A study established a deep convolutional neural network (DCNN) system that significantly improved diagnostic accuracy for early gastric cancer (EGC). In independent tests, the model achieved an area under the curve (AUC) of 0.917 and sensitivity of 93.38% for image datasets, while video tests showed an AUC of 0.930 and 96.92% sensitivity. Novice endoscopists reached near-expert accuracy levels with DCNN support, and overall diagnostic times decreased, demonstrating the system’s potential in clinical settings.
Journal Article by Feng J, Zhang Y (…) Huang X et 6 al. in Surg Endosc
© 2025. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.