Establishment and Clinical Application of an Artificial Intelligence Diagnostic Platform for Identifying Rectal Cancer Tumor Budding

• DOI: 10.3389/fonc.2021.626626
• Publication Date: 2021-03-08T05:40:37Z
• Authors: Shanglong Liu; Yuejuan Zhang; Yiheng Ju; Ying Li; Xiaoning Kang; Xiaojuan Yang; Tianye Niu; Xiaoming Xing; Yun Lu
• Keywords: Tumor budding

Tumor budding is considered a sign of cancer cell activity and the first step tumor metastasis. This study aimed to establish an automatic diagnostic platform for rectal pathology by training Faster region-based convolutional neural network (F-R-CNN) on pathological images budding. Postoperative section 236 patients with from Affiliated Hospital Qingdao University, China, taken January 2015 2017 were used in analysis. The site was labeled Label image software. learning set trained using R-CNN test verify outcome. evaluated through receiver operating characteristic (ROC) curve. Through budding, preliminarily established. precision–recall curves generated precision recall nodule category set. area under curve = 0.7414, which indicated that effective. validation yielded ROC 0.88, indicating established artificial intelligence performed well at diagnosis deep can help pathologists make more efficient accurate diagnoses.