Reduced detection rate of artificial intelligence in images obtained from untrained endoscope models and improvement using domain adaptation algorithm

Abstract

A training dataset that is limited to a specific endoscope model can overfit artificial intelligence (AI) to its unique image characteristics. The performance of the AI may degrade in images of different endoscope model. The domain adaptation algorithm, i.e., the cycle-consistent adversarial network (cycleGAN), can transform the image characteristics into AI-friendly styles. We attempted to confirm the performance degradation of AIs in images of various endoscope models and aimed to improve them using cycleGAN transformation. Two AI models were developed from data of esophagogastroduodenoscopies collected retrospectively over 5 years: one for identifying the endoscope models, Olympus CV-260SL, CV-290 (Olympus, Tokyo, Japan), and PENTAX EPK-i (PENTAX Medical, Tokyo, Japan), and the other for recognizing the esophagogastric junction (EGJ). The AIs were trained using 45,683 standardized images from 1,498 cases and validated on 624 separate cases. Between the two endoscope manufacturers, there was a difference in image characteristics that could be distinguished without error by AI. The accuracy of the AI in recognizing gastroesophageal junction was >0.979 in the same endoscope-examined validation dataset as the training dataset. However, they deteriorated in datasets from different endoscopes. Cycle-consistent adversarial network can successfully convert image characteristics to ameliorate the AI performance. The improvements were statistically significant and greater in datasets from different endoscope manufacturers [original → AI-trained style, increased area under the receiver operating characteristic (ROC) curve, P-value: CV-260SL → CV-290, 0.0056, P = 0.0106; CV-260SL → EPK-i, 0.0182, P = 0.0158; CV-290 → CV-260SL, 0.0134, P < 0.0001; CV-290 → EPK-i, 0.0299, P = 0.0001; EPK-i → CV-260SL, 0.0215, P = 0.0024; and EPK-i → CV-290, 0.0616, P < 0.0001]. In conclusion, cycleGAN can transform the diverse image characteristics of endoscope models into an AI-trained style to improve the detection performance of AI.

A novel artificial intelligence-based endoscopic ultrasonography diagnostic system for diagnosing the invasion depth of early gastric cancer

“…CycleGAN, a variant of GAN, enables the conversion of images from one style to another without matched pairs using a parameter of cycle consistency, which checks that a converted image can be converted back to its original form [ 29 ]. In recent years, GAN-based techniques, including CycleGAN, have been reported as a way to overcome this domain shift problem [ 34 ], including in gastrointestinal endoscopy [ 35 ]. In this study, the AI model developed at Osaka University did not exhibit sufficient performance in the external validation cohort in its original state.…”

Artificial Intelligence: A Reliable Tool to Detect the Elongation of the Styloid Process

“…Such segregated images were further processed by image preparation, image embedding, and a ML approach. For size standardization, the images were prepared using a cropping tool (snipping tool in Microsoft Windows 11 OS software (Microsoft Corporation, Redmond, Washington, United States)) [ 11 ]. A single image with bilateral elongation of the styloid process was considered two images after cropping.…”