深度神经拼块模型通过增强MRI知识迁移从平扫MRI预测肾脏影像生物标志物
本研究纳入200例具有配对动脉期增强与平扫T1加权MRI的患者,将增强MRI上手动分割的皮质、髓质、肾窦标签刚性配准至平扫MRI作为参考标准。采用分层三维Deep Neural Patchworks模型训练并独立测试,评估分割Dice系数与体积一致性。结果显示平扫MRI可实现准确的全肾体积测量,为无需对比剂的肾脏生物标志物提取提供支持。
要点
- 平扫T1加权MRI经增强标签迁移训练后,全肾分割Dice达0.950(左)与0.953(右)。
- 全肾体积与参考标准一致性高,MAE为8.76 mL(占均值2.5%),CCC为0.983。
- 模型对皮质体积轻度高估、髓质体积低估,临床应用时需注意分区比例偏移。
核心结论
- 增强监督的知识迁移使常规平扫MRI能够实现准确的全肾体积测量,支持在无明显结构变形的肾脏中进行无对比剂生物标志物提取。
研究设计与样本
- 回顾性单中心研究,纳入200例具有配对动脉期增强与平扫T1加权MRI的患者。
- 在增强MRI上手动分割皮质、髓质、肾窦,刚性配准至平扫MRI作为体素级参考标签。
- 分层三维Deep Neural Patchworks模型在100例上训练(90例训练/10例验证),在独立100例测试集上评估。
- 参考标准为转移至平扫MRI的增强掩膜,性能评估采用分割Dice系数与体积一致性指标(Pearson、MAE、Lin CCC、Bland-Altman)。
关键影像表现/征象
- 增强MRI提供的皮髓质对比可用于肾皮质、髓质、肾窦的区分,作为平扫MRI模型训练的监督信号。
- 平扫T1加权MRI经模型自动分割后,全肾轮廓与增强转移参考掩膜高度一致,左肾Dice为0.950,右肾为0.953。
诊断标准与关键数值
- 全肾分割性能:左肾Dice系数0.950,右肾Dice系数0.953。
- 全肾体积一致性:MAE 8.76 mL(占均值2.5%),Lin CCC 0.983,偏倚-1.56 mL,95%一致性界限为-28.81至25.69 mL。
- 肾窦体积保持较高一致性,尽管其Dice离散度较大。
- 皮质体积轻度高估、髓质体积低估,预测分区比例与参考值在上述数值范围内偏移。
临床应用与阅片要点
- 常规平扫T1加权MRI可用于自动提取全肾体积,适用于增强检查有禁忌或不适宜的患者。
- 当报告基于自动分割的皮质与髓质体积时,应注意皮质高估与髓质低估的系统性偏移。
- 该方法适用于无明显结构变形的肾脏,阅片时应先排除重度肾积水、较大占位或术后改变等情况。
局限性
- 研究为回顾性单中心设计,样本量为200例,模型泛化性能尚未在多中心前瞻性数据中验证。
- 参考标准来自增强MRI手动分割后刚性转移至平扫MRI,配准误差可能影响标签准确性。
- 结论仅适用于无明显结构变形的肾脏,未评估在复杂肾脏病变中的适用性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
RATIONALE AND OBJECTIVES: Contrast-enhanced (CE) MRI provides clear corticomedullary contrast for renal compartment delineation but may be contraindicated or undesirable in routine practice. We aimed to enable automated extraction of renal imaging biomarkers from routine non-contrast-enhanced (NCE) T1-weighted MRI by transferring CE-derived compartment labels.
MATERIALS AND METHODS: This retrospective single-center study (January 2017 to December 2021) included 200 patients with paired arterial-phase CE and NCE T1-weighted MRI. Cortex, medulla, and sinus were manually segmented on CE MRI and rigidly transferred to NCE MRI to provide voxel-level reference labels. A hierarchical 3D Deep Neural Patchworks (DNP) model was trained on 100 examinations (90 training/10 validation) and evaluated on an independent test set of 100 examinations using the transferred CE masks on NCE as reference. Performance was assessed using Dice similarity of segmentations and biomarker agreement using volumes (Pearson, MAE, Lin's CCC, and Bland-Altman).
RESULTS: Whole-kidney segmentation Dice was 0.950 (left) and 0.953 (right). Total kidney volume showed high agreement with minimal bias (MAE 8.76 mL, 2.5% of mean; CCC 0.983; bias -1.56mL; 95% limits of agreement -28.81 to 25.69 mL). Cortex volume was modestly overestimated and medulla volume underestimated, shifting predicted compartment fractions toward cortex (74.7% vs. 72,1% in ground truth; medulla 21.5% vs. 24.3%; sinus 3.8% vs. 3.6%. Sinus volume maintained high concordance despite higher Dice dispersion.
CONCLUSION: CE-supervised knowledge transfer enables accurate, well-calibrated total kidney volumetry from routine NCE MRI and supports contrast-free renal biomarker extraction in kidneys without major structural distortion.
引用
- Kästingschäfer, Fink, Rau, et al. Deep Neural Patchworks Predict Renal Imaging Biomarkers From Non-contrast MRI via Knowledge Transfer From Arterial-Phase Contrast-Enhanced MRI. Acad Radiol. 2026. PMID: 42509052. doi:10.1016/j.acra.2026.06.071
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。