College of Biomedical Engineering and Instrument Science, Zhejiang University, Hangzhou, Zhejiang, China.
School of Medicine, Sir Run Run Shaw Hospital, Department of Endocrinology, Zhejiang University, Hangzhou, Zhejiang, China.
J Magn Reson Imaging. 2023 Jan;57(1):296-307. doi: 10.1002/jmri.28275. Epub 2022 May 30.
Pancreatic fat accumulation may cause or aggravate the process of acute pancreatitis, β-cell dysfunction, T2DM disease, and even be associated with pancreatic tumors. The pathophysiology of fatty pancreas remains overlooked and lacks effective imaging diagnostics.
To automatically measure the distribution of pancreatic fat deposition on Dixon MRI in multicenter/population datasets using nnU-Net models.
Retrospective.
A total of 176 obese/nonobese subjects (90 males, 86 females; mean age, 27.2 ± 19.7) were enrolled, including a training set (N = 132) and a testing set (N = 44).
FIELD STRENGTH/SEQUENCE: A 3 T and 1.5 T/gradient echo T dual-echo Dixon.
The segmentation results of four types of nnU-Net models were compared using dice similarity coefficient (DSC), positive predicted value (PPV), and sensitivity. The ground truth was the manual delineation by two radiologists according to in-phase (IP) and opposed-phase (OP) images.
The group difference of segmentation results of four models were assessed by the Kruskal-Wallis H test with Dunn-Bonferroni comparisons. The interobserver agreement of pancreatic fat fraction measurements across three observers and test-retest reliability of human and machine were assessed by intragroup correlation coefficient (ICC). P < 0.05 was considered statistically significant.
The three-dimensional (3D) dual-contrast model had significantly improved performance than 2D dual-contrast (DSC/sensitivity) and 3D one-contrast (IP) models (DSC/PPV/sensitivity) and had less errors than 3D one-contrast (OP) model according to higher DSC and PPV (not significant), with a mean DSC of 0.9158, PPV of 0.9105 and sensitivity of 0.9232 in the testing set. The test-retest ICC of this model was above 0.900 in all pancreatic regions, exceeded human.
3D Dual-contrast nnU-Net aided segmentation of pancreas on Dixon images appears to be adaptable to multicenter/population datasets. It fully automates the assessment of pancreatic fat distribution and has high reliability.
3 TECHNICAL EFFICACY: Stage 3.
胰腺脂肪堆积可能导致或加重急性胰腺炎、β细胞功能障碍、T2DM 疾病的发生发展，甚至与胰腺肿瘤相关。脂肪胰腺的病理生理学仍然被忽视，缺乏有效的影像学诊断。
使用 nnU-Net 模型自动测量多中心/人群数据集 Dixon MRI 上胰腺脂肪沉积的分布。
回顾性。
共纳入 176 名肥胖/非肥胖受试者（90 名男性，86 名女性；平均年龄 27.2±19.7），包括训练集（N=132）和测试集（N=44）。
磁场强度/序列：3T 和 1.5T/梯度回波 T 双回波 Dixon。
使用 Dice 相似系数（DSC）、阳性预测值（PPV）和敏感度比较四种类型 nnU-Net 模型的分割结果。金标准是两位放射科医生根据同相位（IP）和反相位（OP）图像进行的手动勾画。
采用 Kruskal-Wallis H 检验和 Dunn-Bonferroni 比较评估四种模型的分割结果的组间差异。通过组内相关系数（ICC）评估三位观察者之间的胰腺脂肪分数测量的观察者间一致性和人机测试-再测试可靠性。P<0.05 被认为具有统计学意义。
三维（3D）双对比模型的性能明显优于二维（DSC/敏感度）和 3D 单对比（IP）模型（DSC/PPV/敏感度），并且与 3D 单对比（OP）模型相比，错误更少，测试集的 DSC 为 0.9158，PPV 为 0.9105，敏感度为 0.9232。该模型在所有胰腺区域的测试-再测试 ICC 均高于 0.900，超过了人类。
Dixon 图像上的 3D 双对比 nnU-Net 辅助胰腺分割似乎适用于多中心/人群数据集。它可以完全自动评估胰腺脂肪分布，具有很高的可靠性。
3 级技术功效：3 级。
