Kaandorp Misha P T, Zijlstra Frank, Federau Christian, While Peter T
Department of Radiology and Nuclear Medicine, St. Olav's University Hospital, Trondheim, Norway.
Department of Circulation and Medical Imaging, NTNU - Norwegian University of Science and Technology, Trondheim, Norway.
Magn Reson Med. 2023 Jul;90(1):312-328. doi: 10.1002/mrm.29628. Epub 2023 Mar 13.
The development of advanced estimators for intravoxel incoherent motion (IVIM) modeling is often motivated by a desire to produce smoother parameter maps than least squares (LSQ). Deep neural networks show promise to this end, yet performance may be conditional on a myriad of choices regarding the learning strategy. In this work, we have explored potential impacts of key training features in unsupervised and supervised learning for IVIM model fitting.
Two synthetic data sets and one in-vivo data set from glioma patients were used in training of unsupervised and supervised networks for assessing generalizability. Network stability for different learning rates and network sizes was assessed in terms of loss convergence. Accuracy, precision, and bias were assessed by comparing estimations against ground truth after using different training data (synthetic and in vivo).
A high learning rate, small network size, and early stopping resulted in sub-optimal solutions and correlations in fitted IVIM parameters. Extending training beyond early stopping resolved these correlations and reduced parameter error. However, extensive training resulted in increased noise sensitivity, where unsupervised estimates displayed variability similar to LSQ. In contrast, supervised estimates demonstrated improved precision but were strongly biased toward the mean of the training distribution, resulting in relatively smooth, yet possibly deceptive parameter maps. Extensive training also reduced the impact of individual hyperparameters.
Voxel-wise deep learning for IVIM fitting demands sufficiently extensive training to minimize parameter correlation and bias for unsupervised learning, or demands a close correspondence between the training and test sets for supervised learning.
开发用于体素内不相干运动（IVIM）建模的先进估计器，通常是出于想要生成比最小二乘法（LSQ）更平滑的参数图的愿望。深度神经网络有望实现这一目标，但其性能可能取决于关于学习策略的众多选择。在这项工作中，我们探讨了无监督和监督学习中关键训练特征对IVIM模型拟合的潜在影响。
使用来自胶质瘤患者的两个合成数据集和一个体内数据集来训练无监督和监督网络，以评估泛化能力。根据损失收敛评估不同学习率和网络规模下网络的稳定性。在使用不同训练数据（合成数据和体内数据）后，通过将估计值与真实值进行比较来评估准确性、精确性和偏差。
高学习率、小网络规模和提前停止会导致拟合的IVIM参数出现次优解和相关性。在提前停止后继续训练可解决这些相关性并减少参数误差。然而，大量训练会导致噪声敏感性增加，其中无监督估计显示出与LSQ相似的变异性。相比之下，监督估计显示出更高的精确性，但强烈偏向训练分布的均值，导致参数图相对平滑但可能具有欺骗性。大量训练还减少了各个超参数的影响。
用于IVIM拟合的逐体素深度学习需要足够广泛的训练，以最小化无监督学习中的参数相关性和偏差，或者需要监督学习中训练集和测试集之间的紧密对应。
