A Retrospective, Single Center, Double Arm, Blind, Controlled, Pivotal Trial to Evaluate Classification by the Depth of Invasion of Early Gastric Cancer Sensitivity and Specificity of Gastric Cancer Image Computer Aided Detection and Diagnosis Software 'WAYMED Endo' and Endoscopists in Endoscopic Images
A Retrospective, Single Center, Double Arm, Blind, Controlled, Pivotal Trial to Evaluate Classification by the Depth of Invasion of Early Gastric Cancer Sensitivity and Specificity of Gastric Cancer Image Computer Aided Detection and Diagnosis Software 'WAYMED Endo' and Endoscopists in Endoscopic Images
The purpose of this clinical trial is to prove that the prediction capability of 'WAYMED endo' is superior to that of the endoscopists in classifying EGC based on the depth of invasion categories in gastro-endoscopic images.
The computer-aided detection·diagnosis software is an Artificial Intelligence (AI) software used to assist medical specialists in diagnostic decisions by automatically classifying EGC based on the depth of invasion categories in gastro-endoscopic images and displaying the results and possibilities on the User Interface (UI).
This clinical trial aims to evaluate the sensitivity and specificity of 'WAYMED endo' compared to that of endoscopists in classifying EGC based on the depth of invasion categories in gastro-endoscopic images. It is designed as a retrospective, single-center, double-arm, double-blind (endoscopist, investigational medical device applicator), controlled, and pivotal trial. Medical data collected retrospectively from subjects who underwent Esophagogastroduodenoscopy (EGD) and biopsy are screened. As a result of screening, medical data that meet all inclusion/exclusion criteria are enrolled and assigned to the trial and control groups.
In the trial group, the investigational medical device is applied to the images, while the endoscopists interpret the images in the control group. The Reference Standard Establishment Committee records the reference standard results as either "Mucosa (mucosal invasion)" or "Submucosa (submucosal invasion)", based on the depth of invasion of the lesion, and marks the detected lesion area with an oval on the image. The reference standard results are blinded, so they cannot be disclosed to the endoscopists or the investigational medical device applicator.
The primary endpoint includes the sensitivity (%) and specificity (%) of "WAYMED endo" and the endoscopists in classifying EGC based on the depth of invasion categories ("Mucosa" or "Submucosa") as confirmed by the reference standard. The secondary endpoint includes the accuracy (%) of "WAYMED endo" and the endoscopists in accurately classifying all early gastric cancer images as either "Mucosa" or "Submucosa", based on the depth of invasion categories as confirmed through pathological examination.
Inclusion Criteria:
1. Patients aged 19 years or older who underwent EGD 2. Confirmed the presence of gastric cancer through the Electronic Medical Record (EMR), including reports of EGD or pathology
Exclusion Criteria: