Minireviews
Copyright ©The Author(s) 2021.
Artif Intell Gastroenterol. Jun 28, 2021; 2(3): 85-93
Published online Jun 28, 2021. doi: 10.35712/aig.v2.i3.85
Table 4 Artificial intelligence implications in the treatment and prognosis of inflammatory bowel disease patients
Ref.
Purpose
AI/DL model
Design
Result
Waljee et al[28], 2018To predict corticosteroid-free biologic remissionRandom Forest modelingRetrospectiveAt week 52, patients predicted to fail succeeded 6.7% of the time
Waljee et al[29], 2019To predict long-term response to ustekinumabRandom Forest modelingRetrospectivePer week-8 model, only 11% predicted to fail achieved remission
Waljee et al[30], 2010To predict response to thiopurinesRandom Forest modelingRetrospectiveThe model was superior to metabolite measurement in predicting non-responders.
Waljee et al[31], 2018To externally validate previously developed thiopurine algorithmRandom Forest modelingRetrospectiveThe algorithm accurately predicted objective remission with AUROC 0.76
Waljee et al[32], 2017To identify patients in objective remission on thiopurines and analyze if these patients had fewer clinical events per yearRandom Forest modelingRetrospectiveAUROC for algorithm-predicted remission was 0.79 vs 0.49 for thiopurine metabolite proving model superiority