support","tools","differential","diagnosis","colitis","guimarães","finkler","reichert","zimmer","gruenhage","krawczyk","lammert","keller","casper"],"title":"Artificial-intelligence-based decision support tools for the differential diagnosis of colitis","year":2023}