Impact of Annotation Level on Multisequence MRI Models for Preoperative Microvascular Invasion Prediction in Hepatocellular Carcinoma

作者：Yifan Pan, Rongping Ye, Jiayi Li, Yamei Liu, Zhaodi Huang, Qiuyuan Yue, Lanmei Gao, Chuan Yan, Yueming Li · 发表于：Radiology Imaging Cancer · 年份：2026 · DOI：10.1148/rycan.250407 · 被引用次数：1 · 研究领域：Hepatocellular Carcinoma Treatment and Prognosis、MRI in cancer diagnosis、Radiomics and Machine Learning in Medical Imaging

Bounding box annotation provides a time-efficient, clinically viable alternative to voxel-level masks for preoperative microvascular invasion prediction in hepatocellular carcinoma, offering statistically comparable performance while improving annotation efficiency.