**3.2 AI in AAG diagnostics**

The atrophic gastritis can benefit from the applications of AI in the diagnosis as well. It is often hard to distinguish between the different types of gastritis. One of the most promising applications is the recent report by Franklin et al. that utilized a CNN machine learning model that can distinguish between cases of HPG and autoimmune gastritis with accuracy equal to GI pathologists [80]. This could be beneficial particularly in AAG since it is hard to diagnose pathologically depending on the expertise of the clinician.

*Autoimmune Diseases of the GI Tract Part II: Emergence of Diagnostic Tools and Treatments DOI: http://dx.doi.org/10.5772/intechopen.106185*

## **3.3 AI in celiac disease diagnostics**

Diagnosis of celiac disease (CD) is difficult because its symptoms are shared with many other diseases. However, AI can be used to further facilitate the diagnosis of CD. Joceli et al. proposed a web-based Clinical Decision-Support System (CDSS) using ML algorithms to identify CD. The database used for testing and training the algorithms consisted of clinical data of patients with 35 attributes of CD-related symptoms recorded per case. For the training set, a total of 178 cases were recorded out of which 46% were diagnosed with CD. For the testing set, a total of 38 cases were recorded out of which 37% were CD positive. The study used different variations of 13 algorithms equating the total number of models to 270. The algorithms were trained on the training set, and the best variation of each algorithm was used on the testing set. The selection criteria were the area under the curve of the receiver operating curve (AUC ROC). The results were compared with clinical diagnosis and the golden standard, and the results showed that the best algorithm was able to diagnose the CD cases with great accuracy. This preliminary work shows the prospective of using AI can be used to aid physicians in their diagnosis of diseases like CD [81].
