Comparison between single and serial computed tomography images in classification of acute appendicitis, acute right-sided diverticulitis, and normal appendix using EfficientNet

• DOI: 10.1371/journal.pone.0281498
• Publication Date: 2023-05-24T17:29:44Z
• Authors: So Hyun Park; Young Jae Kim; Kwang Gi Kim; Jun-Won Chung; Hyun Cheol Kim; In Young Choi; Myung-Won You; Gi Pyo Lee; Jung Han Hwang
• Keywords: Diverticulitis

This study aimed to develop a convolutional neural network (CNN) using the EfficientNet algorithm for automated classification of acute appendicitis, diverticulitis, and normal appendix evaluate its diagnostic performance. We retrospectively enrolled 715 patients who underwent contrast-enhanced abdominopelvic computed tomography (CT). Of these, 246 had 254 215 appendix. Training, validation, test data were obtained from 4,078 CT images (1,959 823 1,296 cases) both single serial (RGB [red, green, blue]) image methods. augmented training dataset avoid disturbances caused by unbalanced datasets. For appendix, RGB method showed slightly higher sensitivity (89.66 vs. 87.89%; p = 0.244), accuracy (93.62% 92.35%), specificity (95.47% 94.43%) than did method. also yielded (83.35 80.44%; 0.019), (93.48% 92.15%), (96.04% 95.12%) Moreover, mean areas under receiver operating characteristic curve (AUCs) significantly appendicitis (0.951 0.937; < 0.0001), diverticulitis (0.972 0.963; 0.0025), (0.979 0.972; 0.0101) with those each condition. Thus, could be accurately distinguished on our model, particularly when