A Lesion-Based Convolutional Neural Network Improves Endoscopic Detection and Depth Prediction of Early Gastric Cancer

Abstract

In early gastric cancer (EGC), tumor invasion depth is an important factor for determining the treatment method. However, as endoscopic ultrasonography has limitations when measuring the exact depth in a clinical setting as endoscopists often depend on gross findings and personal experience. The present study aimed to develop a model optimized for EGC detection and depth prediction, and we investigated factors affecting artificial intelligence (AI) diagnosis. We employed a visual geometry group(VGG)-16 model for the classification of endoscopic images as EGC (T1a or T1b) or non-EGC. To induce the model to activate EGC regions during training, we proposed a novel loss function that simultaneously measured classification and localization errors. We experimented with 11,539 endoscopic images (896 T1a-EGC, 809 T1b-EGC, and 9834 non-EGC). The areas under the curves of receiver operating characteristic curves for EGC detection and depth prediction were 0.981 and 0.851, respectively. Among the factors affecting AI prediction of tumor depth, only histologic differentiation was significantly associated, where undifferentiated-type histology exhibited a lower AI accuracy. Thus, the lesion-based model is an appropriate training method for AI in EGC. However, further improvements and validation are required, especially for undifferentiated-type histology.

Keywords: artificial intelligence; convolutional neural networks; early gastric cancer; endoscopy.

Figure 1

Receiver operating characteristic (ROC) curves of lesion-based visual geometry group (VGG)-16 for the test dataset with their areas under the curves (AUCs). (A) Early gastric cancer (EGC) detection model and (B) EGC depth prediction model.

Figure 2

Classification results of lesion-based VGG-16. The green lines indicate the actual early gastric cancer (EGC) regions. The blue lines indicate the activated regions at testing. The first two rows are images precisely classified to their own classes, whereas the last row shows misclassified images. (A) EGC detection. (B) EGC depth prediction.