浙江医学2025，Vol.47Issue(20)：2182-2186,5.DOI:10.12056/j.issn.1006-2785.2025.47.20.2025-1215
基于MRI的AI分割算法精准评估常染色体显性多囊肾病患者肾体积的研究
MRI-based AI segmentation for precisely assessing renal volumetry in autosomal dominant polycystic kidney disease
摘要
Abstract
Objective To develop a MRI-based artificial intelligence(AI)segmentation algorithm to try to achieve fully-automated accurate measurement of kidney volume in patients with autosomal dominant polycystic kidney disease(ADPKD)and to validate its clinical reliability.Methods Clinical data of 121 ADPKD patients at Hangzhou TCM Hospital Affiliated to Zhejiang Chinese Medical University and 29 at Zhejiang Xiaoshan Hospital were retrospectively collected.Based on axial T2 weighted imaging with fat suppression sequences,a polycystic kidney segmentation model(Swin-UNetMC)architecture based on Swin-UNet is constructed.Random stratified sampling is used for performance validation to achieve automated kidney segmentation and volume measurement.Its agreement with the conventional ellipsoid equation method was analyzed.Results The model achieved a Dice similarity coefficient of 0.982[intersection over union(IoU)=0.96]in the training set and 0.969(IoU=0.94)in the test set.The AI-based segmentation algorithm demonstrated a high correlation with the traditional method(r＞0.98),albeit with a systematic bias,resulting in a range of differences in total kidney volume of approximately-333.3 mL to 278.0 mL.Conclusion The MRI-based AI segmentation algorithm proposed in this study enables fully-automated and accurate kidney volume measurement in ADPKD patients.It demonstrates good agreement with conventional methods,alongside robust stability and efficient segmentation performance.关键词
常染色体显性多囊肾病/肾脏总体积/人工智能/深度学习/磁共振成像Key words
Autosomal dominant polycystic kidney disease/Total kidney volume/Artificial intelligence/Deep learning/Magnetic resonance imaging引用本文复制引用
黄力敏,徐辉景,李志平,李伟伟,余德洪,崔凤..基于MRI的AI分割算法精准评估常染色体显性多囊肾病患者肾体积的研究[J].浙江医学,2025,47(20):2182-2186,5.基金项目
浙江省中医药科技计划项目(2025ZL083) （2025ZL083）
杭州市生物医药和健康产业发展扶持科技专项立项项目(2024WJC163) （2024WJC163）
