CT联合深度学习技术在腹部脂肪定量中的 研究进展

Research Advances in CT Combined with Deep Learning for Abdominal Fat Quantification

摘要: 随着代谢性疾病负担日益加重以及对肿瘤风险预测指标的重视，腹部脂肪尤其是内脏脂肪组织和皮下脂肪组织的精准量化已成为疾病风险预测与个体化干预的关键指标。传统CT测量方法受限于主观性强、效率低下和可重复性差等问题，难以满足临床对高通量和自动化脂肪参数提取的需求。近年来，深度学习模型与CT的结合，实现了CT图像中腹部脂肪组织的高精度分割与测量，并将腹部脂肪作为代谢性疾病和肿瘤的独立影像生物标志物。本文就CT联合深度学习技术在腹部脂肪测量中的进展进行综述。

Abstract: With the increasing burden of metabolic diseases and the growing emphasis on tumor risk prediction indicators, the accurate quantification of abdominal fat, especially visceral adipose tissue and subcutaneous adipose tissue, has become a key indicator for disease risk prediction and individualized intervention. Traditional CT measurement methods are limited by strong subjectivity and low efficiency and poor reproducibility, making it difficult to meet clinical demands for high-throughput, automated extraction of fat parameters. In recent years, the combination of deep learning models and CT has achieved high-precision segmentation and measurement of abdominal adipose tissue in CT images, and has established abdominal fat as an independent imaging biomarker for metabolic diseases and tumors. This article reviews the research progress of CT combined with deep learning technology in abdominal fat measurement.